JP4709744B2 - Electronic compass system - Google Patents

Description

  The present invention relates generally to electronic compass for vehicles, and more particularly to an electronic compass with improved data filtering and / or heading determination.

Cross-reference to related applications :
This application is filed on August 2, 2002, claiming priority under 35 USC §119 (e) for US Provisional Patent Application No. 60 / 360,723, filed March 1, 2002 Which is a continuation-in-part of US patent application Ser. No. 10 / 210,910.

  This application also claims priority under 35 USC §119 (e) for US Provisional Patent Application No. 60 / 449,828, filed February 24, 2003.

  The entire disclosure of each of the above patent applications is incorporated herein by reference.

  Electronic compass is becoming increasingly popular as an accessory in automobiles. The general structure of a typical electronic compass circuit 10 is shown in FIG. More specifically, a typical electronic compass circuit includes a magnetic sensor circuit 12 including a Y-axis sensor 13 and an X-axis sensor 14. The magnetic sensor circuit 12 is coupled to a processing circuit 15, which operates under the control of software code to process the data supplied by the sensor circuit 12, and based on such processing The compass circuit is calibrated and the vehicle heading is determined based on the data provided by the sensor circuit 12. The processing circuit 15 is coupled to a non-volatile memory 16 that stores calibration data so that the compass does not need to be recalibrated every ignition cycle. The calibrated vehicle heading is transmitted from the processing circuit 15 to the heading display 18 and displayed to the vehicle occupant. The heading display is typically incorporated into an overhead console or rearview mirror assembly. A user input switch 20 may also be provided, which allows the user to interact with the processing circuit 15 to change the information displayed on the display 18 and allow manual recalibration. , And / or the geographic zone in which the vehicle is currently traveling can be entered. In addition, a power supply circuit 22 is also provided for receiving 12 volt power from the vehicle battery or ignition and converting that power to power levels useful for various components of the compass circuit 10.

  In this conventional system, a Y-axis sensor 13 is provided to sense a magnetic field perpendicular to the traveling direction of the vehicle, while an X-axis sensor 14 is provided to sense a magnetic field that is in line with the traveling direction of the vehicle. Both sensors 13 and 14 are usually mounted parallel to the ground surface. In such an attached state, when the magnetic field component is not sensed by the Y-axis sensor 13 and the positive magnetic field component is sensed by the X-axis sensor 14, the processing circuit 15 determines that the vehicle is facing north. To do. Similarly, if the magnetic field component is not detected by the Y-axis sensor 13 and the negative magnetic field component is detected by the X-axis sensor 14, the processing circuit 15 determines that the vehicle is facing south. Similarly, if the magnetic field component is not detected by the X-axis sensor and the positive magnetic field component is detected by the Y-axis sensor, the processing circuit 15 determines that the vehicle is facing east. If no magnetic field component is detected by the X-axis sensor and a negative magnetic field component is detected by the Y-axis sensor, the processing circuit 15 determines that the vehicle is facing west. If equal positive magnetic field components are sensed by both the X and Y axis sensors, the processing circuitry determines that the vehicle is facing northeast. If equal negative magnetic field components are sensed by both the X and Y axis sensors, the processing circuit determines that the vehicle is facing southwest. If the positive magnetic field component sensed by the X-axis sensor is equal to the absolute value of the negative magnetic field component sensed by the Y-axis sensor, the processing circuit determines that the vehicle is facing northwest. If the absolute value of the negative magnetic field component sensed by the X-axis sensor is equal to the value of the positive magnetic field component sensed by the Y-axis sensor, the processing circuit determines that the vehicle is facing southeast. If the output level of the magnetic sensor is plotted against the X and Y axes when the vehicle turns through a 360 ° loop under an ideal environment, as shown as circle A in FIG. The plot forms a circle.

  Such an electronic compass generally displays only eight different headings (N, NE, E, SE, S, SW, W and NW), and the magnetic field components sensed by the X and Y axis sensors are always Since it is non-zero and not always equal, the compass processor generally calculates a heading angle φ relative to the X and Y axes, and this heading angle is an angle threshold that defines the boundary between each of the eight different heading representations. Compare with Thus, the circular plot A is effectively divided into eight 45 ° angle segments corresponding to eight different display headings, as shown in FIG. Thus, the compass processing circuit only determines in which segment the heading angle φ is present in order to determine which of the eight headings should be displayed.

  As described above, the ideal environment is that the output levels of the X and Y axis sensors 13 and 14 form a circular plot A for the X and Y axis sensors, with the complete circle center at the origin of the coordinate system. When it comes. In practice, however, the plot of the output of the X and Y sensors in the X and Y coordinate planes does not form a perfect circle, and the center of such a circle often does not coincide with the origin of the coordinate plane. . More specifically, the plot may be slightly oval and the X and Y directions may deviate from the origin as shown by plot B in FIG. When the actual plot is not a perfect circle and its center point deviates from the origin, the processing circuit cannot determine a suitable heading using a simple heading angle calculation. Such deviations and distortions in the circular plot are usually caused by the influence of the vehicle's ferrous material that can alter the magnetic field sensed by the X and Y axis sensors. To facilitate heading calculations, the compass circuit is calibrated to take into account the effect of the vehicle on the sensed magnetic field.

  Not only does the compass circuit have to be calibrated first, but the compass circuit is also affected by the fact that the effect on the magnetic field caused by the vehicle's ferrous material changes over time and because of the only temporary effect on the magnetic field from the outside. You must also recalibrate continuously. For example, adding an antenna that attaches to the roof can cause fluctuations in the magnetic field readings when passing objects with large amounts of iron material, such as near railroad tracks, bridges or large buildings, or when a vehicle passes through a car wash. May cause. Therefore, calibration of electronic compass circuits and continuous recalibration have received much attention.

U.S. Pat. No. 4,953,305 to Van Lente et al. Describes an electronic compass system having an automatic continuous configuration function. This patent discloses a calibration technique that accumulates data from sensors and converts them to data points on an XY coordinate plane as the vehicle travels through multiple 360 ° loops. The processing circuit has a cumulative data maximum value along the Y axis (Y _max ), a minimum value along the Y axis (Y _min ), a maximum value along the X axis (X _max ), and a minimum along the X axis. Determine the value (X _min ). From the maximum and minimum values along the X axis, the span along the X axis can be calculated between X _min and X _max . Similarly, from the maximum and minimum values along the Y axis, the span along the Y axis between Y _min and Y _max can be calculated. If these spans are not equal, the processing circuitry can adjust the gain of one or both of the X-axis and Y-axis sensors until the time that these spans are equal to each other. This process is performed to convert an elliptical plot of the data into a circular plot of the data before further processing. Subsequently, the center point (X _E , Y _E ) of plot B is calculated using the maximum and minimum values from the X and Y sensors (see FIG. 2). X and Y error values (X _E and Y _E ) are then calculated and then used to shift each data point as it is received from the X and Y sensors, respectively. When the compass is first calibrated, it will continue to automatically recalibrate based on the maximum and minimum values accumulated along the X and Y axes.

  One problem with the automatic calibration routine disclosed in the '305 patent is that the vehicle generally travels through a number of 360 ° loops to obtain enough data for the system to be confident that the calibration is accurate. That is what you need. This poses the problem to the vehicle manufacturer that each vehicle must be driven through a number of loops before the vehicle is loaded onto the vehicle carrier and delivered to the dealer. Unfortunately, assembly plants often do not have enough space to operate each vehicle in such a loop, and if so, the process takes valuable time. If the vehicle is delivered to a dealer without being driven through a sufficient loop, the customer will either purchase the vehicle in a non-compassed state or otherwise test drive the vehicle. In this case, the customer mistakenly believes that the compass is malfunctioning and therefore makes an unnecessary warranty claim against the compass.

  A number of patents disclose various solutions to the problem. US Pat. No. 6,192,315 to Jeshki et al. First calibrates the compass based on the expected vehicle magnetism for the particular model on which the compass is installed before installing it on the vehicle. A calibration routine is disclosed. This initial calibration is used until the vehicle acquires sufficient data by traveling through a number of 360 ° loops. When sufficient data is obtained, the compass is switched to the more recently acquired calibration data, and the compass is then continuously recalibrated using the technique of the '305 patent.

  Olson et al., US Pat. No. 5,737,226, discloses a calibration technique in which raw data obtained from a sensor determines whether a processing circuit determines that the compass is no longer accurately calibrated. is doing. In this case, the processing circuit uses the assumed radius to obtain two endpoints that are separated by a predetermined angle or more. Using that assumed radius, two potential center points for the circle are given. The '226 patent obtains an intermediate data point between two endpoints and uses it to use which of the two center points to calibrate and subsequently determine the vehicle heading It is disclosed to identify what to do.

  US Pat. No. 6,301,794 to Parks et al. Discloses a calibration routine that recalibrates the compass whenever three data points are obtained that meet specified criteria. Once three data points are obtained that meet the specified criteria, including the averaging and spacing criteria, the center of the circle is calculated using the equation for the circle to ensure that the circle contains three data points.

  Al-Ater, U.S. Pat. No. 4,807,462, discloses a compass calibration routine that calibrates a compass based on the acquisition of three data points. The center of the circle used for calibration is determined by determining the intersection of two perpendicular bisectors of two lines connecting adjacent points of three data points.

  Each of the above-mentioned patents discloses a calibration routine that calibrates the compass more quickly, but some of the disclosed techniques recalibrate too often, and therefore, calibration errors due to temporary magnetic field disturbances. Either it is overactive in that it tends to be, or it does not respond quickly enough to changes in the field dispersion of a more permanent nature. In addition, each of the calibration routines described above calculates the center of the circle assuming that three to four points are placed exactly around the circle. As will be explained in detail below, it is conceivable that any of these points will deviate from the circumference of a circle that actually fits well with the data obtained. Furthermore, none of the above-mentioned patents disclose a calibration routine that takes into account the vehicle pitch or the strength of the vertical component of the geomagnetic field vector. Thus, if any of the above-mentioned compass is installed in a structure that is movable with respect to the vehicle, such as the housing of a rearview mirror assembly, these systems are resistant to movement of the housing. Cannot provide a quick and accurate response.

  US Pat. Nos. 6,023,229 and 6,140,933 commonly assigned to Bagno et al. Attach a compass sensor to a rearview mirror housing and attach the rearview mirror housing to the vehicle to which it is attached. Various techniques are disclosed that allow pivoting horizontally and vertically. More particularly, various mechanisms have been disclosed for detecting the mirror housing and thus the sensor when it is tilted. When the mirror housing tilt is detected, a signal is sent to the compass processing circuit to indicate that the tilt has occurred, otherwise the processing circuit does not assume that there is a sudden change in the magnetic field vector. The processing circuit then determines a vector of differences between the data points obtained immediately before the slope signal and the data points obtained immediately after the slope signal, which are used for the error compensation signal. The '229 patent discloses a mechanism in which a third magnetic sensor aligned with the Z axis is provided. The output of the Z-axis sensor is used to determine if a tilt has occurred when a sudden change in the output of the X and Y sensors is first sensed. In response to such a sudden change in the X and Y sensor outputs, the processing circuit identifies the error vector or resumes calibration based on whether a sudden change is also detected in the Z-axis sensor. . However, this compass system does not use a Z-axis sensor to determine heading or identify the center of the circle used for calibration.

  In accordance with a first embodiment of the present invention, a vehicle electronic compass system senses three components of a geomagnetic field vector and generates a magnetic signal representative of the sensed three components. A circuit, a pitch and roll sensing circuit for measuring the pitch and roll of the magnetic sensor circuit, and a processing circuit coupled to the magnetic sensor circuit and the pitch and roll sensing circuit, receiving an output signal, Compensating sensed components for the measured pitch and roll, calculating vehicle headings as a function of at least two of the compensated sensed components, and calculating the calculated headings And a processing circuit for generating a heading signal representing.

  In accordance with another embodiment of the present invention, an electronic compass system for a vehicle senses at least two components of a geomagnetic field vector and generates a magnetic signal representative of the sensed at least two components. A sensor circuit and a processing circuit coupled to the magnetic sensor circuit, receiving an output signal from the magnetic sensor circuit, determining a relative strength of a geomagnetic field vector, and generating an output signal received from the magnetic sensor circuit; Determine if there is too much noise as a function of the relative strength of the geomagnetic field vector, and if there is not too much noise in the output signal, the vehicle as a function of the sensed component And the calculated heading or previous heading if there is a lot of noise in the output signal. It comprises a processing circuit for generating a to heading signal.

  According to another embodiment of the present invention, an electronic compass assembly for a vehicle includes a circuit board that defines a plane corresponding to a mounting surface, and a magnetic sensor circuit that is mounted on the circuit board. A magnetic sensor circuit for sensing at least two components and generating an output signal representative of the sensed at least two components, the magnetic sensor circuit comprising at least two magnetic field sensing elements each having a sensitivity axis And at least one of the magnetic field sensing elements has a sensitivity axis that is (a) non-perpendicular and non-parallel to the plane of the circuit board, and (b) another one of the at least two magnetic field sensing elements A processing circuit arranged to face one of two directions, non-perpendicular to the sensitivity axis, and further coupled to the magnetic sensor circuit for receiving an output signal The heading of the vehicle is calculated as a function of the sensed components, and also has a processing circuit for generating a heading signal representing the computed heading.

  According to another embodiment of the present invention, a vehicle rearview mirror assembly is configured to be mounted on a vehicle, and includes a mirror mounting structure having a mirror housing, a mirror mounted on the mirror housing, and the mirror mounting structure. A circuit board supported by the body and a magnetic sensor circuit attached to the circuit board for sensing at least two components of the geomagnetic field vector and generating an output signal representative of the sensed at least two components A magnetic sensor circuit including three magnetic field sensing elements housed in a common integrated package, and a plurality of leads for attachment to the circuit board extending therefrom, and coupled to the magnetic sensor circuit A processing circuit that receives an output signal from the magnetic sensor circuit and functions as a function of the sensed component. The calculated Ddingu and has and a processing circuit for generating a heading signal representing the computed heading.

  According to another embodiment of the invention, an electronic compass subassembly for a vehicle includes a circuit board that includes a connector for connecting electrical components, and is attached to and electrically coupled to the circuit board. A processing circuit for communicating with the electrical component, further coupled to the magnetic sensor circuit for receiving an output signal representative of at least two components of the geomagnetic field vector as a function of the sensed component Processing circuitry for calculating a vehicle heading and generating a heading signal representative of the calculated heading, the processing circuitry using at least two different signal formats associated with different types of electronic components Electrical components configured to communicate with each other and connected to the connector Electronic compass subassembly capable of selecting a particular signal format used by the processing circuit to communicate with and.

  According to another embodiment of the present invention, an electronic compass system for a vehicle senses at least two components of a geomagnetic field vector and generates a magnetic signal representative of the sensed at least two components. A sensor circuit and a processing circuit coupled to the magnetic sensor circuit, receiving an output signal from the magnetic sensor circuit, calculating a vehicle heading as a function of its sensed component, and calculating the calculated heading A processing circuit coupled to the processing circuit for receiving the heading signal and displaying the calculated heading, the processing circuit being configured to generate a heading signal representative of at least one geographic region Displays the geographic display and the geographical magnetic dispersion zone within the geographic area where the compass is currently calibrated Further comprising a, a display configured to.

  According to another embodiment of the present invention, an electronic compass system for a vehicle senses at least two components of a geomagnetic field vector and generates a magnetic signal representative of the sensed at least two components. A sensor circuit and a processing circuit coupled to the magnetic sensor circuit, receiving an output signal from the magnetic sensor circuit, calculating a vehicle heading as a function of its sensed component, and calculating the calculated heading A processing circuit for generating a heading signal representative of the vehicle, the processing circuit being coupled to an electrical component of the vehicle for receiving an indication of a time zone in which the vehicle is currently traveling, and further comprising the processing circuit Compensates the heading calculation as a function of the geographical magnetic dispersion for the time zone in which the vehicle is currently traveling.

  In accordance with another embodiment of the present invention, an electronic compass system for a vehicle senses at least two components of a geomagnetic field vector and generates a magnetic signal representative of the sensed at least two components. A sensor circuit and a processing circuit coupled to the magnetic sensor circuit, receiving an output signal from the magnetic sensor circuit, calculating a vehicle heading as a function of the sensed component, and calculating the calculated heading Generating a heading signal representative, storing the calculated heading history samples at periodic intervals, and analyzing the stored history samples to determine which vehicle has the highest value in any of the four sets of sensed component values. The value of the sensed component that the processing circuit identified as corresponding to the headings of east, west, south, and north Processing circuitry for calculating the variance of four sets of values of the sensed component from which the vehicle travels most frequently and compensating for the heading calculation as a function of the calculated variance; I have.

  According to another embodiment of the present invention, an electronic compass system for a vehicle senses at least two components of a geomagnetic field vector and generates a magnetic signal representative of the sensed at least two components. A sensor circuit and a processing circuit coupled to the magnetic sensor circuit, receiving an output signal from the magnetic sensor circuit, and deriving a noise level from the output signal received from the magnetic sensor circuit. Calculate as a function of the root mean square value, and if the noise level exceeds the threshold noise level, determine if there is significantly more noise in the output signal received from the magnetic sensor circuit, If a lot of noise is not present in the output signal, the vehicle heading is calculated as a function of the sensed component and If significantly more noise Te is present in the output signal includes a processing circuit for generating a heading signal representing the its computed heading or previous heading, the.

  According to another embodiment of the present invention, an electronic compass system for a vehicle senses at least two components of a geomagnetic field vector and generates a magnetic signal representative of the sensed at least two components. A sensor circuit and a processing circuit coupled to the magnetic sensor circuit, receiving an output signal from the magnetic sensor circuit, and deriving a noise level from the output signal received from the magnetic sensor circuit Calculated as a function of the mean square error of the value, and if the noise level exceeds the threshold noise level, determine if there is significantly more noise in the output signal received from the magnetic sensor circuit; If significant noise is not present in the output signal, vehicle heading is calculated as a function of the sensed component. And a processing circuit for generating a heading signal that represents the calculated heading or previous heading if significant noise is present in the output signal.

  According to another embodiment of the present invention, the magnetic sensor comprises: a conductive windshield in which conductive materials are combined; and an electronic compass system having a magnetic sensor circuit disposed in the vicinity of the conductive windshield. In a vehicle in which the circuit senses at least two components of the geomagnetic field vector and generates an output signal representative of the sensed at least two components, the improvement is a processing circuit coupled to the magnetic sensor circuit. A processing circuit for sampling an output signal from the magnetic sensor circuit, calculating a vehicle heading as a function of a sensed component, and generating a heading signal representative of the calculated heading; and Compensation to compensate for the effect of the electric field generated by the conductive windshield on the magnetic field sensed by the magnetic sensor circuit And a stage.

  According to another embodiment of the present invention, a method of attaching a magnetic sensing element to a circuit board includes attaching a first magnetic sensing element to a first portion of the circuit board, and a second to a second part of the circuit board. Attaching a magnetic sensing element; bending a second portion of the circuit board to redirect the second magnetic sensing element relative to the first magnetic sensing element; Securing two portions relative to the first portion of the circuit board to maintain the redirected positions of the first and second magnetic sensing elements.

  In accordance with another embodiment of the present invention, an electronic compass system for a vehicle senses at least two components of a geomagnetic field vector and generates a magnetic signal representative of the sensed at least two components. A sensor circuit and a processing circuit coupled to the magnetic sensor circuit are provided. The processing circuit receives an output signal from the magnetic sensor circuit, calculates a vehicle heading as a function of its sensed component, generates a heading signal representative of the calculated heading, monitors the output signal, Identifies vehicle heading changes that represent approximately 90 ° vehicle turn, stores the sensed components for vehicle headings that traveled immediately before and after approximately 90 ° vehicle turn, and processing circuitry supports east, west, north, and south headings To calculate a variance of the stored sensed component from the four sets of sensed component values identified as to be compensated and to compensate the heading calculation as a function of the calculated variance Is provided.

  In accordance with another embodiment of the present invention, an electronic compass system for a vehicle senses at least two components of a geomagnetic field vector and generates a magnetic signal representative of the sensed at least two components. A sensor circuit and a processing circuit coupled to the magnetic sensor circuit, receiving an output signal from the magnetic sensor circuit, calculating a vehicle heading as a function of the sensed component, and calculating the calculated heading A processing circuit for generating a heading signal representative of the vehicle, wherein the processing circuit determines where in the plurality of geographic zones the vehicle is currently traveling and for the geographic zone in which the vehicle is currently traveling Compensates for the heading calculation as a function of geographic magnetic dispersion, and the processing circuit includes a vehicle heading and each vehicle heading. By monitoring the transit time in grayed, by monitoring the change in the vehicle position for a plurality of geographic zones, which geographic zone the vehicle is determined whether the currently traveling.

  In accordance with another embodiment of the present invention, an electronic compass system is provided for use in a vehicle having a vehicle device that comprises one of a convertible top and a sunroof. The electronic compass system includes a magnetic sensor circuit for sensing at least two components of the geomagnetic field vector and generating an output signal representative of the sensed at least two components, and a process coupled to the magnetic sensor circuit. A processing circuit for sampling an output signal from a magnetic sensor circuit, calculating a vehicle heading as a function of the sensed component, and generating a heading signal representative of the calculated heading; Compensation means for compensating for the effect of changes in the open or closed state of the vehicle apparatus on the magnetic field sensed by the magnetic sensor circuit.

  In accordance with another embodiment of the present invention, an electronic compass system is provided for use with a vehicle having at least one vehicle accessory that may adversely affect the operation of the compass system. The electronic compass system includes a magnetic sensor circuit for sensing at least two components of the geomagnetic field vector and generating an output signal representative of the sensed at least two components, and a process coupled to the magnetic sensor circuit. A processing circuit for sampling an output signal from the magnetic sensor circuit, calculating a vehicle heading as a function of the sensed component, and generating a heading signal representative of the calculated heading; The processing circuit immediately compensates for the effect of the vehicle accessory on the magnetic field sensed by the magnetic sensor circuit when detecting the effect of the vehicle accessory on the magnetic field without requiring a signal from the vehicle accessory. To do.

  These and other features, advantages and objects of the present invention will be further understood and apparent to those skilled in the art from the following description, claims and appended drawings.

  FIG. 3 shows an electronic compass circuit 100 constructed in accordance with the present invention. The compass circuit 100 senses the vertical component of the geomagnetic field vector and has a magnetic sensor (or magnetometer) circuit 102 having at least two sensor elements 104 and 106 for generating an output signal representative of the sensed component. It has. In certain embodiments of the present invention, a third sensor 108 is preferably provided to sense the third vertical component of the geomagnetic field vector. The electronic compass circuit 100 further includes a processing circuit 110 coupled to the magnetic sensor circuit 102 for receiving an output signal from the magnetic sensor circuit 102.

  The processing circuit 110 calculates the vehicle heading as a function of the sensed component of the geomagnetic vector. The processing circuit 110 also generates a heading signal representing the calculated heading. This heading signal can also be supplied to other electronic components of the vehicle and is preferably supplied to a heading indicating device 114 that provides an indication of the vehicle's heading to the vehicle occupant. The heading indication device can be configured to provide a visual or audible indication of vehicle heading. Preferably, the heading indication device 114 is a display device for visually displaying the heading of the vehicle. However, the heading signal generated by the processing circuit 110 is supplied to other electrical components in the vehicle, such as a navigation system, via a separate wiring or via the vehicle bus interface 120 coupled to the vehicle bus. Obviously you can. Such a navigation system can, for example, display the headings directly or change the direction of the displayed map based on the current heading of the vehicle. The heading signal may be sent to the navigation system, may be sent to a heading indication device (ie display), or another vehicle via a dedicated line, vehicle or local area bus, eg LIN or CAN bus. It may be sent to accessories or components. An interface such as an RS485 interface may be used to connect to a dedicated line.

  As described above, the heading signal is preferably supplied to a heading indication device 114 in the form of a display device. The display device can display the headings in an alphanumeric format (ie, N, NE, E, SE, S, SW, W, and NW). An example of a suitable display is disclosed in commonly assigned US Pat. No. 6,346,698, the entire disclosure of which is incorporated herein by reference. The display disclosed in this patent not only displays heading information, but also displays the current external temperature provided by the external temperature sensor 122 at the same time. Alternatively, if temperature or other information is to be provided to the same display rather than simultaneously, a suitable conventional user input switch 116 is provided so that the user can compass, temperature and / or other display information. May be scrolled, toggled between them, or selectively activated and deactivated.

  Another form of display is disclosed in commonly assigned US Pat. No. 6,356,376, the entire disclosure of which is incorporated herein by reference. More particularly, a graphic compass display for providing heading instructions in a graphic format is disclosed. Any of the other forms of display disclosed in this patent may be used, including vacuum fluorescent displays, LED displays, organic LED displays, liquid crystal displays, light emitting polymer displays, and the like.

  The magnetic sensor circuit 102 may be of any conventional structure modified to provide the output from the additional magnetic sensor 108 as appropriate. The magnetic sensor circuit may be a magnetic flux gate, magnetic induction, magnetic impedance, or magnetoresistive sensor. An example of a magnetoresistive sensor is disclosed in US Pat. No. 5,632,092, the entire disclosure of which is incorporated herein by reference. Preferably, the magnetic sensor circuit 102 is a commonly assigned US patent entitled “MAGNETOMETER HAVING A DYNAMICALLY ADJUSTABLE BIAS SETTING AND ELECTRONIC VEHICLE COMPASS INCORPORATING THE SAME” filed on November 20, 2001. It is configured based on any embodiment disclosed in US Pat. No. 6,653,831. Note that the micro-processing circuit used in the magnetic sensor circuit can be suitably programmed to integrate with such a circuit in processing circuit 110 to perform all the necessary functions of both circuits. Alternatively, the microprocessing circuitry can be maintained separately.

  The electronic compass circuit 100 may further include a non-volatile memory 112 that may be external to the processing circuit 110 or incorporated into one of its components. The processing circuit 110 uses the non-volatile memory 112 to provide the best fit approximation data and other information that needs to survive between ignition cycles when the system is powered by a vehicle ignition. Remember.

  The electronic compass circuit 100 can also be coupled to a global positioning system (GPS) receiver 118. Information from the GPS 118 may be sent directly to the processing circuit 110 or may be sent indirectly via the vehicle bus and bus interface 120 or via the local area bus. Information from GPS 118 or information from other positioning systems such as GLONASS or LORAN may be used by processing circuit 110 for various purposes. More particularly, this information can be used to identify in which geographic zone the vehicle is currently traveling, so that the processing circuit 110 can detect the actual N sensed in that particular geographic zone. An appropriate magnetic field offset corresponding to the magnetic N pole offset from the pole can be used. In addition, information from the GPS 118 can be used to derive vehicle speed, distance traveled, or simply whether the vehicle is currently moving. Information from the GPS 118 can also be used to calculate vehicle headings, and such vehicle headings can be re-approximate or compared to those calculated using the magnetic sensor circuit 102. It can also be determined whether recalibration is necessary. This use of GPS information by the compass circuit is commonly used by Robert R. Turnval, entitled “REARVIEW MIRROR WITH INTEGRATED MICROWAVE RECEIVER,” filed on June 28, 2000, the entire disclosure of which is incorporated herein by reference. Assigned US Patent No. 6,407,712 and US Patent Application Publication No. 2002 / 0032510A1 by Robert R. Turnval et al. Entitled "VEHICLE REARVIEW MIRROR ASSEMBLY INCORPORATING COMMUNICATION SYSTEM" filed April 5, 2001 Are disclosed in detail. Robert R entitled "AUTOMOTIVE MIRROR WITH INTEGRATED LORAN COMPONENTS" filed on June 15, 2001, a system in which a LORAN receiver and / or antenna is incorporated into a rearview mirror assembly, the entire disclosure of which is incorporated herein by reference. U.S. Pat. No. 6,539,306, commonly assigned to Turnbal et al.

  As described above, the processing circuit 110 can be coupled to the vehicle bus via the vehicle bus interface 120. Information provided to the vehicle bus that is useful to the processing circuit 110 includes vehicle speed, travel distance, whether the vehicle is stationary or moving, and whether the vehicle steering wheel is being rotated. Other useful information can be provided by inclinometers or other forms of roll sensors provided on the vehicle.

  As shown in FIG. 3, the compass circuit 100 can further include a power supply circuit 124 coupled to a power supply line from either the vehicle battery or the ignition. This power supply 124 converts the power supplied by the vehicle into a voltage useful for the various electronic components of the compass circuit. If the voltage from either the vehicle battery or the ignition exceeds 12 volts, a power circuit as disclosed in commonly assigned US Pat. No. 6,262,831, the entire disclosure of which is incorporated herein by reference. Can be used.

  As will be described below, all or part of the electronic compass circuit 100 is mounted on or in the vehicle rearview mirror assembly. However, all or part of the compass circuit 100 may be mounted anywhere in the vehicle, such as an overhead console, a console mounted on the front windshield, a vehicle A-pillar console, a vehicle instrumentation panel, or others. It may be attached to the position. For example, the sensors 104, 106 and optionally 108 may be mounted on or in the rearview mirror assembly, while the other part of the compass system is somewhere on the vehicle, such as an overhead console or instrument panel May be arranged. The sensor output can be used in other systems of the vehicle, such as navigation systems, headlamp control systems and / or telematic systems.

  FIG. 4 is a front view of a rearview mirror assembly 140 incorporating a heading indicating device 114 in the form of at least an alphanumeric display. FIG. 5 is a side view of such a rearview mirror assembly, partially broken away to show attachment of sensors 104, 106 and optional sensor 108 to circuit board 150 in mirror housing 144 of assembly 140. FIG. . FIG. 6 is a side view of the rearview mirror assembly 140 showing the sensors 104, 106, and 108 disposed within the mounting foot 146 of the mounting structure 145 of the rearview mirror assembly 140. FIG. 7 is a top view of the mirror assembly 140 shown in FIGS. As shown in each of FIGS. 4-7, the rearview mirror assembly 140 includes a mirror housing 144 pivotally coupled to a mounting structure 145, which is then mounted inside the windshield of the vehicle. Or it can be attached to a roof structure or headliner that extends along the top of the windshield of the vehicle. The housing 144 includes a bezel 142 disposed around the front of the housing 144 to hold the mirror element 128 in position.

  The mirror element 128 is preferably an electrochromic mirror element that changes its reflectance in response to an applied voltage signal. As shown in FIG. 3, the processing circuit 110 of the compass circuit 100 is coupled to an electrochromic (EC) drive circuit 126 which in turn drives the electrochromic mirror element 128 and the external electrochromic mirror element to drive voltage. give. The processing circuit 110 can then be programmed to respond to the output signal of the ambient light / glare sensor 130. More particularly, the glare sensor is arranged to sense light from the rear of the vehicle, while the ambient light sensor is generally opposite the mirror housing 144 to sense the ambient light level in front of the vehicle. Placed in. By responding to the light level sensed by the sensor 130, the processing circuit 110 then controls the EC drive circuit 126 and thus the reflectivity of the electrochromic mirror element 128 and the external electrochromic mirror element. Can do. Suitable EC drive circuits are well known in the art, an example of which is disclosed in commonly assigned US Pat. No. 6,247,819, the entire disclosure of which is hereby incorporated by reference. The rearview mirror assembly 140 may include a suitable user input switch 116 to allow the user to activate or deactivate the automatic control of the reflectivity of the electrochromic mirror. The instruction LED 132 and the like are provided for instructing whether or not the electrochromic mirror is in an automatic state.

  User input switch 116 may take any form conventionally used for rearview mirror assemblies or other accessories in a vehicle. Suitable structures for user input switches are incorporated herein by reference in their entirety, commonly assigned US Pat. Nos. 6,407,468 and 6,420, both entitled “REARVIEW MIRROR WITH BUTTONS INCORPORATING DISPLAY”. , 800. Another suitable structure is disclosed in commonly assigned US Pat. No. 6,471,362, entitled “MIRROR WITH IMPROVED BUTTON CONSTRUCTION”, the entire disclosure of which is incorporated herein by reference.

  As shown in FIG. 4, the heading indication display 114 can be provided behind the mirror element 128, and a transparent window is formed on the reflecting surface of the mirror 128 so that the heading can be seen through the mirror. However, the mirror 128 may be configured to incorporate any of the features disclosed in the commonly assigned US Pat. No. 6,356,376 described above. This patent discloses various mirror structures in which a display placed behind a mirror can be viewed without completely eliminating the reflectivity in front of the display. A suitable display 114 or other heading indicating device may be provided at or near the bezel 114, at or near the mounting structure 145, or near or away from the rearview mirror assembly 140. It will be further apparent that accessories may be provided on the instrument panel, overhead console, etc. Additional information displays, such as the passenger side inflatable restriction indication display disclosed in US Pat. No. 6,170,956, the entire disclosure of which is incorporated herein by reference, may be incorporated into the mirror assembly.

  As shown in FIG. 5, the magnetic sensor elements 104, 106 and 108 can be attached to a circuit board 150 provided in the mirror housing 144. The circuit board 150 can be provided vertically, horizontally, or in other directions. These sensors include an X-axis sensor 104 that aligns its axis generally parallel to the vehicle's longitudinal axis (ie, the direction of travel), and a Y-axis sensor 106 that aligns its axis generally horizontally and transversely to the vehicle and Generally oriented to align perpendicular to the axis of the X-axis sensor 106. Thus, both X-axis and Y-axis sensors 104 and 106 are aligned such that their axes are in the horizontal plane. The Z-axis sensor 108, if provided, can be mounted so that its axis is generally vertical. The axis of the Z-axis sensor 108 is preferably perpendicular to the axes of the sensors 104 and 106. As will be described in detail below, these sensors can be attached to circuit board 150 in any direction relative to the circuit board and vehicle by taking into account changes in their direction relative to the circuit board and / or vehicle. Also, these sensors do not necessarily have to be mounted so that their axes are perpendicular to each other, but this is the case when such a configuration is taken into account by coordinate transformation or the like.

  When the sensor is mounted on the mirror housing 144, particularly if the Z-axis sensor 108 is not used, the mirror housing 144 and thus tilted into the mirror housing 144 to sense when the sensor element is tilted by hand. A sensor (not shown) is preferably provided. Examples of suitable tilt sensors are disclosed in commonly assigned US Pat. Nos. 6,023,229 and 6,140,933, the entire disclosures of which are incorporated herein by reference. However, as described below, when the Z-axis sensor 108 is provided, the tilt sensor is not necessary.

  FIG. 6 shows another structure in which the sensors 104, 106 and 108 are mounted within the mounting foot 146 of the mounting structure 145. The effect of attaching the sensors 104, 106 and 108 to the mounting foot 146 is that they are maintained in a fixed relationship to the vehicle.

  Another suitable mounting embodiment for a magnetic sensor is disclosed in commonly assigned US Pat. No. 6,587,573 entitled “SYSTEM FOR CONTROLLING EXTERIOR VEHICLE LIGHTS”, the entire disclosure of which is incorporated herein by reference. ing. In this patent, the compass sensor is disclosed as being attached to a support structure that is adjustably disposed within a housing attached to a mirror mounting foot. The support structure disclosed in this patent is also used to support a camera system for use in a headlamp control system and / or a rain sensing system. A support member is adjustably disposed within the housing to ensure that the camera is mounted generally horizontally with respect to the vehicle. Because the angle of the windshield of different vehicles can vary significantly, the angle of the camera and compass sensor within the fixed housing can be adjusted, eliminating the need to redesign each of these mirror mounting structures for different vehicles, This structure can be used on a variety of vehicle model platforms. The effect of mounting the compass sensor in this way not only ensures that the X and Y sensors are mounted horizontally to the vehicle, but also uses, for example, images obtained from the camera when the vehicle is positioned in front of the target. And to ensure proper orientation of the compass sensor after the vehicle is attached to the windshield.

  24A-24D illustrate a method of attaching the compass sensors 104, 106, and 108 to the circuit board 150. FIG. More specifically, the Y-axis and Z-axis sensors 106 and 108 are mounted parallel to the surface of the circuit board 150 and perpendicular to each other. The X-axis sensor is first attached in parallel to the Y-axis sensor 106 at the portion 150 a of the circuit board 150. Portion 150a is perforated or otherwise separated from the remainder of the circuit board by a notched slot except for two wires or traces that electrically couple sensor 104 to the remainder of circuit board 150. As shown in FIG. 24C, the portion 150 a is bent from the rest of the circuit board 150 so that it and the sensor 104 are vertically aligned with the surface of the circuit board 150. Thus, the sensor 104 can be mounted using a conventional mounting process and is mounted so that its axis is perpendicular to the axes of both of the other two sensors 106 and 108. The sensor 104 and upright circuit board portion 150a can then be secured in this position using any one of a variety of techniques. According to one such technique, a retainer 700 having a plurality of resilient legs 702 is slid onto and around the upstanding portion 150a and the sensor 104, and the resilient legs are snapped into their corresponding holes 150b. be able to.

  Although the above method has been described for mounting three sensors on a circuit board, the same technique is used when only two sensors are mounted (ie, when only sensors 104 and 106 are used). can do.

  FIGS. 25-27 illustrate another method for attaching a compass sensor to a circuit board, such as the circuit board of the mirror assembly 140. In this way, the compass sensors 104, 106 and 108 are integrated into a common integrated sensor package 720 such that their axes are perpendicular to each other. By integrating the sensors into a common package 720, the integrated package 720 can then be attached to the circuit board 150 using a conventional circuit population machine. Further, the package may have one or more alignment protrusions and / or grooves to ensure a consistent mounting direction with respect to the circuit board. In the embodiment shown in FIGS. 25-27, the integrated sensor package 720 includes sensors 106 and 108 such that the axes of the sensors 106 and 108 are parallel to the circuit board 150 when the package 720 is attached to the circuit board. The three-dimensional shape is oriented in the package 720. As described above, the sensors 104 to 108 are preferably of a magnetic induction type, but may be of a magnetic impedance type, a magnetoresistive type, or the like. Package 720 is shown as having six leads (two for each sensor), but may be composed of only four leads (one for each sensor and a common terminal for all sensors). .

  As shown in FIG. 27, when three sensors are attached to a three-dimensional package with their axes perpendicular to each other, a significant percentage of the package volume is not required for sensor integration. This portion is shown as region 722 of sensor package 720. Accordingly, as described below, such a region 722 may be eliminated to reduce the volume in the vehicle accessory (ie, mirror housing) required to accommodate the compass sensor. An example of such a cut solid integrated compass package will be described below with reference to FIGS.

  In addition to removing excess package volume, aspects of the present invention free the designer from the constraint of mounting the sensor so that the sensor axis is parallel or perpendicular to the circuit board to which the sensor is connected. That is. If the sensors are arranged so that their axes are perpendicular to each other, the sensors can be mounted in any direction with respect to the circuit board. As will be described in detail below, the directional dispersion is modified so that the X-axis sensor is aligned perpendicular to the circuit board (and / or parallel to the direction of travel of the vehicle) and the Y-axis sensor is in the horizontal plane. The coordinate transformation can be performed in the compass processor code so that it is aligned parallel to (and / or horizontally and vertically in the direction of travel of the vehicle). Therefore, when the two concepts are combined, the sensor can be attached in the same manner as shown in FIGS. 28 and 29, that is, the solid of the conventional embodiment is cut to remove the excess region 722 and to pull it. Repeatedly, an integrated sensor package 750 resembling a pyramid shape can be obtained. Such a package structure is effective in that the volume is small enough to be able to extend between the support ribs of the mirror housing without requiring modification.

但し、α＝４５°、β＝−３５．２６４３９°、γ＝０°
従って、次のようになる。

To compensate for sensor direction shifts, a coordinate transformation to processing circuit 110 code can be programmed. Such a transformation can be performed using Euler's rotation function. This function consists of (a) rotating the original frame S around the x axis by an angle α to the reference frame S ′, (b) rotating the reference frame S ′ around the y ′ axis by an angle β. And (c) rotating the reference frame S ″ about the z ″ axis by an angle γ to the reference coordinate frame S ″ ′. Such a transformation can be represented by the following linear equation.

However, α = 45 °, β = −35.26439 °, γ = 0 °
Therefore:

  This same coordinate transformation can be used to compensate for compass sensor tilt and rotation when the compass sensor is in the mirror housing. For example, the average tilt and rotation of the mirror housing can be factored into α, β and / or γ in the above transformation. Alternatively, the calculation of the transformation is further simplified by aligning the sensors in the integrated sensor package so that at least one of the sensors is more precisely aligned in the desired direction with respect to vehicle travel. be able to.

  One type of sensor package that can be used is a magneto-impedance sensor package developed by Aichi Steel and Aichi MI. This package includes two magneto-impedance (MI) type sensors formed in a direction perpendicular to each other in an integrated circuit. Another Aichi MI sensor integrated circuit that combines three MI sensors is also available. The same sensor packaging technique can be used to combine the sensor with other components of the compass sensor circuit including the microprocessor of the processing circuit 110. Such MI technology differs from magnetic induction technology in that the MI sensor forms an electrical connection to the core of the sensing inductor. Furthermore, the size of the sensor can be very small compared to existing sensors and therefore can be embedded in a small SMT package with the necessary signal processing integrated circuits. Due to its small size, the MI sensor cannot be placed in the mirror housing by itself, but can be placed in a mounting arm / tube or attached to a channel mount. The MI sensor may be incorporated in a multi-sensor assembly including a sensor such as a photo sensor. The core material of the MI sensor may be amorphous or nanocrystalline, cobalt or iron-based. Various core shapes such as strips, ribbons or wires can be used. Wire cores tend to have a higher effective permeability than strips or ribbons, resulting in better sensitivity.

  FIG. 30 illustrates one embodiment of a circuit board assembly 800 and associated compass circuit according to the present invention. As shown, the circuit board assembly 800 includes a main (ie, mother) board 802 and an optional daughter board 804, which is any one of a number of different types of displays 114 and associated display drivers. Circuit (not shown). The processing circuit 110 attached to the motherboard 802 is preferably programmed to use display drive signals suitable for the particular type of display 114 on the daughter board 804 (if any). This can be accomplished by reprogramming the processing circuit 110 with code suitable to drive the particular type of display being used. However, the processing circuit 110 is preferably programmed with code to drive all types of displays that can be used, the appropriate code being selected manually or automatically. Processing circuit 110 can automatically select the appropriate drive code to use to drive any one of more than one type of display (eg, vacuum fluorescent, graphic display, LCD and LED display as described above). To do so, a plug and play function can be added to the display drive circuit of the daughter card 804. Alternatively, some of the connector pins of connectors 806a and 806b that connect daughter board 804 to motherboard 802 have a unique combination of connections to ground or power and processing circuit 110 is connected to that pin connection. Based on this, the type of display may be determined, and thus the daughter card, if any, may be able to select the appropriate code used to drive the particular display on it.

  Also, as shown in FIG. 30, an interface 808 is attached to the mother circuit board 802 and is electrically coupled to the processing circuit 110. Alternatively, interface 808 can be attached to a daughter board such as 804 (with or without display 114). The interface 808 may be an RS485 interface or other serial interface for connecting to a display or other electrical component to transmit vehicle headings. Such a component or display may be provided in place of display 114 and its daughter board 804, or may be provided in addition to such a display. This component or display may be located somewhere in the mirror assembly or vehicle accessory in which the motherboard 802 is housed, or may be located remotely therefrom. Thus, interface 808 may be a LIN or CAN interface to a vehicle bus or a local area bus. The processing circuit 110 is preferably programmed with code to properly format the vehicle heading signal for transmission to all forms of displays and components via all possible forms of interface. Thus, the processing circuit 110 can identify the type of interface and the type of connected component (s) or display (s) and transmit the vehicle heading signal in an appropriate format. .

  If it is desired to attach a GPS antenna to the rearview mirror assembly 140, the antenna is commonly assigned US Pat. No. 6, entitled “MICROWAVE ANTENNA FOR USE IN A VEHICLE”, the entire disclosure of which is incorporated herein by reference. Can be attached based on the teachings of 396,446. Another suitable antenna is disclosed in commonly assigned US Patent Application Publication No. 2002 / 0032510A1, entitled “VEHICLE REARVIEW MIRROR ASSEMBLY INCORPORATING COMMUNICATION SYSTEM”, the entire disclosure of which is incorporated herein by reference.

  Having described the hardware of the electronic compass system, a general overview of the process of the present invention to be executed by the processing circuit 110 is described below. Following the general overview, a first embodiment illustrating one implementation of several concepts of the present invention will be described in detail with reference to FIGS. 11-21.

  As described above, in the preferred embodiment, the magnetic sensor circuit 102 has a Z axis arranged substantially perpendicular to sense a magnetic field component perpendicular to the components sensed by the X and Y axis sensors 104 and 106. A sensing element 108 is provided. Thus, the process of the present invention preferably maps data to a three-dimensional coordinate system. Ideally, the mapped data corresponds to a sphere rather than a circle in a single fixed plane.

To better illustrate the principles of the present invention, FIG. 8 illustrates a compass magnetic sensor coordinate system that has been modified to take into account the true N pole offset from the magnetic N pole and the sensor direction shift relative to the vehicle. explain. More particularly, a portion of the “world sphere” is shown with coordinate planes labeled X _W , Y _W , and Z _W , where the X _W coordinate axis is in the direction of travel of the vehicle. Aligned and generally horizontal (ie tangent to the ground), the Z _W coordinate axis is vertical (ie pointing to the center of the earth), and the Y _W coordinate axis is horizontal and in the direction of travel of the vehicle Arranged vertically and horizontally.

The true north pole (ie, the north pole of the earth's axis of rotation) does not actually correspond to the magnetic north pole, so the magnetic compass is the actual position relative to the particular location on the ground where the compass is located. The deflection angle D corresponding to the angular offset of the magnetic north pole from the north pole must be taken into account. Mappings of declination D for various geographic zones on the ground surface are readily known and available. In order to take account of the declination, the second coordinate plane is shown in broken lines and is denoted as X _V , Y _V , Z _V , where the X _V and Y _V axes are the X _W and Y _W axes. To the Z _W and Z _V axes are simply rotated equally by an angle D.

Assuming ideal conditions, the X-axis sensor is aligned with the direction of travel of the vehicle, the Z-axis sensor is aligned perfectly vertically, and the Y-axis sensor is aligned vertically with both the X-axis and Z-axis sensors. Assuming that the compass sensor is permanently attached to the vehicle in a closed state, the angle φ that the sensed geomagnetic field vector forms with respect to the X and Y axis sensors to derive the true heading of the vehicle It is only necessary to subtract the declination D from. However, when the sensor is mounted on a housing that can move relative to the vehicle, such as the rearview mirror housing 144 (FIGS. 5 and 7), and the mirror housing is rotated about the vertical Z axis to an angle φ (FIG. 7). The X _m and Y _m coordinate axes of the sensor coordinate system rotate by the same angle φ _m relative to the X _V and Y _V coordinate axes, while the Z _m axis remains coincident with the Z _W axis. If the mirror rotation angle φ _m is assumed or can be identified, it can be subtracted from the heading angle along with the deflection angle D to give an accurate heading.

When the mirror housing 144 is tilted about its horizontal horizontal axis (ie, its Y _m axis) in order to accept drivers of different heights, the calculations are significantly more complicated (see FIG. 5). Such inclination of about the Y _m axis is maintained the Y _r axis thereby resulting in a state consistent with the Y _m axis, the X _r and Z _r axis from only X _m and Z _m axis angle theta _m Rotate.

As described above, vehicle magnetic fields, as well as external magnetic fields other than the geomagnetic field, affect the magnetic field sensed by the sensor. If there is no such effect and the sensor is not tilted with respect to the horizontal (ie, θ _m = 0 °), a plot of the data obtained from the sensor will result in the surface of a sphere having the same radius and center point as the circle Across the circle, its center point corresponds to the origin of the coordinate plane, its radius corresponds to the magnitude of the geomagnetic field vector, and the circle is in a plane parallel to the horizontal XY coordinate plane of the sensor Will be present (ie, the z-component value is kept at least locally constant). However, when the mirror is tilted, the circle kept horizontal is tilted with respect to the XY plane by the tilt angle θ _m , and therefore the z component value changes as a function of the direction of the vehicle. In the case of plotting considering only the X and Y sensor outputs, the data on the XY plane looks like an ellipse. The mirror tilt can be taken into account by factoring the assumed mirror tilt angle θ _m based on the average tilt angle of the average individual, or the vehicle is driven through a 360 ° turn angle. It is also possible to measure by monitoring the dispersion of data obtained from the Z-axis sensor.

The effects of the vehicle's magnetism and external magnetic fields typically dramatically shift the center of the local sphere (sc _xs , sc _ys , sc _zs ) from the origin of the sensor coordinate system (x ₀ , y ₀ , z ₀ ). Since the origin has zero values for x, y and z, the compensation for this shift is to use a vector sc extending from the origin to the center of the world sphere, and the corresponding magnetic sensor data points (rdg _xs , rdg _ys , This can be done by subtracting from each vector RDG extending to rdg _zs ) (see also FIGS. 9 and 10). This compensation is done by subtracting the x, y and z values (sc _xs , sc _ys , sc _zs ) of the local point of the local sphere for each data point (rdg _xs , rdg _ys , rdg _zs ) obtained from the sensor. Achieved. The resulting vector V extends from the center of the local sphere to the data point. As shown in both FIGS. 8 and 9, the local sphere has a radius vector RS (with magnitude rs _xsr ) and a center point (sc _xs , sc _ys , sc _zs ), which is assumed first. Then, it is recalculated through the approximate sequence described below. To aid the approximation process, the error vector VE for the point (rdg _xs , rdg _ys , rdg _zs ) can be determined by subtracting the current radius RS from the vector V. Such an error vector VE is calculated for each data point obtained from the sensor and then averaged or otherwise approximated data (ie, center points (sc _xs , sc _ys , sc _zs )). And radius rs _xsr ). As described in detail below, the center point and radius may be changed later to obtain a “best fit” solution for data obtained from the sensor with the lowest average error. The center point and radius that gives the best fit can be stored for later use in ascertaining the heading angle θ _S from the latest sensor reading and thus the vehicle heading to be displayed.

  Commonly assigned US Pat. Nos. 6,023,229 and 6,140,933, which are hereby incorporated by reference in their entirety, include, among other things, when the heading angle was corrected and when the mirror housing 144 was moved. The use of a ball position (or tilt and / or rotation) sensor that can be used to determine and trigger reapproximation or recalibration is described.

In many vehicles, the driver's position is sufficiently consistent, and the change in mirror angle adjustment around the vertical axis, φ _m, is usually relatively small from one driver to the next. Thus, corrections to this change are less important than compensating for changes in the action of this field when the mirror position and / or angular orientation changes relative to the vehicle's disturbance field. This problem is most noticeable for a two-ball mirror mount that performs substantial position adjustment in addition to changes in angular direction. The mirror position sensor is also more complicated than the two-ball mirror mount. In addition, many of the mirrors are used in parts of the world where the vertical component of the geomagnetic field is substantially greater than the horizontal component, thus increasing the tilt θ _m of the mirror housing about the horizontal axis Y _{m in} some cases. Then, the compass reading has an effect several times larger than the rotation φ _m of the same size around the vertical axis Z _m .

  In the case of a driver using a compass attached to the mirror housing, a common scenario is that two or more drivers adjusting the mirror to widely different positions share the driving and travel together. In such cases, travel generally may last for a long time in one direction, so the approximation is evaluated and updated continuously or quasi-continuously, and substantially less than 360 °, and even substantially 180. A mirror that performs effective approximate adjustment based on a range in the running direction that is less than or 90 degrees is desirable. It is desirable to apply criteria for determining when the best-fit approximation data is satisfactory for displaying heading information. It is also desirable for this determination to place more emphasis on the quality at which the approximate geometry used fits the sample data rather than, for example, the minimum threshold for the angle exerted by the accumulated turn. For example, the quality of the fit can be measured by criteria related to least squares. Then, when fitting the approximate geometry, when the mean squared error for each sample data point is generally low and within an acceptable range, and preferably it converges to the dominant minimum Sometimes the approximate geometric shape can be used as a basis for calculating and displaying heading information. This is true when the total change in direction exerted by the sample points is small, perhaps well below 90 °, and the acceptance criterion for the mean square error is a variable based on a set threshold or other factor It's okay. This is an example, many other measures of quality of fit may be used, and an acceptance criterion may be used for other or additional uses of compass data.

  In the scenario above, when the driver leaves the highway and takes the driver's turn and re-enters highway traffic, some maximum direction changes can occur. In order to take advantage of the rapid and often relatively large changes in direction of travel that often occur when the driver starts a segment of the itinerary, a system that captures and sorts the readings quickly and reliably is desired. As will be apparent from the description below, the compass system of the present invention exhibits these desirable features.

  The raw data points read by the sensor may be filtered and / or averaged with multiple readings. Other data such as variance and standard deviation may be included with the reading as a stability indicator. Any data transformation may be applied that includes one or more of the transformations described below. Applying a magnification factor that can be determined and stored as part of the factory calibration to approximate, calibrate, or simply equalize the gain for two or more directional components of the magnetometer reading Good. Data points can optionally be evaluated based on statistical data compiled with readings, and can be evaluated based on how well they fit the current approximate geometry. Optionally, it is also optional to evaluate how the vertical Z-axis component of the reading matches that of the current sample point and possibly additional criteria. If a data point is eligible under these assessments and the approximate geometry is eligible to be used as a basis for processing magnetic data for reading, a heading instruction based on the acquired data point is Generated and optionally displayed. The processed magnetometer data may also be used for other purposes.

  Points are evaluated for logging into the active sample set or for immediate inclusion therein, and currently stored active and inactive points are considered for possible changes or changes in state sets and in the sample set Is evaluated against. The sample set is then assembled and maintained. The development of a good sample set generally improves the accuracy of the compass reading and, in particular, provides a relatively quick response to the shifting of approximate requirements resulting from mirror adjustment when the magnetic sensor is housed in a movable mirror housing. Work as a basis for.

  Sample point analysis is then performed to provide the data necessary to extract heading information from magnetometer readings. A preferred way of doing this is, for example, selecting and possibly refining a pattern that is a kind of fingerprint characteristic of the response pattern of the magnetometer reading to the geomagnetic field due to the changing vehicle direction, and this pattern is changed by the vehicle Statistically fit the set of magnetometer readings recorded on the vehicle when taking a direction. The data from this pattern matching process is then used to extract components of subsequent readings primarily due to geomagnetic fields, particularly from disturbance effects from other magnetic disturbance sources in the vehicle. The extracted readings and / or data obtained from this extraction process is then used to calculate the heading direction. Although not required for each embodiment, the fitting algorithm can generate a number that indicates the quality of the fit, which is to find a pattern change and / or direction that gives an approximate best fit. It is preferably used in a repetitive sequence. The quality of fit indicated by this number is preferably also used as one of the variables on which other judgments are based. For example, the quality of the fit may be used as a partial basis for determining when the approximation is satisfactory to use as a basis for processing and outputting the heading indication. If different starting conditions for the pattern result in convergence to different local minimums in the fitting process, the quality of the fit can be used to determine in part which one should be selected as the best fit. it can. In this case, if one of the local minimums is not a clear best fit, the heading information is calculated and the use of the fit as a basis for refining the sample set is prohibited, eliminating the possibility of judgment. May be. Preferably, the components of the calculation process used for the adaptation algorithm can be serviced for other uses.

  For example, in a preferred embodiment, the distance from the approximate pattern to the reading is calculated in vector form as the error vector VE described above for each applicable reading. This square of distance, i.e. the square of the magnitude of the error vector, is used in the fit calculation, and the components of the error vector effect the next position of the pattern or changes to it in an iterative process to determine the approximate best fit. Used for algorithmic prediction. Furthermore, the calculations used to determine the error vector are used to directly convert the error vector to the desired basic coordinate system. Also, the angle calculated in determining the distance from the pattern to the point is directly useful in determining the heading angle for a given reading and probably determines which readings are adjacent to each other and those It is also useful for approximately measuring the interval. The distance from the approximate best fit pattern to the point adds or rejects the point in the sample set used to determine the best fit pattern, or accepts or rejects the point for use in determining the current heading. May be used as a partial basis for In a preferred configuration, a substantial portion of the same calculation is used to sequentially analyze each point of the sample set as part of the calculation to assess the quality of the fit and determine the next fit approximation in an iterative sequence. And apply the current best fit pattern to the processing of individual readings to make a partial assessment of their quality and determine the associated headings.

  As described above, the approximate geometry is arbitrary, but may be adjusted continuously or periodically to better fit the sample points. For example, when the approximate pattern or geometric shape is a sphere, the radius of the sphere can be adjusted. The radius of the sphere is usually related to the strength of the geomagnetic field or its horizontal component, based on the approximate form used. In any case, this parameter must be stable for a given locality even when disturbances such as changes in the magnetization pattern of the vehicle or adjustment of the mirror to which the sensor is attached occur. In particular, when the sensor is attached to a mirror, it is desirable to adjust the approximation quickly and thus with a minimum amount of data after the mirror position change. For such changes, the best fit radius is unlikely to change significantly, if any, and the sample points are usually well distributed around the approximate pattern to better determine the radius. Need. It is advisable to prefer to keep the previous best-fit radius for the approximate geometry until the sample point distribution and quality are favorable in order to provide data to adjust this value appropriately. By using the previously determined radius, a more limited sample set can be used effectively to respond quickly and fully to the necessary changes in the compass approximation. It is preferable to extend this solution to other parameters, i.e. based on the relative stability of the parameters in the face of approximate changes or in the relatively short changes expected in the vehicle environment. It is preferable to establish a general hierarchy of parameters that can be adjusted to obtain the best fit. The parameter values that are most stable to these changes are preferably retained until data is available that will likely result in improved settings.

  In some embodiments, rather than simply adjusting an existing pattern, switch to a different approximate reference pattern, for example, switch between a circular and elliptical reference pattern, or perhaps a two-dimensional and three-dimensional pattern There may be cases where this is desirable.

  The approximate pattern is then compared to the active sample points and an indication of the quality of the fit is given. An analysis such as a numerical indication of the quality of the fit and a least squares fit is desired. Thereafter, a change to the shape and / or position of the approximate pattern that produces a good fit can be predicted. These approximations go from simply stepping through an array of possible choices to calculating the absolute value or change increment that should be applied to determine the pattern to use for the next best fit approximation. It is a range. In addition, to limit program size and computation time, it is preferable to utilize a predictive process that shares the intermediate numerical results with other necessary calculations such as those used to measure the quality of the fit. This prediction can then be applied repeatedly to proceed to an approximate best fit.

  When starting for the first time or after a major shift in the approximation, the sample set can be constrained to a very small portion of the approximate geometry. In such cases, it is advantageous to add additional constraints to obtain a reasonably accurate approximation with a limited sample set. For example, when an approximate sphere is used whose radius is approximately equal to the horizontal component of the geomagnetic field, the center of the sphere is usually approximately the center of the circle traversed by the magnetic field vector as the vehicle rolls in a complete circle. I have to enter. In such a case, the center of the approximate sphere is approximately in the horizontal plane of the sample point circle. This is also almost the horizontal plane defined by the vertical component of the centroid of the sample point. Thus, if the sample points are not well distributed enough to define this plane, it is effective, for example, to limit the center of the approximate sphere to the horizontal plane that contains the centroid of the sample points (equal weights for each point). May be specified, or different weights may be specified).

  In particular, another problem can arise when sample points within a sample set span a limited portion of the full range and when the sample set does not include at least three points that are reasonably separated from each other. In these cases, it is difficult to distinguish the correct center of the sphere from the center point, which is generally a mirror image of the first one for the points in the sample set. Another starting center location that causes problems with convergence to the appropriate center point is typically on the perpendicular bisector connecting the two center points described above. In this case, for example, in the predictions mentioned elsewhere in this document, convergence first places the point's center of gravity generally near the approximate sphere, and then points to either the correct center or the wrong mirror image center. Converge slowly. A preferred way to deal with this problem is to try a number of start center points that typically fall in a circle around the centroid of the sample points, for example, regularly or when there is any doubt about the quality of convergence. For each start point, an iteration must be performed to achieve an approximate best fit given that start point. Then, the centers corresponding to the approximate best fit at each start position must be compared, and if they are not about the same point, the quality of the fit for each must be compared. If the quality of fit for one is substantially better than the other, the corresponding center must be chosen. Otherwise, the sample set must be refined until the best quality best fit is obtained for a single approximate best fit center point.

  As part of manufacturing calibration, the relative gain of the magnetic sensor is preferably measured for each sensing direction. Calibration constants are then stored and applied to each subsequent reading obtained by the magnetometer to calibrate and at least equalize the relative gain of the magnetometer for each of its sensing axes. Is preferred. When there are conditions that give an essentially elliptical response pattern, it is preferable to prescale the data so that a circle or sphere fitting process can then be used. In particular, when the sensor is mounted on a mirror, it is usually convenient to have the nominal vertical axis parallel to the approximate vertical plane of the circuit board. However, in a vehicle, the circuit board in the mirror is usually not in a vertical position with respect to the intermediate driver. Assuming that one sensor, for example the Y-axis sensor, is parallel to the circuit board and in a horizontal plane, the coordinates are rotated around this axis and the X and Y axes are common for intermediate drivers. It is preferable to establish a coordinate system that is in a horizontal plane and whose transformed Z axis is nominally perpendicular to the middle driver.

  For the first preferred embodiment, a sphere is selected as the approximate pattern. In a preferred configuration, the approximate radius is set to a value such that the circle generally traced by the magnetic reading as the vehicle rolls in a circle approximately forms the equator of this sphere at the best fit position. It is preferable. It is preferable to adjust this radius only when a set of sample points with a relatively high quality and a relatively good distribution is obtained. Then, for example, iterating through a set of trial radii, it is preferable to find the radius where the fit is good and the sample point falls near the equator of the approximate sphere. Except as described below, once a reasonable center point is established, it is preferable to use the previously established center point as a starting point for a new iteration. It is also preferable to use the average of the error vector sums for each sample point as a vector to add to the previous center point to establish a new center point for the next step in the iterative process. This iterative sequence is preferably continued until the center change for a number of successive iterations falls below the threshold value.

  When doubt about the quality of the fit is indicated by a relatively high error indication for the approximation, additional iterations similar to or approximately equal to those described below must be applied here as well.

  In some embodiments, it is preferable to not only calculate an indication of how well the points fit the approximate geometric data pattern, but also store it with each point in the same set of points. This may be a quality indication or an error indication such as the square of the distance from the approximate pattern. It is also preferred to save an indication of the time or at least the relative order in which the sample points were acquired in the sample point set. This may be, for example, a sequence number, date, or mileage. The order of sample point acquisition is maintained by storing sample points in memory on a first-in first-out (FIFO) basis.

  For a given sample point set, a quality / error indication and accompanying description may be stored with each point of the sample point set. The points can then be ordered based on the angle φ calculated for each point. Then, pairs of points that are adjacent to each other in the ordered list, including the pair consisting of the first and entry of the ordered list, can be considered as adjacent points. The magnitude of the difference in angle φ modulo 360 ° relative to the angle associated with each point of this pair can be used as an indication of the spacing between the two members of each pair of adjacent points.

  Any calculation similar to that used to find the center of gravity is useful for determining the center of the distribution of sample points. For each sample point, for example, one hypothetical mass can be specified, and the x component at the center of the mass is made equal to the average of the x components of the points in the set. Similarly, the y and z components are each made equal to the average of the y-components of the member points and the average of the z-components of the member points.

  There are many possible techniques for applying the selection criteria. In general, given a new point that is a candidate to replace one of the existing active sample points, first a decision must be made as to whether the replacement should be performed, and second, the replacement If so, you must determine which points in the existing active sample set should be replaced. One way to apply the criteria is to replace each point of the existing active sample set with a candidate point, for example, temporarily, one after the other, and one at a time. In the case of each temporary replacement of a single sample point with a candidate point, calculations are applied as necessary and the information necessary to compare the results to apply the replacement criteria is logged. The logged results are then compared with each other and similar data for an existing data set to see if any replacement satisfies the join criteria, and if so, to best satisfy the criteria Select which one of the members should be replaced. Then, if there are no satisfactory replacements, the sample set is not exchanged and the candidate points are not used as active members of the sample set. Otherwise, the replacement that produces the most favorable result is retained and the replaced point is discarded or designated as an inactive set.

  The angle for the largest interval and the next largest interval is then logged as each replacement is made, and the results for each replacement are the results for each other and for the largest and next largest intervals in the original sample set. Compare and determine the suitability of any of the possible replacements, and rank options for suitability under this criterion.

  The centroids for the original sample set and the sample set with each replacement are compared and ranked, similar to those used for the interval evaluation described above.

  It is preferred to eventually purge older points and know at least the relative acquisition time when a change is detected in the pattern. This information is useful when it is necessary to separate data that conforms to the new pattern from data that is not. Here, the separation or separation of the set of points may be performed based in part on the order in which they were acquired. Also, whether the data that fits well with the pattern and that the statistical data related to the acquisition indicates that the stable reading has been performed should be given a priority and whether the suitability is poor. Alternatively, it is preferable to reject data acquired under noisy conditions.

  The vertical component of the geomagnetic field and the additional effect on the measured vertical magnetic field component by the vertical component of the geomagnetic field in the vehicle must be relatively stable even when turning in a horizontal plane. Substantial changes in the vertical component of the reading are likely to occur either due to transient conditions or changes that cause a more permanent shift to be approximated. If the change is transient, the measurement point is not good to use as a sample point or to calculate the heading. If the shift lasts for a long time or is semi-permanent, a relatively quick response is needed to properly re-apply the compass. In any case, comparing the perpendicular component of the magnetization intensity for a particular point with the dominant average of the perpendicular component, or perhaps comparing the mean of the perpendicular component for a particular group of points, the particular reading will probably fit the current pattern Useful in determining when not to. The determination of whether such changes are transient or relatively permanent is preferably made in part by accumulating the reading history. If there is a relatively stable shift in the vertical component of the reading, the approximation will probably also shift. Here, it is useful to store a representative reading after the shift is first detected so that the reading can be evaluated for inclusion in a new set of approximate samples, and the approximate shift is It is advantageous to purge or acquire points before they become active. If the vertical component of the reading generally returns to its previous value or dominant value after a short distance or run time, it is assumed that the change is transient and the value with the shift is best purged from the sample set. Is done.

  When the sensor is attached to a movable mirror assembly, mirror adjustments that can change the approximation often occur just before or when the driver begins a new part of the itinerary. For example, when turning back from a private road and turning to a roadway, when turning back from a parking position and turning to a roadway, or when going around a clover leaf-shaped interchange and entering a highway, a turn is probably made. These situations can often provide the best opportunity to quickly collect data points in changing directions. This data is potentially very useful for quickly correcting the compass approximation. However, it is dangerous to quickly purge all existing data before obtaining more points to confirm that the approximation has truly changed. After obtaining and temporarily storing these points, qualifying them later as valid sample points is a problem before there is a reasonable confirmation that they are stable and not transient It has the advantage of acquiring points when data is obtained while taking care to avoid purging good data points with data that can be proven.

  In addition to the above, it is beneficial for the compass system to receive an indication of whether the vehicle is moving. For example, if the vehicle is not moving, the compass system should be configured not to update the displayed heading and to change the stored approximate geometry (or associated point set). Can do. Thus, when the driver adjusts the mirror during the time that the sensor is attached to the mirror housing and the vehicle is stationary, the displayed heading does not change. This is effective because it is difficult for the driver to believe that the compass is operating properly if the heading changes while the vehicle is not moving. The indication of whether the vehicle is moving is in the form of a signal sent from another component in the vehicle (such as GPS) that simply indicates whether the vehicle is stationary or moving. Good. This signal does not need to indicate the speed at which the vehicle is traveling, but merely indicates whether it is moving. However, according to an embodiment of the present invention, a signal representing the speed of the vehicle can be used to determine whether the vehicle is stationary. Furthermore, the compass can be programmed to assume that the vehicle is moving if there is a minimum noise level (eg, greater than 2 mG) in the data read from the sensor over a specified time period. Alternatively, a microphone or vibration sensor may be used to determine whether the vehicle is moving.

  While the general concepts behind the approximation process of the present invention have been described above, a first embodiment that implements some of the above concepts will be described in detail below. However, the present invention is not limited to this specific example.

  Processing circuit 110 is preferably programmed to operate as a state machine. FIG. 11 shows the various operating states and criteria used to change the state. More detailed information regarding the state entry or exit will be further described with reference to the flowchart beginning with FIG. 12A. Referring to the state diagram shown in FIG. 11, the processing circuit first operates in the POWER_ON state 150 when the vehicle ignition is engaged. If the approximate geometric shape of the compass has not been previously calculated, no approximate data has been stored in the non-volatile memory (NVM) 112, and thus the processing circuit 110 is placed in the “approximate” state 152. Put in. The processing circuit remains in the “approximate” state 152 until a specified number of data points that satisfy certain criteria are obtained. From these points, approximate best fit data is obtained and stored in NVM 112, which is then used to calculate and display the vehicle heading. The processing circuit 110 then enters a “learn” (LEARN) state 156. When the vehicle is subsequently turned off and then turned on again, the processing circuit 110 enters the POWER_ON state 150 of the state diagram. However, at this time, valid approximate data is stored in the NVM 112, and thus the processing circuit 110 enters an “INITIALIZE” state 154. During this “initialization” state, the processing circuit 110 processes the data from the sensor until a minimum number of points satisfying a certain criterion is obtained, while simultaneously using the approximate data stored in the NVM 112 to perform vehicle heading. Calculate and display. At this point, the processing circuit 110 enters a “learn” state 156. During this “learning” state, the processing circuit 110 adds additional data that satisfies certain criteria for use in updating approximate data previously obtained in either “approximation” or “initialization” state. Keep accumulating typical data points. As will be described in detail below, if any of these data points are separated by a distance from the outer surface of the approximate geographic pattern (ie, the local sphere), the magnitude of the error vector VE exceeds a predetermined threshold. The processing circuit 110 sets the rewrite Flag to true, clears all previously obtained data points, and re-enters the “initialization” state 154, the circuit meets the specified criteria. Stay there until you get the minimum number of points again.

  When the processing circuit 110 is in the “learning” state 156 and obtains a specified number of data points that exceeds the minimum data point required to enter this “learning” state, the processing circuit 110 is in a “locked” (LOCK) state. 158 is entered and remains there until the magnitude of the error vector VE exceeding the specified maximum value is identified. At this point, the processing circuit returns to the “initialization” state 154, sets the re- mitFlag to “true” and clears all previously obtained data points. Otherwise, the processing circuit 110 remains in the “locked” state 158. Each time the difference between the most recently generated approximate data and the approximate data stored in the NVM 112 exceeds the limit, the old approximate data of the NVM is overwritten with new approximate data.

  FIG. 12A shows the steps performed by the processing circuit 110 during execution of the compass flow control algorithm 200. The first step performed by the processing circuit 110 under this routine is to receive data from the magnetic sensor circuit 102, which is the X, Y of the latest raw magnetometer point obtained from the magnetic sensor circuit 102. And the Z coordinate. The processing circuit 110 then executes step 202, which is a call to the Noise Analysis subroutine 204 shown in FIG.

The Noise Analysis subroutine 204 is used to filter and smooth the raw magnetometer points (rawMagPoint) received from the magnetic sensor circuit 102. In step 206, processing circuit 110 first filters the raw magnetometer points using exponential smoothing based on the following equation:
EX1 _n = α _x1 xX1 _n + (1-α _x1 ) EX1 _n-1
EY1 _n = α _y1 xY1 _n + (1-α _y1 ) EY1 _n-1
EZ1 _n = α _z1 xZ1 _n + (1-α _z1 ) EZ1 _n-1
However, α _x1 = α _y1 = α _z1 = 0.5, and X1 _n , Y1 _n , and Z1 _n are each equal to the X, Y, or Z component value of rawMagPoint. In this equation, EX1 _n−1 , EY1 _n−1 , EZ1 _n−1 each represent the corresponding X, Y or Z component value of the previously filtered raw data point. However, when the first data point is obtained by the compass system, there is no data to smooth the first rawMagPoint, so EX1 _n-1 , EY1 _n-1 , EZ1 _n-1 are X1 _n , It is simply set equal to Y1 _n and Z1 _n . The filtered rawMagPoint (EX1 _n , EY1 _n , EZ1 _n ) is then passed to step 208 where a second exponential smoothing with α _x2 = α _y2 = α _z2 = 0.25 is performed. Used in the next formula.
EX2 _n = α _x2 xEX1 _n + (1-α _x2 ) EX2 _n-1
EY2 _n = α _y2 xEY1 _n + (1-α _y2 ) EY2 _n-1
EZ2 _n = α _z2 xEZ1 _n + (1-α _z2 ) EZ2 _n-1
As described above, in the preferred embodiment, α _x1 = α _y1 = α _z1 = 0.5 and α _x2 = α _y2 = α _z2 = 0.25. However, these values can vary based on the noise level and the desired performance. For example, if α _z1 = 0.125 and α _z2 = 0.0625 are set, stronger and smoothed Z data is obtained, and consequently the responsiveness to noise conditions in the data set is reduced. In these equations, (EX2 _n−1 , EY2 _n−1 , EZ2 _n−1 ) represents the X, Y or Z component value corresponding to that of the previously filtered newMagPoint. By using a second filtering step, gain is added due to the inherent phase delay between the two filters, which increases with frequency. However, it is clear that this second filtering step is optional. Next, in step 210, the processing circuit 110 stores the value of (EX2 _n , EY2 _n , EZ2 _n ) in the variable newMagPoint (newMagPoint = (EX2 _n , EY2 _n , EZ2 _n )).

The change in the output of the first exponential smoothing filter in block 206 is then analyzed by comparing it with the change in the output of the second exponential smoothing filter in step 208. This analysis performed at step 212 specifies a quality index based on the amount of change measured and the length of time that has elapsed since the amount of significant change occurred. Thus, this step effectively calculates the first derivative of the location of the data point filtered twice, which represents the rate of change (ie, as well as velocity) of the location of the filtered data point. This can be written as:
DX1 _n = EX1 _n -EX2 _n-1
DY1 _n = EY1 _n -EY2 _n-1
DZ1 _n = EZ1 _n -EZ2 _n-1

In step 214, the processing circuit 110 uses the following equation to represent the second derivative of the twice-filtered data point position that represents the rate of change of the data point position rate of change (ie, similar to acceleration): Calculate
DX2 _n = DX1 _n -DX1 _n-1
DY2 _n = DY1 _n -DY1 _n-1
DZ2 _n = DZ1 _n -DZ1 _{n-1
However, DX2 n-1, DY2 n} -1, DZ2 n-1 is the previous value of DX2 _n, DY2 _n, DZ2 _n obtained for previous RawMagPoint. Step 206-214 are performed for each X, Y and Z values of RawMagPoint, then the magnitude of the thereby obtained vector _{D2 n (D2 n = SQRT (} DX2 n 2 + DY2 n 2 + DZ2 n 2)) Is compared with a prescribed value HIGH_LEVEL in step 216. When the D2 _n dimensions (MagnitudeD2 _n) exceeds HIGH_LEVEL, the processing circuit 110 executes step 218, where the variable delayCounter is set equal to a constant LONG_DELAY, and variable noiseLevel is equal to the constant NOISY Then, the process flow returns to step 240 (FIG. 12A) of the CompassFlowControl routine 200.

In step 216, MagnitudeD2 _n is in determining the processing circuit 110 does not exceed HIGH_LEVEL proceeds to step 220, to compare the MagnitudeD2 _n constant MED_LEVEL. MagnitudeD2 _{when n} exceeds MED_LEVEL, the processing circuit 110 proceeds to step 222, where the variable DelayCounter, is set equal to the greater of the previous value or the constant MED_DELAY of DelayCounter. In step 222, the processing circuit 110 sets noiseLevel equal to NOISY, and then returns to step 240 of the CompassFlowControl routine 200.

If Magnitude D2 _n does not exceed HIGH_LEVEL or MED_LEVEL, processing circuit 110 proceeds to step 224 where it determines whether Magnitude D2 _n exceeds a constant LOW_LEVEL. If the constant LOW_LEVEL is exceeded at step 224, the processing circuit 110 sets the variable delayCounter equal to the previous value of that delayCounter or the larger of the constant SHORT_DELAY, while setting the variable noiseLevel to NOISY (step 226), Thereafter, the process returns to step 240 of the routine 200.

If MagnitudeD2 _n does not exceed LOW_LEVEL, the processing circuit 110 decrements the delayCounter at step 228, and then proceeds to step 230 to determine whether the delayCounter is greater than zero. If delayCounter is greater than 0, the processing circuit 110 sets noiseLevel to QUIET at step 232 and then returns to step 240 of the routine 200. When the processing circuit 110 determines that the delayCounter does not exceed 0 in step 230, the process proceeds to step 234, and after setting noiseLevel equal to SILENT, the process returns to step 240 of the routine 200. The different values used in steps 216-226 may be, for example, HIGH_LEVEL = 25, MED_LEVEL = 15, LOW_LEVEL = 5, LONG_DELAY = 8, MED_DELAY = 4, and SHORT_DELAY = 2.

  31-34 illustrate another embodiment of a noise analysis routine. FIG. 31 shows a high level block diagram of a noise analysis function that receives raw sensor data for X, Y, and Z sensors and provides filtered X, Y, and Z sensor data along with noiseLevel instructions. When the noiseLevel is “SILENT”, all compass operations are executed. When the noiseLevel is “QUIET”, the change to the PointSet point is prohibited. When the noiseLevel is “NOISY”, the change to the PointSet point is prohibited and the change to the CompassState is also prohibited. The displayed heading is a member of CompassState and is therefore not changed when the “NOISY” state exists. The outputs NOSYY, QUIET, and SILENT are evaluated as “true” when their respective values are greater than zero. FIG. 32 is a detailed block diagram.

  As shown in FIG. 32, the raw data X from the magnetometer 102 is fed to a first exponential filter 820 that is similar to the filtering function of block 206 of FIG. 13, where α = 0.5. The output EX1 of the first filter 820 is supplied to a second exponential filter 822 and a functional block 824 described below. The second filter 822 is similar to the filtering function of block 208 in FIG. 13 and α = 0.25. The output EX2 of the second filter 822 is provided to the functional block 824. This functional block 824 performs the same function as block 212 of FIG. 13 and provides an output DX 1 that is sent to functional block 826. This functional block 826 performs the same function as block 214 of FIG. 13 and generates output DX2. Similar filtering and processing of the raw Y and Z data from the magnetometer 102 is performed and the values DX2, DY2, and DZ2 are sent to block 828.

  Block 828 squares the values of DX2, DY2, and DZ2, and adds the squared values together. The output of block 828 labeled “noise” is applied to block 830, which divides the output “noise” of block 828 by the specified value “noiseFactor”, takes the square root of that value, and specifies it The value “delayIntercept” is added to generate the output “delay”. This output “delay” is applied to the function block 832 and represents a NOISY state. In other words, if the value of “delay” is greater than 0, it can be seen that NOISY is “true” and therefore the NOISY state exists.

  The block 832 calculates the value of the function MIN using the value “delay” and the specified value “maxDelay”, and calculates the value of the function MAX using the value “delay” and zero. The MAX and MIN values are sent to blocks 834 and 838. The functional block 834 recalculates the value of the function MAX using the value “delay”, but replaces the value “peak” with zero. The value of “peak” is derived from the output of block 836, which subtracts the value of “1” from the output of block 834. The output of block 834 represents the state of the QUIET. In other words, if the output of block 834 is greater than zero, the QUIET is “true” and thus it can be seen that a QUIET condition exists.

  The output of block 834 is also applied to block 838, which adds the value of the output of block 834 to the output of block 832. The output of block 838 is inverted by inverter 840, and its output represents the state of SILENT. In other words, if the output of block 840 is greater than 0, it can be seen that SILENT is “true” and therefore a SILENT state exists.

The relationship between the delayed signal, defined as the square root of DX2 ² + DY2 ² + DZ2 ² , and the noise signal is shown in FIG. 33 when noiseFactor = 10, delayIntercept = −3, and maxDelay = 32.

Magnetic noise becomes less of an issue as the radius of the approximate sphere increases. For example, in Zealand, Michigan, USA, the radius is about 180 mG. If it is desired to ignore noise present above 10% of the signal, it will be related to the noise level in the range of 18-20 mG. In Singapore, the radius is about 400 mG. In this case, the noise level is twice or more. It is therefore beneficial to adapt the response of the system based on the horizontal strength of the geomagnetic field (represented by the radius of the approximate sphere). This can be easily achieved by adjusting the value of “delayIntercept”. For example:
Radius delayIntercept
≤128mG -2
≤256mG-3
≦ 512mG -4

The relationship between the delayed signal and the noise signal, defined as the square root of DX2 ² + DY2 ² + DZ2 ² , is shown in FIG. 34 when noiseFactor = 10, delayIntercept = −4, and maxDelay = 32.

Returning to FIG. 12A, at step 240, the processing circuit 110 determines whether it is currently in the POWER_ON state. If so, the processing circuit 110 proceeds to perform step 242 and determines whether the effective approximate radius rs _xsr is stored in the _NVM 112. Initially, such radii are not stored in non-volatile memory unless prior approximation data has previously been stored in the compass. If a valid radius is not stored in the NVM 112, the processing circuit 110 proceeds to step 244 and sets its state to “approximate”. The process then flows via connection “1” to step 246 shown in FIG. 12B.

  In step 246, processing circuit 110 determines whether noiseLevel is SILENT. If not, the processing circuit 110 returns to the CompassFlowControl routine 200, at which point after the new rawMagPoint is attached, it executes the Noise Analysis subroutine 204 described above again. The processing circuit 110 then determines in step 318 that it is in the “approximate” state and returns to step 246 and continues through the loop of steps 200-246 (and step 318) until noiseLevel becomes SILENT. The length of time it takes to obtain this result depends on the consistency of the data received and how it initially varies with the length of the delayCounter set. If noiseLevel is SILENT, processing circuit 110 executes step 248, which is a call to BuildPointSetDistance subroutine 250 shown in FIG.

  In a preferred embodiment, a delay of eg 10 seconds after power on is added, during which time magnetometer data is collected and passed through a noise filter. However, the BuildPointSetDistance routine 250 is not called until a 10 second delay has elapsed. This prevents various magnetic transients from learning the wrong point. These transient states can occur internally due to battery charging, door opening, mirror position adjustment, and the like. Transient conditions can also occur from the outside, such as the presence of gas pumps, metal garage doors, metal beams in parking lots, and the like. After the delay has passed, the BuildPointSetDistance routine 250 is called each time new magnetometer data is obtained until two points are found. This indicates that the magnetic field is likely to change due to the movement of the vehicle. The VehicleUnderway flag is then set to track this event. When this event occurs, the two learned points are discarded, and the BuildPointSetDistance routine 250 is called to start a new point set construction with Point [1]. Alternatively, a vehicle movement flag may be used.

  In the BuildPointSetDistance routine 250, the processing circuit 110 first determines at step 252 whether the variable numPoints is greater than zero. Initially, this variable is set to “0”, so when the routine is first passed, the processing circuitry proceeds to step 254 where the latest value of newMagPoint is referred to below as “PointSet”. Added to Point [1] of the array. PointSet stores multiple magnetometer data points up to the limit specified by the constant TOTAL_POINTS, for example, this constant is equal to 12, so PointSet stores a total of 12 points, and numPoint is "1 ”To TOTAL_POINTS. The data points stored in PointSet are then used to approximate the center point of the local sphere and determine the radius of the sphere, which together form approximate data, which is used to determine the vehicle heading. Gives a relative base for determination.

  In addition to adding newMagPoint to PointSet at step 254, the processing circuit increments the variable numPoints and then returns control to step 280 of the CompassFlowControl routine 200 (FIG. 12B). In this step 280, the processing circuit determines whether the variable numPoints is greater than or equal to the constant MINFIT_POINTS. If not, the processing circuit returns to step 200 where a new rawMagPoint is obtained and subjected to the noise analysis subroutine 204. The processing circuit 110 then proceeds with the new MagPoint obtained from the subroutine 204 and reenters the BuildPointSetDistance subroutine 250 if the noiseLevel is SILENT. Otherwise, the processing circuit 110 continues to achieve a new rawMagPoint until noiseLevel becomes SILENT again.

Upon entering the BuildPointSetDistance subroutine 250 a second time, the value of numPoints is no longer equal to zero, so the processing circuit executes step 256 to determine whether the value of numPoints is greater than one. Since the value of numPoints does not exceed 1 when this subroutine is passed the second time, the processing circuit then proceeds to step 258 where it determines the distance between newMagPoint and PointSet Point [1]. . Then, in step 260, the processing circuit determines whether this distance exceeds the variable minDist, which is defined as:
minDist = 2xdefRadiusxsin (MIN_ANGLE / 2)
However, MIN_ANGLE = 2xπ / (3xTOTAL_POINTS)
The value of defRadius is, for example, a preset constant of 150, or preferably the value of the approximate radius stored in the NVM 112. Thus, the minimum distance (minDist) is a function of the number of TOTAL_POINTS to obtain and use for approximation purposes. If the distance between the two points thus obtained does not exceed minDist, the subroutine returns to step 200 to acquire a new data point, which is obtained from minDist from the first acquired point. Continue until a second newMagPoint separated by a large distance is obtained, where the processing circuit executes step 262 to add newMagPoint to the PointSet at Point [2], while increasing the value of numPoints To do. The processing circuit then returns to step 200 to obtain and filter a new rawMagPoint.

  After obtaining the first two points of PointSet, the processing circuit goes through the steps described above and re-enters the BuildPointSetDistance subroutine 250 if noiseLevel remains at SILENT. When the BuildPointSetDistance subroutine 250 is entered, the processing circuit proceeds to Step 264 because the value of numPoints exceeds 1.

In step 264, the processing circuit finds the two points closest to the recently obtained newMagPoint among the two points stored in PointSet. When first passing this path of the flow diagram, the two points correspond to the two previously obtained points Point [1] and Point [2]. Then, in step 266, the processing circuitry determines whether the distance between the closest point in PointSet and newMagPoint exceeds the value of the variable newPointDist, which is defined as:
newPointDist = 2xdefRadiusxsin (NEW_POINT_ANGLE / 2)
However, NEW_POINT_ANGLE = 2xπ / TOTAL_POINTS
If this distance exceeds the value of newPointDist, the processing circuit increments the value of numPoints and then adds newMagPoint to the PointSet at position Point [numPoints] (step 268). The processing circuit then returns to step 200 where it obtains and processes rawMagPoint. Again, assuming that NoiseLevel remains at SILENT, the BuildPointSet Distance subroutine 250 is executed again. This subroutine continues to be called at step 248 until the value of numPoints is greater than or equal to the constant MINFIT_POINTS set equal to 4, for example.

  Assuming subroutine 250 is entered again and in step 266 the processing circuit determines that the closest distance does not exceed the value newPointDist, the processing circuit adjusts the position of the point closest to the direction of newMagPoint in step 270. To do. For example, if any of the X, Y and Z values of newMagPoint are greater than the X, Y and Z values of the nearest point, such X, Y and Z values of the nearest point are It can be increased by a value of “1”. Similarly, if any of the newMagPoint X, Y or Z values are lower than any of the nearest points, the corresponding X, Y or Z value of the nearest point is set to, for example, “1”. It can be reduced by a value. Thereafter, in step 272, the processing circuit finds the distance between the PointSet point whose position has just been adjusted and the closest point. Then, in step 274, if the processing circuit determines that this distance is less than the value minDist, it proceeds to step 276 and resets the position of the point whose position has just been adjusted to its original position. Otherwise, the processing circuit simply returns to step 200 without performing step 276 first. These latter two steps are performed to ensure that the PointSet point does not adjust its position so that it is too close to another PointSet point. This ensures that the points are maintained in a well-separated relationship around the local sphere.

If the value of numPoints is greater than or equal to MINFIT_POINTS, the process proceeds to step 282 (FIG. 12B) where the approximate best fit center points (sc _xs , sc _ys , sc _zs ) are set to the average value of all points in PointSet and The best fit radius rs _xsr is set to the constant MIN_RADIUS. This step is an initial step that is used before proceeding through the iterative subroutines AdjustBestFitPoint and AdjustBestFitRadius, which are called in subsequent steps 283 and, as described below, based on the error metric Further adjustments are made to give the best fit approximate local sphere. As shown in FIG. 15, the AdjustBestFitPoint subroutine 284 first determines a value (ve _xs , ve _ys , ve _zs ) for each point currently stored in PointSet (step 286). The value of ve _xs for a given point is the X component of the vector VE. As described above and shown in FIGS. 8-10, the vector VE has a vector V extending from the center of the approximate local sphere to a specific data point in PointSet and a calculated approximate radius vector RS (with magnitude rs _xsr Error vector which is the difference between The values ve _ys and ve _zs correspond to the Y and Z component values of the vector VE, respectively.

After step 286, the processing circuit executes step 288 and adds the average value (ve _xs , ve _ys , ve _zs ) for all points in PointSet to the current approximate center point (sc _xs , sc _ys , sc _zs ). . Then, when executing the AdjustBestFitRadius subroutine 290, the processing circuit determines a value v _xsr for each point in _PointSet (step 292). This value v _xsr corresponds to the magnitude of the vector V as described above with respect to FIGS. The processing circuit then sets the approximate radius rs _xsr to the average value of v _xsr for all points in _PointSet (step 292). Then, in step 296, the variable vErrXrLimit is set equal to rs _xsr / 2, and the variable centerShiftLimit is set equal to rs _xsr / 4. The significance of these variables is described in detail below.

In order to calculate the error vector of newMagPoint, the magnetic field strength RDG shown in FIG. 9 (or newMagPoint in the flow diagram)
RDG = (rdg _xs , rdg _ys , rdg _zs ) = newMagPoint
And the magnetic field strength at the center of the approximate sphere SC = (sc _xs , sc _ys , sc _zs )
To obtain the vector V.
V = RDG-SC = (v _xs , v _ys , v _zs )
The vector V is then rotated around the Z _S axis to give new coordinate spaces X _i , Y _i , Z _i .
φ _s = arctan (v _ys / v _xs )
v _xi = v _xs xcos (φ _S ) + v _ys xsin (φ _S )
It then rotates around the Z _i axis to give new coordinate spaces X _sr , Y _sr , Z _sr .
θ _s = arctan (v _zs / v _xi )
v _xsr = v _xi xcos (θ _s ) + v _z xsin (θ _s ), v _ysr = 0, v _zsr = 0
The term v _xsr obtained thereby is the magnitude of the vector V. The error vector VE is defined as follows.
VE = V-RS
In order to determine the size of the error vector VE, the size of the approximate sphere is subtracted from the size of the vector V.
ve _xsr = v _xsr −rs _xsr
Finally, the individual components of the error vector are determined by performing a rotation back to the coordinate space X _s , Y _s , Z _s .
ve _xs = ve _xsr xcos (θ _s ) xcos (φ _s )
ve _ys = ve _xsr xcos (θ _s ) xsin (φ _s )
ve _zs = ve _xsr xsin (θ _s )

  As will be apparent to those skilled in the art, the AdjustBestFitPoint subroutine 284 and the AdjustBestFitRadius subroutine 290 adjust the approximate center point data based on the error metric to obtain the best fit of the approximate geometric pattern (in this case a sphere) to the resulting data points. Determine repeatedly. Then, after execution of these two subroutines, the process returns to step 300 of FIG. 12B where the currently calculated approximate center point and radius are stored in the NVM 112 to overwrite previously stored approximate data, The process then proceeds to step 302 where the processing circuit enters a “learn” state.

  Following step 302, the processing circuitry executes step 304 and calls a heading subroutine 306 shown in FIG. Heading subroutine 306 has an input that is the value currently stored in newMagPoint, which is generated by noise analysis subroutine 204. The processing circuit then determines in step 308 whether the value noiseLevel has been set equal to NOISY. If so, the processing circuit exits the heading subroutine and returns to step 200 of the CompassFlowControl routine to obtain a new rawMagPoint.

If the processing circuit determines in step 308 that noiseLevel is not set to NOSYY (ie, either QUIET or SILENT), the processing circuit executes step 310 and stores the approximate center stored in NVM 112. Using the points (sc _xs , sc _ys , sc _zs ) and the radius rs _xsr , calculate the heading angles φ _s , ve _xsr and error vectors (ve _xs , ve _ys , ve _zs ) for _newMagPoint . The method used to determine these values is as described above, except that the approximate center point stored in NVM 112 is used in place of the current approximate center point.

Then, in step 312, the processing circuit determines _whether the value ve _xsr calculated for newMagPoint exceeds the variable vErrXrLimit previously set in step 296 of the AdjustBestFitRadius subroutine 290 (FIG. 16). _If the value of ve _xsr exceeds this limit, the process returns to step 200 of FIG. 12A without displaying a heading. Otherwise, if this error vector is sufficiently small, the processing circuit calculates the true heading at step 314. The true heading is equal to φ _s −φ _m −D (when the magnetic sensor is mounted in the mirror housing 144), where φ _m is the rotation angle of the mirror (FIG. 7), and D is The declination described above with reference to FIGS.

  After executing the heading subroutine 306 following the call of step 304, the process returns to step 200 and the processing circuitry is now in the “learn” state set in step 302. The routine again obtains a new rawMagPoint at step 200 and executes the noise analysis subroutine 204 as previously performed for each new rawMagPoint. The processing circuit then determines in step 240 that it is no longer in the POWER_ON state and in step 318 determines that it is not in the “approximate” state. Then, at step 320, it is determined that it is not in the “initialization” state, and in this case, at step 322, it is determined whether it is in the “learning” state. If the processing circuit is in the “learning” state, the process proceeds to step 350 (FIG. 12C), otherwise it is in the “locked” state, and in this case, the process proceeds to step 470 in FIG. 12D.

If the vehicle ignition is turned off while the vehicle is in a “learning” state or other state, the processing circuitry will first begin when power is restored by turning on the vehicle ignition. Enter POWER_ON state. The processing circuit begins at step 200 and proceeds to step 240 where it determines that it was actually in the POWER_ON state and proceeds to step 242. Here, assuming that the effective radius rs _xsr is stored in the _NVM 112, the processing circuit proceeds to step 324, sets the state to the “initialized” state, and then proceeds to step 326. In step 326, the approximate center point (sc _xs , sc _ys , sc _zs ) is retrieved from _{NVM 112} along with radius rs _xsr and used as the current approximate center point and radius. The process then proceeds to step 328 shown in FIG. 12B.

  In step 328, the processing circuit determines whether the variable noiseLevel is SILENT. If not, the processing circuit calls heading subroutine 306 at step 330, returns to step 200 to process a new rawMagPoint, and proceeds to step 240 to determine that it is no longer in the POWER_ON state. In this case, the process proceeds to step 320 through step 318 where it is determined that it is not in the “approximate” state, where it is determined that it is in the “initialized” state. The process then returns to step 328. This continues to loop until noiseLevel is SILENT. If noiseLevel is SILENT, step 332 is executed and the subroutine BuildPointSetDistance 250 is called. This BuildPointSetDistance subroutine 250 is then executed continuously, and as long as noiseLevel is SILENT, the heading subroutine is called until numPoints is greater than or equal to the constant MINFIT_POINTS (step 334). The execution of BuildPointSetDistance is necessary because all values of PointSet are cleared when vehicle ignition is turned off. Clearing the value of PointSet each time the vehicle is turned off means that the compass system will quickly move when a new driver subsequently enters the vehicle and shifts the position of the mirror housing 144 to which the magnetic sensor is attached. Allows re-approximation.

If a sufficient number of points are obtained in PointSet, the processing circuitry proceeds to step 336 and the AdjustBestFitPoint and AdjustBestFitRadius subroutines are called. Then, in step 338, the processing circuit calculates the difference between the current approximate center point (sc _xs , sc _ys , sc _zs ) and the point currently stored in the NVM 112. If the difference between these points exceeds the variable centerShiftLimit set in step 296 (step 340), the processing circuitry proceeds to execute step 300 after setting the rewriteFlag to “true” in step 342. , Where the currently calculated approximate center point and radius are stored in the NVM 112 to overwrite previously stored approximate data. The process then proceeds to step 302 where the processing circuit enters a “learn” state. If, at step 340, the processing circuit determines that the difference between the currently calculated approximate center point and the point currently stored in the NVM 112 does not exceed the variable centerShiftLimit, the processing circuit Skip 300 and go to step 302. In step 302, the state is set to "learn" and in step 304 the heading subroutine is called. Following this heading subroutine, the process returns to step 200 of the CompassFlowControl routine or a new rawMagPoint is obtained and processed. Control then flows through step 322 to determine if the processing circuit is in a “learning” state. The processing circuit then proceeds to step 350 (FIG. 12C) and determines whether the value noiseLevel is SILENT. If not, the heading subroutine is called at step 352 and if the noiseLevel is QUIET, the vehicle heading is updated; otherwise, the heading is not updated and the routine returns to step 200 where the noiseLevel is Process a new rawMagPoint and subsequent rawMagPoint until it becomes SILENT. When noiseLevel becomes SILENT, the processing circuit executes step 354 and calls the Update Angle Bucket subroutine 356 shown in FIG.

  The Update Angle Bucket subroutine 356 serves the purpose of establishing a number of angle buckets that numerically correspond to the total number of points in the PointSet stored in the constant TOTAL_POINTS. If the total number of points is 12, for example, there are 12 anglebuckets, each corresponding to a mutually exclusive 30 ° segment of a 360 ° circle. anglebuckets are specified as a linear array in the form of anglebucket [1] to anglebucket [TOTAL_POINTS]. Thus, for example, angle Bucket [1] corresponds to an angle from 0 ° to 29 °, angle Bucket [2] corresponds to an angle from 30 ° to 59 °, and so on. The more anglebuckets, the narrower the angular range corresponding to anglebuckets. Conversely, the smaller the anglebuckets, the wider the angular range associated with each anglebucket. The value stored in the angle Bucket array corresponds to the index of the associated Point [index] stored in PointSet. When “0” is stored in the anglebucket, it indicates that there is no related point in the PointSet stored in this specific anglebucket. As will be described in detail below, when a negative value is stored in the anglebucket, the PointSet Point [index] corresponding to the absolute value of the negative number stored in the anglebucket is different from the anglebucket. Means that the different anglebuckets are already full. Note that only one point is allowed to be associated with an anglebucket. This ensures that the points are well spaced around the approximate circle. Its purpose is to fill each anglebucket with a different single point in the PointSet so that each anglebucket has a positive value.

The heading angle φ _s associated with each sensor data point of the current sample PointSet is used to arrange these points in the order starting with the point with the smallest associated heading angle and ending with the point with the largest associated heading angle. Is preferred. The angle must be expressed in modulo 360 ° with angle increments including 0 ° processed as usual in compass readings. Other units such as radians or conventional units may be used to represent angle measurements.

The first step of the Update Angle Bucket subroutine 356 is to zero all angle Buckets (step 358). This allows anglebuckets to re-point the PointSet point to fresh. Then, in step 360, the angle φ _s is calculated for the first point of PointSet (ie Point [1]) and the value of φ _s is used to identify which anglebucket this point corresponds to. Then, in step 362, the processing circuit determines whether the selected anglebucket has the value “0”. If so, the PointSet point index is stored in this angle Bucket (step 368). For example, when the first point is Point [1], the index value “1” is stored in the selected anglebucket. Then, in step 366, it is determined whether there are more points in the PointSet to be specified in the angleBucket. If so, processing circuit 110 calculates φ _s for the next point in PointSet and uses this angle to select the corresponding anglebucket (step 360). If the value “0” is stored in the selected anglebucket, the process again proceeds to step 364 where the PointSet point index is stored in the selected anglebucket. This process continues until the processing circuitry determines in step 366 that there are no more points in PointSet to specify. In this case, the processing circuit performs step 368 to determine if there are any stored indexes from step 370. If not, the Update Angle Bucket subroutine 356 ends and the process returns to step 376 (FIG. 12C).

In the Update Angle Bucket subroutine 356, if the PointSet point selected in step 360 has an angle φ _s and the selected angle Bucket has been previously selected relative to another point, the value of the angle Bucket is no longer Not “0” and this is determined in step 362 and the processing circuit then proceeds to step 370 and stores the index of the most recently selected point in PointSet for later use in steps 368 and 374. To do. The routine then accumulates an additional index of any point corresponding to the already specified anglebucket while continuing to look for further points in PointSet at step 372. Then, once all the points have been specified, the process continues to step 368 where it is determined if there are any stored indexes that occurred in step 370. If so, the processing circuit places the negative value of the stored index into an angle Bucket with a value of zero. Therefore, for example, if Point [2] is already specified in angle Bucket [5], the value “2” is stored in angle Bucket [5]. Then, if Point [7] has an angle φ _s corresponding to the angle range of angle Buckets [5] that has already stored the value “2”, the process stores the index “7” at step 370. This is identified at step 368. The process then puts “−7” into one of the unused anglebuckets whose value is “0”. Note that there is always an anglebucket with a value of “0”, since the number of anglebuckets corresponds to the maximum number of points that can exist in the PointSet. The significance of the number of negative indexes in non-compatible anglebuckets will be described below. Once all the points in PointSet have been assigned to anglebucket in some way, the process returns to step 376 of FIG. 12C.

  In step 376, it is determined whether the current value of numPoint is less than TOTAL_POINTS. If numPoints is equal to TOTAL_POINTS, the process proceeds to step 378 and a determination is made whether there are anglebucket entries (ie, negative entries) that are less than zero. If not, the process proceeds to step 380 and the processing circuitry is set to the “locked” state. The approximate center point and radius are then stored in NVM 112 (step 382), and in step 384 the heading subroutine is called. The process then returns to step 200 (FIG. 12A).

  If it is determined in step 376 that numPoints is less than TOTAL_POINTS, or if it is determined in step 378 that a negative value is stored in any anglebuckets, the processing circuit proceeds to step 386. Proceeding, the BuildPointSetAngle subroutine 388 shown in FIG. 19 is called.

The BuildPointSetAngle subroutine 388 is called after the UpdateAngle Bucket subroutine 356 is executed. However, the BuildPointSetAngle subroutine 388 uses the most recently obtained newMagPoint as an input, while the UpdateAngle Bucket subroutine uses only the points stored in the PointSet. The purpose of the BuildPointSetAngle subroutine is to specify a newMagPoint as a point in the PointSet when possible and a new point in the angle Bucket. The process begins at step 390 where an angle φ _s is calculated for newMagPoint and is used to select an angle Bucket with an angle range that includes the angle φ _s . In step 392, it is determined whether a negative value is stored in the selected anglebucket. If so, the processing circuit identifies which point in the PointSet is associated with that anglebucket by looking at the absolute value of the value stored in that anglebucket and is stored in that point in the PointSet. The X, Y, and Z values for the selected point are replaced with the X, Y, and Z values of newMagPoint (step 394). The processing circuit further stores the positive value of the index of the replaced point in the selected anglebuckets, so that the newMagPoint is associated in that particular anglebucket. The process then returns to step 408 of FIG. 12C. If it is determined at step 392 that the selected anglebucket does not have a negative value, then at step 396 a determination is made as to whether the selected anglebucket is equal to zero. If this is not equal to zero, meaning that another point has already been specified for the anglebucket, the position of the point in the PointSet specified by the anglebucket is about step 270 of the BuildPointSet Distance subroutine 250 (FIG. 14). As described above, adjustment is made in the direction of newMagPoint (step 398). Otherwise, if the selected anglebucket is equal to zero, the processing circuit determines in step 400 whether the anglebucket has a negative value. If not, the value of numPoints is incremented, newMagPoint is added to PointSet at its index numPoints, and the selected anglebucket is set to the value of numPoints, so here newMagPoint is associated with the selected angleBucket (Step 402).

  If, in step 400, the processing circuit determines that there are several anglebuckets with a negative index stored, it executes step 404 with the absolute value of the negative index stored in the anglebucket and the newMagPoint data. The point of the identified PointSet was written over the data previously stored at that identified point of the PointSet and the index of the overwritten point previously had the value “0” It is stored in the selected anglebucket. Next, the value “0” is specified for the anglebucket having a negative index. The process then returns to step 408 of the flow shown in FIG. 12C, which is a call to the CalcFittingSphere subroutine 410 shown in FIG.

The first step in this subroutine is to determine the average θ _s for all points in PointSet (step 412). The θ _s for each PointSet point is found by calculating θ _s = arctan (v _zs / v _xi ) for each PointSet point. Note that the angle θ _s represents the deviation of the sensor coordinates from the X and Y planes. The change in the angle θ _s becomes zero when the X and Y planes are parallel to the ground surface. The inclination of the X and Y planes causes a change in θ _s . Ideally, the circle formed by the PointSet points should lie completely in a plane as close as possible to the X, Y plane. In practice, however, each point may have a different Z component that is sensed by the Z sensing element 108, giving different θ _s values.

After calculating the average value θ _s for all points in PointSet in step 412, the processing circuit determines in step 414 whether numPoints is equal to TOTAL_POINTS. If so, the average value θ _s is compared with the value LOW_LIMIT in step 416. If the average value θ _s does not exceed LOW_LIMIT, the process proceeds to step 420 and returns to step 430 (FIG. 12C) after the AdjustBestFitPoint and AdjustBestFitRadius subroutines are called. Otherwise, the process proceeds to step 418 and the approximate radius rs _sr is set to the constant MIN_RADIUS. Thereafter, in step 420, the AdjustBestFitPoint subroutine is called, and then the process returns to step 430 of FIG. 12C.

If the value of numPoints is not equal to TOTAL_POINTS in step 414, the processing circuit determines in step 422 whether the value of numPoints is greater than or equal to the value of FIT_POINTS, for example, the value of TOTAL_POINTS is 12. Sometimes eight. If the value of numPoints is greater than or equal to FIT_POINTS, the processing circuit determines in step 424 whether the average of θ _s is greater than the value MED_LIMIT. If not, the process returns to step 430 of FIG. 12C, otherwise proceeds to step 418, where the approximate radius rs _xsr is set to MIN_RADIUS and in step 420 the AdjustBestFitPoint and AdjustBestFitRadius subroutines are called. Later, the process returns to step 430. If the value of numPoints is not equal to TOTAL_POINTS and does not exceed FIT_POINTS, the processing circuit determines in step 426 whether the average value θ _s exceeds the value of HIGH_LIMIT. If the average value θ _s does not exceed the value of HIGH_LIMIT, the process proceeds to step 430 of FIG. 12C, otherwise it returns after performing steps 418 and 420 first. As will be apparent to those skilled in the art, different limits on the mean value θ _s apply based on the number of points in the PointSet. The more points, the lower the tolerance limit. In this case, the radius of the sphere is too large. This is because the PointSet points form a ring far away from the equator of the sphere (ie, the radius of the sphere is greater than the radius of the point circle at PointSet). By reducing the radius of the sphere to MIN_RADIUS, the circle radius of the point circle fits within the point circle on the sphere and the size of the sphere is reduced. The smaller the sphere, the closer the PointSet point will be near the equator of the sphere, which is desirable. The AdjustBestFitPoint and AdjustBestFitRadius subroutines then move the radius and center back to the point circle (if not already there). The different limits used in steps 416, 424 and 426 may be, for example, LOW_LIMIT = 5 °, MED_LIMIT = 10 °, and HIGH_LIMIT = 20 °.

Returning to FIG. 12C, in step 430, the processing circuit finds the maximum value of ve _xsr in _PointSet . If the maximum ve _xsr exceeds vErrXrLimit (step 432), the processing circuit calls the ResetPointSet subroutine 436 at step 434. The ResetPointSet subroutine 436 is shown in FIG. 21 and begins at step 438 where the rewriteFlag is set equal to “true”. Then, in step 440, each point in PointSet is reset to zero. The process then returns to step 442 of FIG. 12C, the state of the processing circuit is set to “initialized”, and the process returns to step 200 to begin reaccumulating points for PointSet.

If the maximum ve _xsr does not exceed the limit at step 432, the processing circuit determines at step 444 whether the value of numPoints is less than FIT_POINTS. If numPoints is less than FIT_POINTS, the process returns to step 200 of FIG. 12A after the heading subroutine is called at step 446. However, if numPoints is greater than or equal to FIT_POINTS, the processing circuit determines in step 448 between the current approximate center point (sc _xs , sc _ys , sc _zs ) and the approximate center point currently stored in the NVM 112. Calculate the difference. If this difference is greater than centerShiftLimit (step 450), the processing circuit proceeds to step 452 and after saving the most recent approximate center point and radius to NVM 112 by overwriting the previous approximate data, In step 446, the heading subroutine is executed, and then the process returns to step 200.

  If the difference between the approximate center points does not exceed centerShiftLimit, the processing circuit checks in step 454 whether rewriteFlag is set to “true”. If not, the process proceeds to step 446 and returns to step 200 after the heading subroutine is called. Otherwise, if the reflectFlag is “true”, the reflectFlag is set to “false” in step 456 and the most recently obtained approximate center point and radius are overwritten in step 454 to the approximate data of the NVM 112. Written. Again, following step 452, the heading subroutine 446 is called and then returns to step 200.

If the processing circuit is set to the “locked” state, it is determined when it does not respond positively to the determination that it is in any of the other states as it proceeds through the CompassFlowControl routine 200 shown in FIG. 12A. Is done. Accordingly, the process proceeds to step 470 of FIG. 12D, where a determination is made whether noiseLevel is SILENT. If not, the process returns to step 200 of FIG. 12A after calling the heading subroutine in step 472. This continues until noiseLevel is SILENT, at which time step 474 is executed and the Update Angle Bucket subroutine 356 is called. Thereafter, the BuildPointSetAngle subroutine 388 is called at step 476 and then the CalcFittingSphere subroutine 410 is called at step 478. Again, the maximum value of ve _xsr for the PointSet point is determined in step 480 and the maximum ve _xsr is compared to vErrXrLimit in step 482. If the maximum ve _xsr exceeds this limit, the processing circuitry proceeds to step 486 and after calling the ResetPointSet subroutine 436, enters the “initialization” state at step 488 and then returns to step 200 of FIG. 12A. .

If the maximum ve _xsr does not exceed the limit at step 482, the processing circuit executes step 490 to calculate the difference between the current approximate center point and the approximate center point currently stored in the NVM 112. To do. If the difference between these points exceeds the centerShiftLimit in step 492, the most recent approximate center point and radius are stored in the NVM 112 instead of the previous approximate data (step 496), and then the heading subroutine is executed in step 494. Execute, and then return to step 200. If the difference between the approximate center points does not exceed centerShiftLimit, step 496 is bypassed and the heading subroutine is called at step 494 and then returns to step 200.

  Some modifications to the method are described below. FIG. 35 shows a circle with 12 angle buckets, each defined by a unique 30 ° arc. Points are added to the angle bucket as follows:

  1. If the target angle bucket already contains a point, the existing point is changed by moving it in the direction of the new point.

  2. If the angle bucket is empty, the adjacent angle bucket is tested. If an adjacent bucket contains a foreign point (defined as a point belonging to another angle bucket), the foreign point is removed from the adjacent bucket and a new point is placed in the target angle bucket.

  3. If adjacent buckets are properly filled, test all buckets to see if they contain extraneous points. If a foreign point is found, it is removed and a new point is placed in the target angle bucket.

  4). If no extraneous point is found, a new point is added to the PointSet and the target angle bucket.

  A new approximate sphere is calculated when PointSet is changed. If there are fewer than 12 points and the number of full angle buckets is 8 or more, the difference between the new approximate sphere and the approximate sphere stored in the NVM is calculated. If this difference is greater than 1/4 of the stored sphere radius, a new approximate sphere is stored as follows.

  1. The new approximate sphere radius is averaged with the previous eight radii to determine the approximate sphere radius to be stored in the NVM.

  2. If the number of full angle buckets is 8 or less, the center point of the approximate sphere saved in the NVM is the difference between the new approximate sphere and the sphere stored in the NVM. Is divided by 8 and added to the sphere stored in the NVM. This has the effect of moving the old NVM center point towards the new approximate center point by 1/8 of the distance between the two center points.

  3. When the number of full angle buckets is 9, the moving distance is 1/4 of the difference.

  4). When the number of full angle buckets is 10, the moving distance is ½ of the difference.

  5. If the number of full angle buckets is 11 or 12, the new approximate sphere center point simply replaces the old center point saved in the NVM.

  As long as the number of points in PointSet is kept below 12, a change in the approximate sphere stored in NVM can only occur after a new point has been added to PointSet. When the approximate sphere in the NVM is updated, a new point must be added before the NVM can be changed again. In other words, if there are 8 points in PointSet and the new approximate sphere is greater than ¼ of the radius from the stored approximate sphere, the update to the stored sphere is performed in a prescribed manner. Is called. Even if the separation is greater than 1/4 of the radius, no additional updates to the NVM can be made until the number of points in the PointSet is 9 or more. This restriction is no longer imposed once all 12 Points of PointSet are specified. When twelve full angle buckets occur first, a copy to NVM is automatically triggered as described above only if the separation is greater than 1/4 of the radius.

  Although the electronic compass of the first embodiment of the present invention has been generally described as having three sensing elements, the various aspects described above and combinations of these aspects are compass circuits having only two sensing elements. Represents a new improvement to. Therefore, the present invention is not limited to a system having three sensing elements.

  Two-dimensional or three-dimensional compass compensation techniques are specified to correct compass errors due to hard and soft magnetic field effects. These errors are often large but are usually very slow and can therefore be adequately corrected by the use of various adaptive algorithms. Vehicle magnetic field changes can be caused by external sources such as railroad tracks, power transmission lines, car wash machines, roof antennas and the like.

  There is often a compromise between the speed of adaptation and the stability of the calibration point, defined as the point at which the compass is considered to be correctly compensated. The method described above and described in US Patent Application Publication No. 2003/0167121 A1 provides very quick compensation with minimal loss of stability.

  Non-fixed compass sensors, such as sensors located on a moving vehicle, experience the dynamic action of pitch and roll. Compared to the more static hard and soft magnetic field effects, pitch and rolls cause continuous errors that are not often corrected for land vehicles due to cost and signal processing constraints. The use of pitch and roll sensors is a common practice in fields that require more accurate heading information. More specifically, aviation and marine electronic compass designs provide an electronic gimbal by using two additional sensors, one that measures the compass pitch and the other that measures the roll. Is a common practice. The electronic gimbal corrects the compass heading error in real time by the pitch and roll of the sensor. Commonly assigned US Pat. No. 6,140,933 discloses the use of an inclinometer in an automobile compass to correct the tilt of the mirror housing to which the sensor is mounted.

  Finally, in the case of a magnetic field sensor placed in the mirror housing, there is a pitch and roll static effect such that the change in housing tilt is a static pitch and / or roll change. Given sufficient time, most adaptive compass compensation techniques provide a level of correction for errors associated with such static pitch and roll changes. Here, a 3-axis compass is particularly well suited. This is because the tilt of the sensor from the horizontal causes a large shift in the sensing field in both the horizontal and vertical directions. If this event can be detected quickly, it can be quickly corrected as taught above and in US Patent Application Publication No. 2003/0167121 A1.

  In the event of static changes in the pitch and / or roll of a two-axis compass, compass compensation, often referred to as calibration values, is described in US Pat. Nos. 6,023,299, 6,140,933, and As disclosed in US Pat. No. 6,418,376, moderate modifications can be obtained by changing only the difference between the magnetic fields experienced before and after the occurrence of static changes. However, in the case of a 3-axis compass, updating the calibration value by the amount of change measured may increase heading error in some cases rather than minimizing the error. If a position sensor is available, another technique must be found for updating the compass calibration when a change in sensor position is detected.

  The prior art teaches using a position detector to determine that a static change in sensor pitch / roll has occurred. Similarly, electronic gimbals are a well-understood technique for correcting dynamic changes in sensor pitch / roll. The variable pitch / roll sensor is made by using an analog device ± 2g dual axis accelerometer ADXL202E. Appropriate filtering to obtain both dynamic and static pitch / roll changes, allowing correction of heading errors due to dynamic pitch / roll changes and compensation changes due to static pitch / roll changes Additional benefits can be obtained.

  Figures 36 and 37 show the effect of a 10 degree tilt of the sensor about the y-axis for two different geographic locations.

The following table contains the above two positions and shows three additional positions.

  Its purpose is to determine a new calibration point C1 given an existing calibration point C0 and two points, P0 found just before the shift and P1 found just after the shift. If the main field and the slope are known, it is easy to calculate the actual value of C1. However, only the horizontal field R is known. The geographical location of the compass is not known, so the main field and slope are unknown. FIG. 38 shows what is known.

Using C0, P0 and P1, the following vector can be defined.
A = (P1 _x −C0 _x , P1 _y −C0 _y , P1 _z −C0 _z ), and B = (P1 _x −P0 _x , P1 _y −P0 _y , P1 _z −P0 _z ),
_{However, C0 = (C0 x, C0} y, C0 z), P0 = (P0 x, P0 y, P0 z),
P1 = (P1 _x , P1 _y , P1 _z )

  The unit vector in the direction of vector A is defined as N = A / | A |, and the projection of B in the direction of A is defined as p = (A · B) / | A |. Therefore, the point C1 '= C0 + p * N. Substitution results in the solution C1 ′ = C0 + (A · B) / | A || A |.

  Alternatively, if the radius R is also known, the unit vector can be defined as U = −A / | A | in the direction of −A, C1 ′ = P1−R * U or C1 ′ = P1 −R * A / | A |

In summary, there are three methods:
Method 1, shift method: Given C ″, C1 is determined from the vector defined by P1 and P2 from C0.
Method 2, projection method: C1 ′ = C0 + (A · B) / | A || A |
Method 3, radius method: C1 ′ = P1-R * A / | A |

  In the previous technique, only method 1 has been described. Method 3 is the most accurate, but finds the magnitude of the vector defined by P0 and C0, or R by some other means as described above and in US Patent Application Publication No. 2003/0167121 A1. It is necessary to know. Method 3 also requires a square root to determine | A |, but does not require a square root to calculate | A || A |.

In this example, the rotation is about the y-axis, so the value of y does not change. Here, the point coordinates shown are point = (x, z), and P1 results from a 10 ° rotation.

A newly calculated calibration point C1 ′ is indicated from each method along with the magnitude of the vector between the expected point C1 and the calculated point C1 ′. This magnitude error is compared with the length of the expected vector defined by points C0 and C1 to get a percent error.

  Obviously, methods 2 and 3 are significant improvements over method 1.

  If pitch and roll sensors are available, the output of these sensors can be used to correct both dynamic and static tilt of the compass sensor. FIG. 39 is a block diagram of a circuit that can be coupled to the microprocessor of the compass processing circuit 110.

  Again, the analog device ADXL202E is one possible dual-axis tilt sensor 900 that can convert acceleration to tilt. Suitable values are R1 = 125 kΩ, R2 and R3 = 425 kΩ, C1 and C2 = 0.47 μF, C3 and C4 = 4.7 μF. These values are selected for stable operation and are intended to match the filter characteristics applied to the magnetometer data. Thus, any time the measured pitch or roll error is matched with the corresponding measured magnetic field.

An accelerometer is most sensitive to tilt when its sense axis is perpendicular to gravity (parallel to the ground surface). When the output signal from the accelerometer is converted to acceleration that varies between -1g and + 1g, the power output is calculated as follows.
Pitch = asin (Ax / 1g)
Roll = asin (Ay / 1g)

  The measured pitch and roll can be incorporated into the methods taught above and in US Patent Application Publication No. 2003/0167121 A1. Immediately after calculating the mean magnetic field (after the noise filter) and before updating the PointSet point or determining the current heading, rotate the coordinates to the earth plane using the measured pitch and roll angle Then return.

Using the right hand coordinate system and defining the right hand rotation as a positive angle rotation, a coordinate transformation is performed using the following equation, where the filtered magnetometer points are (Hx, Hy, Hz) Then, the point that is rotated back to the surface of the earth is as follows.
Hex = Hx * cos (pitch) −Hy * sin (roll) sin (pitch) −Hz * cos (roll) sin (pitch)
Hey = Hy * cos (roll) -Hz * sin (roll)
Hez = Hx * sin (pitch) + Hy * sin (roll) cos (pitch) −Hz * cos (roll) sin (pitch)

  The second embodiment of the present invention has a number of similarities to the first embodiment. For example, both embodiments can be implemented using the same hardware. However, the specific example of the second embodiment described below is different in that a two-dimensional approximate geometric pattern is plotted as long as only the X and Y axis sensors are used. However, the second embodiment described below can be modified to merge the third sensor and / or manipulate the sensor data in three dimensions.

  Similar to the first embodiment, the second embodiment derives the vehicle heading using an approximate geometric pattern that is the best-fit pattern determined through multiple iterations. When data is obtained from the magnetic sensor, the data is first processed to determine whether the data point is stable. The stable data points are then processed to create a list of reference data points maintained in a reference list with the most recent reference data point at the top. As old reference points are revised, they are moved to the top of the list. Thus, the point that has been dropped is dropped out from the bottom of the stack, as described below. The reference points are used to determine the best-fit geometrical approximation pattern, and thus the calibration points used to determine the vehicle heading. The approximate geometry pattern is based on the X and Y values of the reference point, the assumed radius R, and the elliptic constant K used to allow the approximate geometry to be either circular or elliptical, Calculated after many iterations. Once the best-fit approximate pattern is determined (ie, the pattern that gives the lowest error metric value), the confidence level is determined using the error metric. This confidence level can be further determined based on the number of reference points in the list and the distance and / or time of the vehicle run. As described in detail below, this confidence level is then used to determine if the magnetic data point just read is stable enough to be used as a reference point for subsequent calibration of the compass system. Set the threshold to The confidence level can also be used to determine whether stable magnetic data points can be added to the reference list or averaged at points already in the list. Details of this process are described below with reference to the remaining accompanying drawings.

  FIG. 22A is a flow diagram of the main process for a specific example of the second embodiment. The main process 500 begins at step 502 where the hardware is initialized. This includes I / O, memory, magnetometer and display initialization. Next, at step 504, the processing circuit 110 of the compass circuit 100 (FIG. 3) determines whether the data stored in the non-volatile memory (NVM) 112 is valid. If not (as during the initial start-up of a new vehicle), the process proceeds to a decalibrate subroutine 506, described in detail below with reference to FIG. 22B. Otherwise, if there is valid data in the NVM 112, the processing circuit 110 determines whether enough reference points (eg, four or more reference points) are accumulated and stored for a good calibration solution. . If not, the calibration cancellation subroutine 506 is executed. Otherwise, the process proceeds to step 510 where event driven and background tasks are started. When the process returns from the decalibration routine 506, the process proceeds to step 510. Events and background tasks performed during step 510 include a magnetic field processing task (FIG. 22C), a stable point processing (FIG. 22D), and a display update task (FIG. 22E). These tasks are described in detail below.

  Following execution of the event driven and background tasks, the processing circuit 110 proceeds to step 512 to execute an idle task or routine. Such an idle task or routine, described below with reference to FIGS. 22F and 22G, includes the process of identifying the best fit geometric approximation pattern.

  Steps 510 and 512 are performed continuously as long as the vehicle ignition remains on. The process begins at step 502 when vehicle ignition is turned off. Accordingly, the calibration cancellation subroutine 506 is only executed following the operation of the vehicle ignition.

  As shown in FIG. 22B, the calibration cancellation subroutine 506 begins at step 520 where the number of learned reference points is set to zero. This allows the newest calibration solution to be based on fresh reference points. This is significant when the magnetic sensor is attached to the mirror housing. This is because a new driver may enter the vehicle between the ignition cycles and tilt or rotate the mirror housing. Another effect is to discard the transient data obtained when the vehicle moves along the assembly line during manufacturing.

  Next, at step 522, the processing circuit 110 sets the error metric to a high value. This reduces the threshold requirement for magnetic data points to qualify as reference points due to the low level of confidence, and thus a better list of reference points used to calibrate the compass system. This is done to ensure rapid population.

  In step 524, the processing circuit 110 sets the variables to default levels, for example, sets the assumed radius R and elliptical constant K to default values. In the preferred embodiment, the radius is set to 200 mG and the elliptical constant is set to K = 128. How to use these variables will be described in detail below.

  Information is then written to NVM 112 at step 526. Such information includes the current values of radius and elliptical constant, and can also include the number of data points and error metric values. Thereafter, the calibration cancellation routine 506 is completed, and the process returns to step 510 of the main process 500 shown in FIG. 22A.

As described above, one of the event driving and background tasks performed at step 510 of the main process 500 includes the magnetic field processing task 530 shown in FIG. 22C. This magnetic field processing task 530 is executed at a periodic sampling rate of, for example, 4 to 20 Hz. As shown, the first step in such processing is to read data from the magnetometer sensor 532. This raw data is then filtered and then the mean value and second derivative of the x and y values are calculated in step 534. More specifically, in this process, the filtered values of raw data x and y (x _f and y _f ) are calculated as follows:
x _f = (x + x _avg ) / 2
y _f = (y + y _avg ) / 2
However, x _avg and y _avg in the above two formulas are continuous weighted average values calculated before being calculated as follows.
x _avg = (x _f + x _avg * 6) / 7
y _avg = (y _f + y _avg * 6) / 7
The new first derivatives dx _new and dy _new are then calculated as follows:
dx _new = x _f −x _avg
dy _new = y _f -y _avg
Accordingly, the second derivatives d ² x and d ² y are calculated as follows:
d ² x = dx _new −dx _prev
d ² y = dy _new −dy _prev
Here, dx _prev and dy _prev are the first derivatives calculated previously. When the above calculation is complete, the values of dx _new and dy _new are stored in dx _prev and dy _prev respectively, and the average values x _avg and y _avg combine the new filtered data points using the above equation. To be adapted.

  Then, in step 536, the processing circuit 110 evaluates the confidence level of the existing reference data point set. As described above, the confidence level is determined as a function of the error metric, the number of data points in the reference data point set, and may be a function of the travel distance and / or travel time of the vehicle. More specifically, the processing circuit 110 determines that the reliability is greater when the number of reference points in the reference list is less than a specified number (eg, 4 points) or the average error is greater than a first threshold (eg, 8000). Set the sex level to low. Average error is a measure of how well the approximate geometry pattern fits the current data set and is described in detail below in connection with FIG. 22G. If neither of the two conditions for specifying a “low” confidence level is satisfied, the processing circuit 110 has a number of reference data points in the reference list within a predetermined range (eg, 5 to 6 points). Or whether the average error is greater than a second threshold (eg, 4000) but less than the first threshold. If either condition is true, the processing circuit sets the reliability level to “intermediate”. If none of the conditions for specifying a “low” or “intermediate” confidence level is true, the processing circuit sets the confidence level to “high”.

  Then, in step 538, based on the confidence level, thresholds are set for time delay, point spacing, determination of whether the raw data is noisy, motion, and heading stability. For example, if the confidence level is “low”, the noise threshold is set to 25 mG and the time delay is set to 2 seconds, and if the confidence level is “medium”, the noise threshold is If it is set to 15 mG and the time delay is set to 6 seconds and the confidence level is “high”, the noise threshold is set to 5 mG and the time delay is set to 8 seconds. The time delay is used to ensure that the data is stable enough to be used as a reference point, so the higher the level of confidence in the existing reference point and solution, it will replace the existing reference point. It becomes difficult. Furthermore, the lower the confidence level, the faster the system will acquire and accept a new reference point to calculate the solution with a higher confidence level. The threshold for determining whether the vehicle is moving may be a fixed value of 2 mG, for example, or may be variable as a function of the confidence level. The time delay is similar to that used in the embodiment described with reference to FIG. 13 when determining whether the data was below the noise threshold for a time period long enough to be considered stable. May be used. The threshold for point spacing based on the confidence level is determined as described below in connection with step 552 of FIG. 22D.

In step 540, the processing circuit 110 calculates a noise level. The noise level is calculated as follows.
Noise level = SQUAREROOT [(d ² x) ² + (d ² y) ² ]
Since the calculation of the square root function is a time consuming process for a low-cost processor, evaluating the square of the noise level can give a similar but less time consuming result. Alternatively, the noise level may be calculated using the root mean square or mean square error. For example, the noise level may be set equal to 1/2 log [(d ² x) ² + (d ² y) ² ].

  In step 542, processing circuit 110 determines whether the vehicle is moving based on the calculated noise level and the threshold determined in step 538. Thus, the vehicle is considered moving when the noise level exceeds a threshold (eg, 2 mG). The noise sign may vary for each vehicle model, and the noise threshold to be used in connection with the vehicle for which the compass is installed is installed after installation or if the vehicle model is known in advance. It may be preloaded in the factory before. Processing circuit 110 may perform known digital signal processing techniques to characterize when noise is due to vehicle movement. The processing circuit may also require the noise level to exceed a threshold for a predetermined time, such as 2 seconds. Then, at step 544, the processing circuit 110 determines whether the heading is stable based on whether the noise flag is set. The noise flag is initially set when the calculated noise level exceeds the established noise threshold, and until the noise level drops below the noise threshold and stays below that threshold for an established time delay , Kept in a set state.

  In step 546, the processing circuitry determines whether the vehicle is moving, whether the data is stable, and whether the data is free of noise. If each of these conditions is true, the processing circuit 110 executes a stable point processing subroutine 548. Otherwise, the magnetic field processing subroutine 530 ends and the controller returns to the main process 500. The noise level and average value are compared to a threshold to determine if the data is noisy. If noise or movement is detected, a countdown timer is initialized. Noise or loss of stability resets the counter as described above with reference to FIG.

  The stable point processing subroutine 548 is shown generally in FIG. 22D. As shown in FIG. 22D, the first step (step 550) of the subroutine is the closest and second closest to the list of reference points relative to the stable data points obtained in the magnetic field processing subroutine 530. Is to find the point. Then, in step 552, the processing circuit establishes a minimum spacing for the new data point based on the confidence level, the current radius, and the number of points included in the list of reference points. Conditions resulting from a large minimum spacing include a high confidence level, a large radius, and a large number of reference points in the list. For example, for the “low” confidence level, the minimum spacing is set to the current radius / 3 (typically about 67 mG), and for the “medium” confidence level, the minimum radius is the current radius / 2 (typically about 100 mG) and for the “high” confidence level, the minimum radius is set to the current radius x 3/4 (typically about 150 mG). Then, in step 554, the processing circuitry determines whether the spacing distance between the nearest point in the reference list and the new stable data point is greater than the minimum spacing established in step 552. If this distance is greater than the minimum interval, the process proceeds to step 556 and a new stable data point is added to the reference list as a new reference point. It is preferred to store the most recent reference point at the top of the list, but each reference point is shifted down by one position in the memory stack, and the last reference point is deleted if the memory stack overflows . Otherwise, each of the previous reference data points is maintained. Then, in step 558, a new data point is added to the top of the stack, and in step 560 the list of reference points is stored in the NVM and then returns to the magnetic field processing subroutine 530 and then to the main routine 500. Return.

  If the processing circuit determines in step 554 that the distance between the closest point and the new data point is less than or equal to the minimum spacing requirement, the processing circuit determines in step 562 that the closest data point is the reference. Determine if already at the top of the list of points. If so, the process proceeds to step 568. Otherwise, steps 564 and 566 are performed first. In step 564, the processing circuitry reconfigures the stack to move the closest reference point to the top of the stack. In step 566, the NVM flag is set to indicate that an NVM write is subsequently required.

  In step 568, the processing circuit determines whether the second closest point is too close to the new data point. This is defined as a function of the second closest data point within a fixed distance (eg, 75 mG) from the new data point, or a minimum interval (eg, the current value of the minimum interval) and / or radius. This can happen if you are within distance. One reason for not averaging such points is that the S-curve pulls the two points together, resulting in lost resolution. Thus, if the second closest point is too close, the process proceeds to step 572. Otherwise, if the second closest point is not too close, the processing circuit first executes step 570 and then proceeds to step 572. In step 570, the current data point is averaged with the closest point now at the top of the stack, and the average value is returned to the first position in the stack.

  In step 572, the processing circuitry determines whether the NVM flag has been set or whether the top point of the stack has been moved by the minimum milligauss number. Use the NVM flag to minimize the number of writes to the NVM and extend the life of the NVM. If either of these conditions is true, the processing circuit returns after writing to the NVM in step 560, or otherwise simply enters the main process routine 500 without writing to the NVM. Return. The minimum milligauss number to which the point is moved may be, for example, 30 milligauss.

  The process shown in FIG. 24D is performed each time a stable data point is acquired, so that the list of reference points is continuously updated with the most recent data. This ensures that the compass calibration is continuously updated using the most recent stable reference point.

  An additional event driven and background task performed at step 510 of the main process routine 500 is the display update subroutine 580 shown in FIG. 22E. This subroutine is preferably executed on a periodic basis, for example every 2 seconds, using a short time average of the data. More filtering may be performed using a long-time average.

  The display update subroutine 580 begins by determining at step 582 whether sufficient reference points have been obtained for a good solution. For example, if a sufficient reference point (for example, 4 points) for ensuring accurate calibration cannot be obtained, the displayed heading is not updated. If enough reference points are obtained, the process proceeds to step 584 and the processing circuitry determines whether the noise flag has been set. If the noise flag is set, the system does not update the heading display using the noise data. Thus, the display is not updated until the determination is made at step 584 that the received data is no longer noisy. Then, in step 586, a determination is made whether the average value is located too far from the approximate geometric pattern. If the average value is too far from the approximate geometric shape pattern, the heading is not updated. On the other hand, if the average value is sufficiently close to the approximate pattern, the process proceeds to step 588. Typical limits for the allowable distance from the circumference of the approximate geometric pattern are 0.5R and 1.5R. Alternatively, the error equation in the best fit ellipse search routine 604 may be used only at this point and its value compared to a predetermined threshold.

In step 588, the vehicle heading is calculated from the calibration points for the average of the most recent data. This is done by multiplying or dividing the ΔY value (y _avg −y _cal ) of the average data point by a value of K / 128, and then calculating the inverse tangent commonly known in the art. Is called. The newly calculated heading is then displayed at step 590 and the process then returns to the main process routine 500.

  The idle task executed at step 512 of the main process routine 500 is described below with reference to FIGS. 22F and 22G. More specifically, as shown in FIG. 22F, the first step 600 is to determine whether enough reference points are stored in the reference list for a good solution. Again, in the preferred embodiment, four data points are generally sufficient for a good solution. If there are not enough reference points, the process returns to the main process routine 500, which continues until sufficient reference points are obtained for a good solution. If enough reference points are obtained, step 602 is executed and important variables are copied to a temporary working copy. For example, this includes a working copy of the reference point list, the best fit geometry pattern radius, the pattern elliptic constant K, and the center calibration point of the pattern. Next, the best fit ellipse search routine 604 is executed. The best fit ellipse search routine 604 is described below with reference to FIG. 22G.

The best fit process 604 starts by setting the best error metric value to a high value such as 10,000 (step 606). An initial estimate is made for the calibration center point of the approximate circle by comparing the mean values X and Y of the reference points stored in the reference list (step 608). Initially, a default radius value of R = 200 mG is used (step 610) and a default elliptic constant of K = 128 is used (step 612). A preferred implementation of the second preferred embodiment defines an approximate geometric pattern using the following equation:
(X _avg −x _cal ) ² + (K / 128 (y _avg −y _cal )) ² = R ²

As will be apparent to those skilled in the art, the above equation is for a circle except for the value K / 128, which introduces an elliptic constant into the equation. By initially setting K to 128, the initial guess for the approximate geometry is for a circle with a radius of 200 mG and a center point corresponding to the average X and Y values of the reference points in the reference list. As shown in step 614, the center point (x _cal , y _cal ), radius R, and elliptic constant K yields the lowest error metric and thus the best fit against the reference points included in the reference list A thorough iterative search is performed for the elliptic constant giving As shown in step 616, an error value is determined for each reference point and added to determine an error metric. More specifically, the error of a particular reference point for the approximate geometry is calculated using the following equation:
Error = ABS [(xx _cal ) ² + (K / 128 ( _{yy cal} )) ² −R ² ]

In step 616, the processing circuitry flags which of the reference points has the highest error and the amount of this error is saved. Then, in step 620, an error metric is calculated by adding the errors from all reference points. Then, in step 622, the effects of the worst data point are removed by subtracting the error from the error sum. Then, in step 624, the processing circuitry determines whether the error sum thus calculated is lower than the best error sum obtained so far. If the error is lower than the previous best fit, the values of x _cal , y _cal , R and K are stored (step 626) and then go to step 628 where x _cal , y _cal , R And a determination is made whether all values of K have been tested. If not, the process proceeds to step 630 and loops back through steps 616-628 to attempt the next set of x _cal , y _cal , R and K values. This loop continues until all values are tested in step 628, in which case the loop is complete (step 632) and the sum of errors for each reference point is divided by the number of reference points in the reference list. Thus, an average error based on the best fitting algorithm is calculated. This value is used to determine the reliability level as described above. The process then returns to step 634 of idle routine 512.

In determining which value to try next in step 630, in the preferred embodiment, the value x _cal is first increased in 10 mG increments to a value of 400 mG higher than the average value of the x portion of the reference data point; Thereafter, the value of x _cal is decreased in 10 mG increments until a value of −400 mG is reached. Then, using the value of x _cal that has reached the best fit (ie, the lowest error metric), the value of y _cal is similarly increased and decreased. As these steps are performed, the value of x _cal is increased by 10 mG in 1 mG steps and then decreased by 10 mG in 1 mG steps around the best fit reference point found so far. The value of x _cal giving the lowest error metric is then stored and used, while y _cal is similarly increased and decreased in 1 mG increments. Thus, once the best values x _cal and y _cal are determined, the value of radius R is similarly increased and decreased to obtain the best radius fit. Similarly, the value of K is increased and decreased to adjust the ellipticity of the approximate pattern and get the best fit.

  In the preferred embodiment, the reference list is limited to 8 reference points. However, it is clear that the number of reference points included in the list can vary based on the desired resolution and the processing speed of the processing circuit. In general, the more reference points included in the reference list, the longer the processing time required to identify the best-fit approximate pattern. Furthermore, while the presently preferred best fit algorithm has been described above, other best fit algorithms that provide either faster and / or sounder approximations may be used.

  When the best-fit ellipse search subroutine 604 is completed, the processing circuit returns to step 634 (FIG. 22F) of the idle subroutine 512. In this step 634, the best-fit calibration point, radius, elliptical constant and error metric are copied back to the memory location used by the main processing routine and the vehicle heading is calculated and displayed using the new solution. The In step 636, it is determined whether the calibration point has been shifted by an amount sufficient to legally write new information to the non-volatile memory. If so, at step 638, after information is written to the NVM, processing returns to the main processing routine. Step 636 is performed to limit the number of writes to the NVM and thus extend the lifetime of the NVM.

FIG. 23 illustrates certain effects obtained by implementing the present invention. As shown in FIG. 23, data points AH represent reference data points obtained sequentially, and data point A is a reference data point obtained first. As described above, these reference points are stored in a list of eight or more reference points. Also, as described above, the best-fit approximate geometric pattern can be determined from about four data points. Thus, in the example shown in FIG. 23, the first four reference data points A-D yield a best-fit approximate pattern M having a center calibration point (x _cal , y _cal ).

  When reference data E is obtained, this reference data point E is added to the top of the reference list. However, since the distance from the approximate geographic pattern M to the reference point E is very large, the reference data point E is excluded from the reference data points used to repeatedly select or reselect the approximate pattern M in some cases. Also good. The determination of whether or not to exclude the reference data point may be made based on whether the error metric increases greater than a predetermined threshold or whether the calibration center point of the approximate pattern M moves greater than a predetermined distance.

  Subsequently, when the reference data point F is added to the reference list, the reference data point F may be excluded based on the error metric increase resulting from using points E and F. Similar results apply when the reference data point G is obtained. However, when reference data points H are obtained, points E-H constitute a sufficient number of points to ensure their own solution of the best-fit approximation pattern. Thus, at this point, points AD are excluded from the solution, or if there are two good approximation patterns for each set of points AD and EH, the two solutions And can then be used to determine vehicle heading. The best-fit geometric pattern N that best fits point E-H is used because point E-H is the most recently obtained of the eight reference points. When the situation as described above occurs, two separate sets of criteria lists are stored and updated separately. Thus, for example, if an additional reference data point that best matches the reference data point E-H is subsequently obtained, this new reference data point replaces the reference data point A in one of the two lists, but not necessarily It does not replace the other list. This allows two separate solutions to be stored simultaneously, but one of the two solutions that best fits the most recent reference data point is selected. Such a scenario is probably when a magnetic sensor is attached to the rearview mirror housing and two different drivers use the vehicle to adjust the mirror position to two different respective positions. This can also occur when there is a temporary magnetic disturbance of the vehicle causing a temporary shift in the position of the obtained data point. Thus, when a reference data point is obtained that more closely matches the first pattern of two approximate patterns (eg, pattern M), the system returns to the first approximate pattern to change the mirror housing position or vehicle magnetism. Adapt quickly to changing conditions.

  As described above, as the reliability level of the approximate solution increases, the filtering threshold for the new reference point increases correspondingly. Thus, it may take a long time to learn the last few reference points in the reference list, or to replace or change the reference points in the reference list. As a result, as described above with reference to FIG. 23, the response of the system to a sudden change in sensed vehicle magnetic state may not be as rapid as desired. By using two separate reference lists, one with a lower filtering threshold than the other, the system learns new reference points more quickly and is able to respond to sudden changes in the sensed vehicle magnetic state. Can be adjusted.

  When a compass sensor is placed near an electrically actuated device in a vehicle, such a device can create a disturbing magnetic field that can cause inaccuracies in the reading of the compass sensor. Some patents, such as US Pat. No. 4,953,305, address this problem by continuously recalibrating the compass system, but such recalibration compensates for the disturbing effects. Often it takes a long time to adjust and then it takes a long time to adjust it back when the vehicle accessory is no longer activated. U.S. Pat. Nos. 5,511,319 and 6,286,222 receive a signal over a wired connection that indicates when such a faulty device is activated, and the compass processing circuit A compass system is disclosed that allows the effects caused by a device to be compensated more quickly based on expected changes to the sensed magnetic field as a result of operation of the device. However, it is difficult to make such a prediction precisely, so the compass sensor reading may not be accurate when such a vehicle device is activated. For example, changes in load resistance, manufacturing variations, battery voltage variations, sensor sensitivity changes, and mechanical alignment changes can cause failures. Also, such a wired connection is not always possible or practical. For example, not all vehicle devices output a signal indicating when it is operating. In addition, vehicle equipment controllers such as sunroofs and convertible tops do not generate signals indicating the open or closed state of their sunroofs or convertible tops, but generate signals while they are opening or closing. Only if the sunroof or convertible top is opened and closed automatically. In particular, modern metal convertible hardtops cause significant changes in vehicle magnetism based on whether they are open or closed.

  One other type of vehicle device that has recently been added to vehicles and that can cause disturbances in the sensed magnetic field is the use of conductive windshields in the vehicle. Such a conductive glass is a transparent conductive material having sufficient resistance to generate heat when an electric current is passed to heat the windshield and thus clear the snow, ice or fog windshield. Includes many very thin wires or small strips of material. A typical on / off signal for such a conductive glass or electronic accessory is shown in FIG.

  The present invention uses a number of compensation means to overcome problems associated with inaccurate predictions of the effects of these electrical vehicle accessories on the sensed magnetic field. The first of such compensation means compensates by changing the control mechanism that operates these accessories in relation to timing the compass samples with respect to the on / off signals supplied to these vehicle accessories. The controller to perform. More specifically, as shown in FIG. 41, when the power supplied to a vehicle accessory that may cause a failure is in a low or off state, the compass takes samples at the regular sampling rate described above. However, when a vehicle accessory that can cause a failure is turned on, the power level of the activation signal increases. However, the present invention contemplates periodically pulsing the vehicle accessory between a high activation state and a low activation state and timing the compass sensor reading during the low power of the periodic activation signal. . To accomplish this, a signal giving the status of the vehicle accessory is sent to the compass processing circuit via either a simple wire or bus connection. The compass processing circuit 110 uses this signal to synchronize sampling by the compass sensor. The sample rate can vary between the on state and the off state of the accessory. If multiple accessories are used, the processing circuit may receive a separate signal for each accessory or process the signal to determine when all accessories are off or in a low power cycle. May be. The accessory pulse activations can be synchronized to each other to ensure a sufficient off-time interval for compass sensor sampling. Various signals used to activate the vehicle accessory may be transmitted to the compass processing circuit via the vehicle bus, local bus, dedicated wire (s) and / or wireless communication.

  Although the compass means has been described as used in connection with conductive windshields, this compensation method compensates for noise from other vehicle accessories such as fans, blowers, lights, windshield wipers, etc. May be used.

  Another form of compensation means for compensating for the action of the electric field produced by the conductive windshield is to change the direction of the current and the part of the windshield to be heated in a cyclical modification, or at least close to the compass sensor. A window controller that reduces the effect by configuring or patterning conductive strips in an area so that every other strip or group of strips conducts current in the opposite direction of its adjacent strips It is. In this way, two different reverse magnetic fields are generated in the windshield and zero out at least near the compass sensor. Similarly, one half of the windshield conductive strip is driven with the opposite polarity to the other half of the windshield conductive strip so that at least each windshield has a compass sensor placed on each side. The generated magnetic field can be zeroed. Such a split system can be electrically coupled in series to keep the number of electrical connections in two such connections. Yet another compensation means is a compensation coil provided on the window or mirror to generate a magnetic field that nulls that generated by the heated window. Yet another compensation means is a conductive or other shield coating that is selectively placed at or around the position of the compass sensor.

  Another compensation means that can be used to compensate for the magnetic field generated by heated windshields, defrosters, windshield wipers and other vehicle accessories is monitoring in the form of a digital signal processor (DSP) or modified compass processing circuit. A monitoring circuit that monitors the output of at least one compass sensor while searching for rises and falls in the level of the sensor output signal in a state consistent with that produced by the magnetic field generated by the vehicle accessory. The DSP or processing circuit preferably monitors the output of two or three compass sensors to search for the code corresponding to the noise source. The DSP or modified processing circuit can then ignore the sensor output signal when the vehicle accessory is operating or otherwise use predictive compensation. Such compensation is to offset the center of the calibration circle by a known amount, or use a different best-fit approximation pattern, as described above with reference to FIG. When the sign of the sensor output caused by the noise source disappears, the processing circuit can shift back to the first best-fit approximation pattern. The monitoring technique described above can eliminate the need for wiring to provide a vehicle accessory activation signal to the compass processing circuit.

  A DSP can also be used to watch the pulse width of the vehicle accessory drive signal superimposed on the sensor reading as a result of the magnetic field generated by the vehicle accessory. The DSP can then limit the analysis of the compass sensor readings when the pulse signal is at a low level.

  Yet another compensation means for compensating for vehicle accessory noise may simply identify the motion of the vehicle accessory or further identify the particular vehicle accessory that is operating, with all compass sensor output signals for each axis. It is a modified compass processing circuit that analyzes a code. The processing circuit can then select an appropriate compensation technique that best compensates for that particular vehicle. The code may be encoded in the activation signal such that a unique pulse sequence is generated when the vehicle accessory is activated. For example, one accessory may have a 90% duty cycle, while another accessory may have a 98% duty cycle. The compass system may also search for changes in battery voltage to determine whether the vehicle and / or vehicle accessory is operating.

  Yet another compensation means for compensating for the magnetic field generated by a vehicle accessory is a controller that drives the accessory with an AC signal and is disclosed in commonly assigned US Pat. No. 6,653,831. And a controller for compensating for noise generated by the AC power line. More specifically, this patent discloses a magnetometer that removes AC power line noise by taking a reading at twice the frequency of the AC power line. Thus, the reading is obtained at 120 Hz. Each two consecutive readings from each sensor are then averaged, which counteracts the effect of the magnetic field generated by the AC power line. Thus, in order to counteract the effects of noise generated by the vehicle accessory, readings are obtained from the sensor at twice the frequency of the signal used to drive or activate the vehicle accessory, and each two consecutive from each sensor. You can average readings.

  As mentioned above, it is desirable to be able to immediately and accurately compensate for vehicle magnetic changes or noise caused by vehicle accessories. This can be achieved by receiving a signal from a noise source via a wired or wireless connection. Such a wired connection may be a dedicated line or a local or vehicle scale bus. A suitable wireless connection such as infrared or RF, for example a BLUETOOTH® connection, can also be used. If a wired or wireless connection is not possible or impractical, the compass circuit monitors the output of the compass sensor for a known noise source magnetic code, and the geometric approximation pattern used by the compass system is known. Or it can be adjusted by a predictable amount to compensate for the identified noise. Such adjustments are preferably made without the need to perform a complete recalibration routine, but the current geometric approximation pattern is shifted by a predetermined offset or is used whenever noise occurs. It may be done by switching to two geometric approximation patterns. A compass system if the vehicle accessory changes state without generating a certain noise when in a certain state, but causes a change to the vehicle magnetism (ie sunroof or compatable top) Can monitor the output of the sensor for a shift in output level corresponding to what is known or expected to occur when the accessory's state changes. Again, the effect of accessory state changes on the output level of the compass sensor can be compensated by adjusting the geometric approximation pattern used by the compass system by a known or predictable amount. Using such a technique, the compass system can accurately and dynamically adjust for the opening and closing of the sunroof or the compatible top without the need for wired connections or external electrical signals. Furthermore, the compensation means of the compass system performs this compensation immediately without detecting re-calibration when detecting the effects caused by the vehicle accessory, and immediately returns to the calibration state when the effects of the vehicle accessory are removed. Can do. In addition, the action of the open sunroof or composable top is a sign in the output signal of the compass sensor, corresponding to the noise that occurs when the motor for opening and closing the sunroof or compassible top is activated. Can be predicted by monitoring. Upon signal deactivation, the compass processing circuit can immediately select the second geometric approximation pattern determined during calibration with the combustible top or sunroof open, and therefore Calibration can be performed immediately when the bable top or sunroof opens.

  The sign magnetic field effect monitored by the monitoring circuit may be scaled as a function of the vehicle battery voltage. Vehicle accessories such as conductive windshields generate sign noise that varies as a function of vehicle battery voltage. Thus, the monitoring circuit can also monitor the battery voltage, scale the magnetic field sign it monitors, and determine whether the vehicle accessory is operating.

  An electronic compass placed on a vehicle typically sets the geographic zone in which the vehicle is currently located, and the compass processing circuitry uses this location identification information to establish offset compensation for that particular geographic zone. You need to be able to do it. Such compensation is required to correct for true north and magnetic north pole offsets. Because the effect of this offset is different in different geographic regions around the world, conventional compass systems have been required to inform the geographic zone in which the vehicle is located.

  New car dealers can manually set the customer's geographic zone, but the customer may move or travel across the country and the vehicle may be driven in other zones. Thus, current compass systems require the user to know when they have crossed to a different geographic zone and remember to manually set that zone on the compass. Many users don't know how to initially set their geographic zone in the compass, or forget to do it when driving to a different zone, so that the user simply sets the geographic zone appropriately There are a number of warranty claims that can only occur if set.

  As described above, the GPS system can be used to identify the position of the vehicle, and the lookup table can be consulted to determine the geographic zone and the appropriate compass correction. However, GPS signals are not always available for compass systems. This is because the GPS antenna and receiver are not built into every vehicle, and the compass system may not be in a location where it can access the vehicle location identification signal. Accordingly, an aspect of the present invention is to increase the accuracy of the compass system when no zone is set. Furthermore, the present invention eliminates the need to set a zone or provide an accurate indication of the zone, as described in detail below.

  Most roads in North America are close to the grid on the east, west, south, and north sides. In other words, most roads are generally straight and extend east-west or north-south. Based on this fact, the compass system of the present invention determines the magnetic variance while traveling along the east / west or south / north road, and then uses this variance to compensate for the compass heading. Can be programmed to be used as a mechanism. The ability of the compass system to determine whether a vehicle is traveling along an east / west or south / north road can be verified based on whether a turn of approximately 90 ° was last made. All geographic zones of the US continent fall within a nominal range of about ± 20 ° within 48 states of the US continent. Thus, if the vehicle is approximately east / west / south / north within the system margin (± 20 °) and a turn of about 90 ° is made, the difference between the heading and the primary heading is Use to shift the variance calculation slowly. Reasonable results can be obtained with sufficient averaging.

  In systems with moving sensors, such as systems where the compass sensor is located in the rearview mirror housing, the geographic zone and mirror aim results can be used to compensate not only for the zone but also for the mirror position. it can. Instead of ± 20 °, the zone plus the expected mirror angle and variation sum may be used.

  Further, if the calibration point is reliable and the vehicle travels in a relatively constant direction for a long time (ie, several hours), the vehicle will be on a long journey and may traverse multiple magnetic change zones. Is reasonable. During such long itineraries, small shifts of variance may be added or subtracted at regular intervals such as 0.5 ° per hour. In the end zone, south / north travel causes variations that can be compensated using similar techniques.

  In the calibration method described above, the four points stored in the RAM defining the stability points are probably 90 ° apart. These four point rotations can be used for mirror rotation and zone correction.

  One advantage of using the technique described above is that the current geographic zones are each about 4.2 °. Thus, current techniques for manually selecting a geographic zone produce only about 4 ° error in the zone. However, the present invention compensates for small variances within each geographic zone and thus can provide more accurate compensation.

  In order to support a statistical analysis of the normal direction of vehicle travel, three plots showing statistical information of the frequency with which the vehicle travels in a direction are shown. FIG. 42A is a plot of the number of seconds a vehicle was measured for travel with various headings. FIG. 42B is a plot of the number of samples or occurrences at various headings when it is required to perform a 90 ° turn before accepting the sample as a filtering technique. Obviously, this removes a large number of samples that exist between 90 ° intervals. FIG. 42C shows the data with further filtering so that all points greater than ± 20 ° from the dominant peak are removed. Clearly, therefore, vehicle travel has four main peaks corresponding to east, west, south, and north. Thus, the dispersion of these peaks from 0 °, 90 °, 180 ° and 270 ° can be measured and used to display the heading after compensating the compass reading.

  Another technique for determining the appropriate compensation for the geographic zone in which the vehicle is located is to provide an improved display graphic showing the selected geographic zone. This allows the user to visually confirm the selected zone. However, because there are so many zones, it is difficult to accurately illuminate the selected zone and it is difficult for the driver to determine exactly which zone is located. To assist the automatic zone detection method described above, a graphic display may be provided that shows the time zone for a particular geographic region (ie, country, continent, etc.). An example of such a display is shown in FIG. As shown, four time zones are shown on a graphic map of the United States, one of these time zones being illuminated or unilluminated, or otherwise illuminated differently from the other zones, Indicates the time zone where the vehicle is currently located. This gives accurate feedback to the user so that the appropriate time zone in which the vehicle is located can be properly selected. Such a display and selection mechanism is significantly more intuitive than the current 15 different geographic magnetic zones used for North America. This is because the time zone is much easier for the driver to understand than the geographical magnetic zone. You may also use automatic sources of time zones, such as RDS radio (CT-clock time) and / or reference (NIST (WWVB, WWV)) or locally generated time reference, Usually adjusted manually by the user for the time zone or automatically using GPS. Such a clock may be implemented on the display in a rearview mirror or other vehicle accessory in which the compass is located, or may be provided at a remote location in the vehicle and timed to the compass system via a bus or individual wire. Zones may be sent. In a preferred form, the graphic display may be configured as disclosed in US Pat. No. 6,356,376, the entire disclosure of which is incorporated herein by reference. Alternatively, the display may display the time zone in an alphanumeric format rather than a graphic display.

  Although a reduced number of zones are displayed to the user, good accuracy is still obtained when combined with the automatic zone detection method described above. The result of the zone calculation is clipped with a reasonable configuration around the selected zone (eg, half the user interface zone size). This protects against odd traffic patterns that can reduce the accuracy of previous methods. Adding a small weight to the vehicle heading samples obtained before and after the first and / or last detected heading change during the ignition cycle gives an additional improvement when memorizing the historical samples and analyzing the stored historical samples Can be obtained. This is because these vehicle heading samples may occur when the vehicle is on a roadway off the street that is not visible on the board and is not frequently seen. Another way to protect against odd traffic patterns is to average the sensor output over time.

  Although it is preferred to calculate the noise level using the second derivative, a noise level can be reduced using a digital filter separately or in addition thereto. In addition, it can benefit from factory calibration and gain (usually in mG) presets.

  The present invention has been described as incorporating an electronic compass system within the mirror housing of a rearview mirror assembly. It will be apparent to those skilled in the art that various other vehicle accessories and components can be incorporated into the rearview mirror assembly, in whole or in part, and in various combinations. Such vehicle accessories and components can be attached to, on or against the mirror housing, mirror mount, mirror mount or attachment to the housing, or in a console or other housing associated with the rearview mirror assembly. It can also be attached to. Furthermore, such vehicle accessories may share components such as processors, sensors, power supplies, wire harnesses and plugs, displays, switches, antennas and the like. Examples of other vehicle accessories, components or features are described in detail below.

  44A-C illustrate another embodiment of a rearview mirror assembly 900a that incorporates any of the electronic compass systems described above. As shown in FIGS. 44A-44C, the mirror assembly 900a includes a bezel 902 and a case 904. The bezel and case are coupled to define a mirror housing for incorporating features in addition to the reflective element 901 and the information displays 905a and 905b. Commonly assigned US Pat. Nos. 6,102,546, D410,607, 6,407,468, 6,420,800, and 6,471,362, incorporated herein by reference. Describes examples of various bezels, cases, and associated button structures that can be used with the present invention.

As shown in FIGS. 44A-44C, the mirror assembly can include first and second microphones 910a and 910b. Examples of microphones for use in the present invention are commonly assigned US patent application Ser. No. 09 / 444,176, US Pat. No. 6,614,911, US Patent Application Publication No. US 2002, incorporated herein by reference. No. 0110256A1 and PCT Application No. WO 03/041285 . Although two microphones are shown attached to the back of the mirror case 904, one or more such microphones can be attached to the top of the mirror assembly (as shown in FIGS. 45A and 45B). It can also be mounted at the bottom of the mirror assembly or anywhere in the mirror case or bezel. The two microphones 910a and 910b are preferably incorporated into recessed portions 912a and 912b on the back of the mirror case, one near each end of the mirror assembly. As shown in FIG. 44A, the microphone is configured such that the acoustic dam 914 extends around a transducer 916 in the microphone housing 918. Additional details of this preferred configuration are disclosed in commonly assigned PCT application No. WO 03/041285 , incorporated herein by reference. Audio systems including microphones may be at least partially integrated with the information display and / or share components with the information display in a common controller. Furthermore, the status of these systems and / or the devices controlled thereby may be displayed on the relevant information display.

As shown in FIGS. 45A and 45B, a single microphone 910 is provided on the upper surface of the mirror assembly 900b. In this configuration, the microphone housing 918 preferably includes two transducers similar to those disclosed in the PCT application WO03 / 041285 and US Patent Application Publication No. US2002 / 0110256A1.

The mirror assembly 900 can include first and second illumination assemblies 920a and 920b. Various lighting assemblies and lighting devices for use in the present invention are commonly assigned U.S. Pat. Nos. 5,803,579, 6,335,548, 6,441, incorporated herein by reference. , 943, 6,521,916, 6,523,976, 6,670,207 and 6,805,474, and commonly assigned US patent application Ser. No. 09/723, No. 675 . Each lighting assembly preferably comprises a reflector, a lens and a lighting device (not shown). There may be two lighting assemblies, one generally arranged to illuminate the front passenger seat area and a second one typically arranged to illuminate the driver seat area. Alternatively, there may be only one lighting assembly that illuminates both seating areas and / or there may be additional lighting assemblies that illuminate the central console area, overhead console area, or the area between the front seats. .

  The mirror assembly 900 may also include first and second switches 922a and 922b. Suitable switches for use in the present invention are commonly assigned US Pat. Nos. 6,407,468, 6,420,800, 6,426,568, which are incorporated herein by reference. This is described in detail in US Pat. Nos. 6,614,579 and 6,471,362. These switches include lighting assembly, display, mirror reflectivity, voice activation system, compass system, telephone system, highway toll gate interface, telemetry system, headlight controller, rain sensor, tire pressure monitoring system, navigation system, lane departure It can be incorporated to control alarm systems, adaptive cruise control systems and the like. Other displays or systems described herein or described in references incorporated by reference may be incorporated at any location within the associated vehicle and may be controlled using switches.

The mirror assembly 900 can also include first and second indicators 924a and 924b. Various indicators for use in the present invention are described in US Pat. Nos. 5,803,579, 6,335,548, 6,441,943, 6,521,916, incorporated herein by reference. No. 6,523,976, 6,670,207 and 6,805,474, and commonly assigned US patent application Ser. No. 09 / 723,675 . These indicators can indicate the status of the display, mirror reflectivity, voice activated system, compass system, telephone system, highway toll interface, telemetry system, headlight controller, rain sensor, security system, and the like. Other displays or systems described herein or described in references incorporated by reference may be incorporated at any location within the associated vehicle and may have a state indicated by an indicator.

The mirror assembly 900 can further include first and second light sensors 926 and 928 that act as glare and ambient light sensors, respectively. Preferred optical sensors for use in the present invention are commonly assigned US Pat. Nos. 5,923,027, 6,313,457, 6,359,274, incorporated herein by reference. Nos . 6,379,013, 6,402,328, 6,679,608, and 6,831,268 are described in detail. The glare sensor 926 and / or the ambient light sensor 928 automatically controls the reflectance of the self-dimming reflective element and the brightness of the information display and / or the backlight. The glare sensor 926 is also used to sense trailer vehicle headlights, and the ambient light sensor is used to detect ambient light conditions when the system is operating. In another embodiment, a sky sensor 930 is generally incorporated and arranged to detect light levels above and in front of the associated vehicle. The sky sensor 930 can be used to automatically control the reflectivity of the self-dimming element, the external light of the controlled vehicle, and / or the brightness of the information display. The mirror assembly may further include a sun-load sensor for sensing the light level toward the vehicle driver and passenger sides of the vehicle to control the vehicle climate control system.

  In addition, the mirror assembly 900 may include first, second, third, fourth and fifth operator interfaces 932a-932e disposed on the mirror bezel 902. Each operator interface is shown to include a backlight information display “A”, “AB”, “A1”, “C”, and “12”. It will be appreciated that these operator interfaces can be incorporated somewhere in the associated vehicle, for example, in mirror cases, accessory modules, instrument panels, overhead consoles, dashboards, seats, center consoles, and the like. Suitable switch structures are commonly assigned US Pat. Nos. 6,407,468, 6,420,800, 6,426,568, 6,614,579, incorporated herein by reference. And No. 6,471,362. These operator interfaces include lighting assemblies, displays, mirror reflectivity, voice activation systems, compass systems, telephone systems, highway toll gate interfaces, telemetry systems, headlight controllers, rainfall sensors, tire pressure monitoring systems, navigation systems, lanes A departure warning system, an adaptive cruise control system, etc. can be controlled. Other displays or systems described herein or described in references incorporated by reference may be incorporated at any location within the associated vehicle and controlled using the operator interface (s). May be. For example, the user may program the display (s) to show predetermined information, may program the display (s) to scroll through a series of information, Alternatively, a set point related to a certain operation device may be input together with a related sensor input to display certain information when a predetermined event occurs. For example, in one embodiment, a given display is unlit until the engine temperature is above the threshold, and then the display is automatically set to display the engine temperature. As another example, a proximity sensor located at the rear of the vehicle can be connected to the controller and coupled to a rearview mirror display to indicate to the driver the distance to the object, i.e. the display It may be configured as a bar having a length proportional to a given distance.

  Specific locations and numbers of these additional features are shown in FIGS. 44A-44C and FIGS. 45A, 45B, but fewer or more individual devices may be used as shown in the references incorporated herein. It will be appreciated that it can be incorporated at any location within the associated vehicle.

  A mirror mount 903 is included for attaching the mirror assembly to the windshield 898 or the vehicle roof structure within the vehicle. Numerous accessories such as a rain sensor, camera, head ride controller, additional microprocessor, additional information display, compass sensor, etc. may be incorporated into the mount 903 or the housing 952 attached to the mount 903. It will be understood. These systems may be at least partially integrated with the information display and / or share components with the information display in a common controller. In addition, the status of these systems and / or devices controlled thereby may be displayed on its associated information display.

  The mirror assembly 900 is shown in FIG. 44A to further include a circuit board 960 to which a compass sensor module (not shown) is attached and a daughter board 962 with an input / output bus interface (not shown). .

  The electrical output signal from either or both sensors 926 and 928 is used as an input to a controller (not shown) to reflect the reflectance of reflective element 901 and / or any one of displays 905a and 905b or It is also possible to control the overall brightness. Details of various control circuits for use therewith are commonly assigned U.S. Pat. Nos. 5,883,605, 5,956,012, 6,084,700, incorporated herein by reference. 6,222,177, 6,224,716, 6,247,819, 6,249,369, 6,392,783, and 6,402,328 Explained. These systems may be at least partially integrated with the information display and / or share components with the information display in a common controller. In addition, the status of these systems and / or devices controlled thereby may be displayed on its associated information display.

  Although the compass sensor module has been described as being attached to the circuit board 960, the sensor module may be disposed in the mount 903, and the accessory module 952 may be located anywhere near the mirror assembly 900 or in the associated vehicle. It will be appreciated that, for example, it may be located under the dashboard, in the overhead console, in the center console, in the trunk, in the engine compartment, etc. The compass system described above may be at least partially integrated with the information display and / or share components with the information display in a common controller. In addition, the status of these systems and / or devices controlled thereby may be displayed on its associated information display.

  The circuit board 960 may include a controller (not shown) such as a microprocessor, and the daughter board 962 may include an information display 905a. The microprocessor, for example, receives the signal (s) from the compass sensor module, processes the signals, and sends the signals to the daughter board to control the display 905a and corresponding vehicle headings. Instruct. As described herein and described in the references incorporated herein, the controller may include an optical sensor (s), a rain sensor (s) (not shown), an automatic vehicle exterior Light controller (s) (not shown), microphone (s), global positioning system (not shown), telecommunications system (not shown), operator interface (s), and The information display (s) can be controlled to receive signal (s) from a number of other devices and provide appropriate visual indications.

  The controller (s) used to control the compass system are mirror reflectivity, external light, rain sensor, compass, information display, windshield wiper, heater, defroster, defogger, air conditioner, telephone system Navigation systems, security systems, tire pressure monitoring systems, garage door opening transmitters, remote keyless entry, telemetry systems, voice recognition systems such as digital signal processor based voice activation systems, and vehicle speed, at least in part Can be controlled. Controller 796 (s) receives signals from switches and / or sensors associated with any of the devices described herein and described in the references incorporated herein and described herein or Other devices described in the references incorporated herein can be automatically operated. The controller may be located at least in part outside the mirror assembly, or may include a second controller somewhere in the vehicle or additional controllers throughout the vehicle. The individual processors can be configured to communicate serially, in parallel, via the BLUETOOTH® protocol, via wireless communication, via the vehicle bus, via the CAN bus, or other suitable communication. A multi-pin connector interface 964 can be provided for such external connections.

Commonly assigned US Pat. Nos. 5,990,469, 6,008,486, 6,130,421, 6,130,448, 6,255, incorporated herein by reference. 639, 6,049,171, 5,837,994, 6,403,942, 6,281,632, 6,291,812, 6,469,739 , No. 6,465,963, No. 6,429,594, No. 6,587,573, No. 6,611,610, No. 6,621,616, No. 6,653,614, No. The external light control systems described in US Pat. Nos. 6,379,013, 6,861,809, and 6,774,988, and US Patent Application Publication Nos. US2004-0143380A1 and US2004-0008410A1 , It can be incorporated on the basis of the invention. These systems may be at least partially integrated with the information display and / or share components with the information display in a common controller. Furthermore, the status of these systems and / or the devices controlled thereby may be displayed on the relevant information display. As disclosed in US Pat. No. 6,587,573, both the compass sensor and imaging sensor array 950 can be housed in an accessory housing 952 attached to a mount 903.

Moisture sensors and windshield fog detection systems are commonly assigned U.S. Pat. Nos. 5,923,027, 6,617,564, 6,313,457, and 6, which are incorporated herein by reference . 681,163 . These systems may be at least partially integrated with the information display and / or share components with the information display in a common controller. Furthermore, the status of these systems and / or the devices controlled thereby may be displayed on the relevant information display.

  Commonly assigned US Pat. No. 6,262,831, incorporated herein by reference, describes a power source for use in the present invention. These systems may be at least partially integrated with the information display and / or share components with the information display in a common controller. Furthermore, the status of these systems and / or the devices controlled thereby may be displayed on the relevant information display.

The mirror assembly can further include one or more antennas 940 for receiving and / or transmitting RF signals. The mirror assembly may further include appropriate receiving, transmitting, and / or processing circuitry. Such antennas are used in cellular telephone systems, BLUETOOTH® transmission / reception systems, remote keyless entry (RKE) systems, trainable garage door opener systems, tire pressure monitoring systems, global positioning satellite systems, LORAN systems, etc. Can be used. Some of these systems can share a common antenna and receive, transmit, process and display circuitry accordingly. Examples of tire pressure monitoring systems incorporated into rearview mirror assemblies are commonly assigned U.S. Pat. Nos. 6,215,389 , 6,431,712, and 6,861,942, incorporated herein by reference . And 6,596,935 . Examples of GPS systems incorporated into the rearview mirror assembly are commonly assigned US Pat. Nos. 6,166,698, 6,297,781, and 6,396,446, incorporated herein by reference. And US Patent Application Publication No. US2002 / 0032510A1 . An example of a LORAN system incorporated into a rearview mirror assembly is disclosed in commonly assigned US Pat. No. 6,539,306, incorporated herein by reference. Examples of both telephone / telematic systems and BLUETOOTH® systems incorporated into a rearview mirror assembly are disclosed in commonly assigned US Patent Application Publication No. US2002 / 0032510A1, which is incorporated herein by reference. An example of a trainable garage door opening system and RKE system incorporated into a rearview mirror assembly is disclosed in US Pat. No. 6,091,343, incorporated herein by reference.

  The mirror may further include an infrared (IR) transmitter / receiver for transmitting / receiving information to / from the mirror assembly and possibly to the vehicle. An example of such a rearview mirror assembly is disclosed in commonly assigned US Pat. No. 6,407,712, incorporated herein by reference.

Further, the mirror assembly can include one or more of the same or different types of displays. Different types of displays include, for example, vacuum fluorescence, LCD, reverse LCD, LED, organic LED, dot matrix, backlight indication, and the like. For displays intended to display a significant amount of information simultaneously, the display disclosed in commonly assigned US Pat. No. 6,346,698, incorporated herein by reference, can be used. Examples of backlight indicating panel displays are commonly assigned U.S. Pat. Nos. 6,170,956, 6,572,233, 6,356,376, and 6, 870,655 . Various displays used in the rearview mirror are disclosed in commonly assigned US Pat. No. 6,356,376, and US Patent Application Publication No. US 2002/0154379 A1, incorporated herein by reference.

  Vehicle accessory wiring in the rearview mirror assembly housing can extend through the mounting bracket and along the windshield under the channel mount (if the mounting bracket is not previously extended to the headliner). An example of a rearview mirror assembly in which accessory wiring in a mirror assembly housing is routed through a mounting bracket is disclosed in commonly assigned US Pat. No. 6,467,919, incorporated herein by reference.

  Although the present invention has been described as being implemented with a sensor disposed within the housing of the rearview mirror assembly, the sensor can also be mounted at the mounting foot or other location of the rearview mirror assembly. Furthermore, any or all of the various components of the electronic compass of the present invention can be mounted anywhere on the vehicle. Certain embodiments of the present invention are novel and include land vehicles (ie, automobiles, trucks, sports utility vehicles (SUVs), trains, motorcycles, bicycles, motorbikes, scooters, snowmobiles, all-terrain vehicles (ATVs). It will be further apparent that it is useful for vehicles such as), military vehicles) and other vehicles such as aircraft, ships and amphibious vehicles.

  Although the present invention has been described as using a process of plotting signals output from a magnetic sensing circuit with reference to each other in a two-dimensional or three-dimensional coordinate system, the signals are processed and analyzed separately, A similar solution may be used in which analysis results are compared and similar results are achieved.

  The above description is considered that of the preferred embodiments only. Various modifications of the invention will be apparent to those skilled in the art. Therefore, the embodiments shown and described in the accompanying drawings are merely illustrative of the present invention and do not limit the scope of the present invention defined in the claims.

It is an electric circuit diagram of the block form of the conventional electronic compass circuit. It is a graph which shows the calibration technique used by the conventional electronic compass system. It is an electric circuit diagram of the block form of the electronic compass circuit comprised by this invention. 1 is a front view of a rearview mirror assembly incorporating at least a portion of an electronic compass circuit of the present invention. FIG. 5 is a partially cutaway side view of a rearview mirror assembly with a magnetic sensing element mounted in a mirror housing. It is a partially broken side view of a rearview mirror assembly in which a magnetic sensing element is attached to an attachment foot of a rearview mirror assembly attachment structure. It is a top view which shows the upper part of a rear-view mirror assembly with which the sensing element was attached in the mirror housing. It is a graph of the local sphere with respect to the world sphere sensed by the compass. Fig. 4 is a graph of a local sphere sensed by a compass for various coordinate axes. Figure 5 is a graph of a sample data set plotted against a local sphere sensed by a compass. It is a state diagram which shows the various operation states of the compass processing circuit based on 1st Embodiment, and those mutual relationship. It is a flowchart of the compass flow control routine performed by the processing circuit of the electronic compass of the first embodiment of the present invention. It is a flowchart of the compass flow control routine performed by the processing circuit of the electronic compass of the first embodiment of the present invention. It is a flowchart of the compass flow control routine performed by the processing circuit of the electronic compass of the first embodiment of the present invention. It is a flowchart of the compass flow control routine performed by the processing circuit of the electronic compass of the first embodiment of the present invention. 12 is a flowchart of a noise analysis subroutine called during the compass flow control routine of FIGS. 12A-12D. 12 is a flowchart of a first point set building subroutine called during the compass flow control routine of FIGS. 12A-12D. FIG. 12 is a flowchart of a best fit center point adjustment subroutine called during the compass flow control routine of FIGS. 12A-12D. 12 is a flowchart of a best fit radius adjustment subroutine called during the compass flow control routine of FIGS. 12A-12D. 12 is a flowchart of a heading calculation subroutine called during the compass flow control routine of FIGS. 12A-12D. 12 is a flowchart of an angle bucket update subroutine called during the compass flow control routine of FIGS. 12A-12D. 12 is a flowchart of a second point set building subroutine called during the compass flow control routine of FIGS. 12A-12D. FIG. 12 is a flowchart of a local sphere fit calculation subroutine called during the compass flow control routine of FIGS. 12A-12D. 12 is a flowchart of a point set reset subroutine called during the compass flow control routine of FIGS. 12A-12D. It is a flowchart of the compass flow routine performed by the processing circuit of the electronic compass of 2nd Embodiment of this invention. It is a flowchart of the compass flow routine performed by the processing circuit of the electronic compass of 2nd Embodiment of this invention. It is a flowchart of the compass flow routine performed by the processing circuit of the electronic compass of 2nd Embodiment of this invention. It is a flowchart of the compass flow routine performed by the processing circuit of the electronic compass of 2nd Embodiment of this invention. It is a flowchart of the compass flow routine performed by the processing circuit of the electronic compass of 2nd Embodiment of this invention. It is a flowchart of the compass flow routine performed by the processing circuit of the electronic compass of 2nd Embodiment of this invention. It is a flowchart of the compass flow routine performed by the processing circuit of the electronic compass of 2nd Embodiment of this invention. It is the graph which plotted the compass sensor data and the approximate geometric shape relevant to it. It is a top view which shows a part of circuit board to which the compass sensor was attached. FIG. 24B is a side view of the circuit board shown in FIG. 24A. FIG. 24B is a side view of the circuit board shown in FIG. 24A with the portion 150a bent upward. FIG. 24B is a side view of the circuit board shown in FIG. 24A in which a retainer is disposed around a bent portion of the circuit board. It is a partially broken side view of a rearview mirror assembly in which a magnetic sensing element is integrated in a common integrated sensor package mounted in a mirror housing. FIG. 26 is an enlarged view of the integrated sensor package of FIG. 25. 27 is a perspective view of the integrated sensor package shown in FIGS. 25 and 26. FIG. 6 is a front view of an integrated sensor package according to another embodiment of the present invention. FIG. FIG. 29 is a top view of the integrated sensor package shown in FIG. 28. 1 is an exploded perspective view of a mother / daughter combined circuit board assembly according to an embodiment of the present invention; FIG. It is a block diagram of the noise analysis function by this invention. It is a block diagram which shows the noise analysis function shown in FIG. 31 in detail. noiseFactor = 10, delayIntercept = -3, and is a graph showing the relationship between the DX2 ² + DY2 ² + DZ2 defined noise signal as the square root of ^two and the delay signal when maxDelay = a 32. noiseFactor = 10, delayIntercept = -4, and a graph showing the relationship between the DX2 ² + DY2 ² + DZ2 defined noise signal as the square root of ^two and the delay signal when maxDelay = a 32. Fig. 6 is a circular plot divided into 12 equally sized angle buckets. An example plot of a compass reading resulting from a vehicle mirror housing shift while located in Zealand, Michigan, USA. An example plot of a compass reading resulting from a vehicle mirror housing shift while located in Taipei, Taiwan. The information known to the compass processing circuit is plotted before and after the mirror housing shift occurs. FIG. 6 is a circuit diagram illustrating a compass system according to another embodiment of the present invention. FIG. 2 is a signal timing diagram of a typical on / off activation signal used in a typical electrical device used in a vehicle. FIG. 5 is a signal timing diagram of a novel on / off activation signal used in an electrical device for use in a vehicle and supplied to the compass circuit of the present invention. It is a plot of the number of seconds a vehicle has been measured for running at various headings. A plot of the number of samples or occurrences at various headings when it is required to perform a 90 ° turn before accepting the sample as a filtering technique. A plot of the number of samples or occurrences at various headings when further filtering is used so that all points greater than ± 20 ° from the dominant peak are removed. It is a front view which illustrates the graphic compass display comprised by this invention. It is a side view of the rearview mirror assembly comprised by this invention. It is a rear view of the rearview mirror assembly comprised by this invention. It is a top view of the rearview mirror assembly comprised by this invention. It is a front view of the rearview mirror assembly comprised by this invention. It is a top view of the rearview mirror assembly comprised by this invention.

Claims (7)

In an electronic compass system for vehicles,
A magnetic sensor circuit for sensing at least two components of the geomagnetic field vector and generating an output signal representative of the sensed at least two components;
A processing circuit coupled to the magnetic sensor circuit, receiving an output signal from the magnetic sensor circuit, determining a relative strength of a geomagnetic field vector, and noise in the output signal received from the magnetic sensor circuit Determining whether a threshold noise level set as a function of the relative strength of the magnetic field vector is exceeded, and if the noise in the output signal is lower than the threshold noise level, the vehicle as a function of the sensed component And a processing circuit for generating a heading signal representing the calculated heading or previous heading if the noise exceeds the threshold noise level;
Electronic compass system with   The processing circuit determines whether noise in the output signal received from the magnetic sensor circuit exceeds the threshold noise level when a change in the output signal exceeds the threshold noise level. Electronic compass system described in 1.   3. The electronic compass system according to claim 2, wherein the processing circuit does not update data used for calibration of the compass system when the noise exceeds the threshold noise level. The processing circuit includes:
Update the heading signal only when the change in the output signal does not exceed the threshold noise level, and when a specified time period has elapsed since the last change in the output signal exceeded the threshold noise level. Only update the data used to calibrate the compass system,
The electronic compass system according to claim 3.   The electronic compass system of claim 4, wherein a time period that must elapse before the data used for calibration of the compass system can be updated is set as a function of the relative strength of the geomagnetic field vector.   The electronic compass system according to claim 1, wherein the processing circuit further updates data used for calibration of the compass system only when the noise does not exceed the threshold noise level.   The electronic compass system according to claim 6, wherein the processing circuit updates data used for calibration of the compass system only when a change in the output signal does not exceed the threshold noise level.

Images