JPH0762687B2 - Vehicle speed detector - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a vehicle speed detection device, and more particularly to a speed detection device capable of highly accurate speed or acceleration detection.

[Prior Art] For a vehicle, the speed v and the acceleration α are extremely important factors for objectively grasping the running state of the vehicle.
The speed detection device of each vehicle is required to accurately detect the speed and acceleration.

In particular, in today's vehicles with advanced electronics, the detected speed v and the acceleration α are widely used for various applications such as an automatic door lock, a power steering, a digital speed controller, an engine ignition timing adjustment, and a navigation controller. Accurate detection of these data is extremely important for improving drivability and the vehicle interior environment.

In particular, in vehicles equipped with an electronic skid controller system (a system that prevents the wheels from locking and rolling sideways during sudden braking),
The speed detecting device is used to independently detect the vehicle speed and acceleration of each wheel of the vehicle, that is, the total of four wheels in the front, rear, left, and right. The detected vehicle speed / acceleration and its subtle changes are compared with the ideal control conditions stored in the computer, and the braking force of each wheel is controlled. Therefore, real-time measurement of vehicle speed / acceleration is extremely accurate. Required to do.

However, the conventional vehicle speed detection device has a high accuracy required by the electronic skid control system (hereinafter referred to as an ESC system), and has a wide range of vehicle speed and acceleration from low speed to high speed (for example, 2.5 Km / h to 250 Km / h). Could not always be detected accurately.

The reason will be described in detail below.

Speed Detection FIG. 14 shows a preferred example of a vehicle speed detecting device used in an ESC system, in which a rotating object 10 is provided on the rotating side of a wheel or a transmission, and the object 10 is detected. A sensor 12 is provided on the vehicle fixing portion side so as to face the vehicle.

The detected body 10 and the sensor 12 may be formed so as to magnetically detect the rotation of the detected body 10 or may be formed to optically detect the rotation.

In the device shown in the figure, the detected object 10 is formed as a timing gear made of a magnetic material, and a large number of timing teeth 10a are projected on the surface thereof at equal intervals.

The sensor 12 is formed by using a permanent magnet, a coil, and the like, and detects the rotation of the timing gear 10 as a change in magnetic resistance.

Therefore, when the wheels rotate and the timing gear 10 rotates, the AC signal Y1 is output from the magnetic sensor 12 to the pulse generation circuit 14 every time each timing tooth 10a passes.

Then, the pulse generation circuit 14 receives the input AC signal Y1
Is converted into a speed detection pulse signal P corresponding to the passage of the timing tooth 10a and output to the speed calculation circuit 16.

Such a pulse signal P is proportional to the rotation speed of the timing gear 10, and the output time interval T thereof increases as the rotation speed increases.
Is shorter and the output time interval T becomes longer as the rotation speed decreases.

Then, the speed calculation circuit 16 calculates and outputs the vehicle speed v and the acceleration α based on the output time interval of the pulse signal P.

By the way, in the device for detecting the vehicle speed v and the acceleration α based on the output time interval of the pulse signal P as described above, the pulse signal P is generated from the pulse generation circuit 14 at the timing when the timing tooth 10a passes in front of the sensor 12. Is required to be output at an accurate phase (timing).

However, the pulse generator circuit 14 that has been widely used
Uses the zero-cross method shown in Fig. 15, and the AC signal
In the section where Y1 changes from the peak value on the mountain side to the peak value on the valley side, when the signal Y1 crosses the zero point, the pulse signal P
Is output.

Therefore, in the vehicle speed detecting device, if only the fundamental wave corresponding to the passage of each timing tooth can be output as the AC signal Y1 from the sensor 12, the pulse signal P is also output in the correct phase, and the vehicle speed v and the acceleration are increased. It becomes possible to accurately detect α.

However, in reality, due to the following causes,
In the AC signal Y1 output from the magnetic sensor 12, components other than the fundamental wave, that is, a swell component with a low output wave and a noise component with a high frequency are mixed, and thus a relatively large error is included in the detected vehicle speed and acceleration. There was a problem of being lost.

(A) Occurrence of waviness That is, when a timing gear is used as the detected object 10, since the timing gear is a pressed product, the rigidity is low. Especially when this is attached to the wheel, both the base side and the timing gear are attached. Installation distortion occurs.

Further, in addition to such mounting distortion, cornering force or the like causes distortion.

As a result, the distance c between each timing tooth 10a and the magnetic sensor 12 provided on the surface of the detected object 10 during one rotation changes corresponding to the strain, and as shown in FIG. The low-frequency swell W is included in the output Y1 of 12 and the phase of the pulse signal P is ε1, depending on the conventional zero-cross method.
An error of ε2, ... Is included.

Therefore, even if the vehicle speed v and acceleration α are constant,
As shown in FIG. 15, when the waviness W rises, the output time interval of the pulse signal P becomes long, and when the waviness W falls, the output time interval of the pulse P becomes short and the speed v
Also, the detection result of the acceleration α includes an error component corresponding to the waviness W.

(B) Generation of noise In addition, when such a speed detection device is used in the ESC system, due to chatter from the brake and other influences, in the AC signal Y1, as shown in FIG. A higher frequency noise component N will be included.

Therefore, even when the velocity v and the acceleration α are constant, as shown in FIG. 16, errors of ε1 and ε2 occur in the phase of the pulse signal P, and the detected velocity v and the acceleration α include a predetermined error component. It will be.

In particular, the error caused by the noise component N becomes larger as the fundamental wave, that is, the sine wave component included in the AC signal Y1 becomes smaller.

Therefore, when the AC signal Y1 contains a large waviness component W in addition to the noise component N, the connection angle θ with the 0 level becomes small as the entire waveform falls due to the waviness, as shown in FIG. The phase error of is increased by the relation of ε∝1 / sin θ.

(C) As described above, in the AC signal Y1 output from the sensor 12, in addition to the fundamental wave corresponding to the tooth profile of the timing gear 10 in a one-to-one relationship, a low frequency swell component W and a high frequency The noise component N of is often included.

Therefore, there is a problem in that the timing at which each timing tooth 10a passes and the timing at which the pulse signal P is output are deviated, and a relatively large error is included in the detected vehicle speed v and the acceleration α.

In order to suppress the occurrence of such a measurement error, various measures have been conventionally taken in terms of both hardware and software.

Countermeasures by hardware processing (a) FIG. 19 shows a speed detection device according to Japanese Patent Laid-Open No. 55-83647. The speed detection device includes an output side of the sensor 12 and a control amplifier 20. A low-pass filter 18 whose cutoff frequency can be controlled by a control signal from the control amplifier 20 is provided between them to remove a component having a frequency higher than that of the fundamental wave.

Therefore, the device can remove the high frequency noise component N as shown in FIG. 16 from the AC output Y1 of the sensor 12, but on the other hand, the low frequency swell component included in the AC output Y1. W could not be removed, and it was not always possible to accurately detect the velocity V and the acceleration α.

(B) Further, in order to remove such a waviness component, a high-pass filter may be provided on the output side of the sensor 12. At vehicle speeds above the crossover frequency of the high-pass filter, the swell component cannot be removed, and at vehicle speeds below the crossover frequency, the swell is eliminated and the fundamental wave itself −
6dB attenuation is done. Wide range of vehicle speed (2.5Km / h ~ 250Km /
In this system that deals with h), it is necessary to change the crossover frequency according to the vehicle speed at that time, and it is extremely difficult to add a high-pass filter to the sensor output having an output amplitude proportional to the vehicle speed. .

Therefore, the development of a velocity detection device capable of accurately measuring velocity and acceleration without being affected by the undulation component W of low frequency and the noise component N of high frequency using such a filter has not yet been developed. Absent.

Measures by Soft Processing Further, FIGS. 20 to 23 show a proposal according to Japanese Patent Laid-Open No. 57-158564, in which the acceleration α is calculated by processing a pulse signal according to a predetermined program. The present invention relates to an acceleration detecting device for obtaining

This acceleration detecting device compensates for the period deviation ΔT (mainly due to poor mechanical accuracy) of the detection pulse signal P output from the speed detecting device, and therefore, the timing gear 10 is used for the period for detecting the acceleration / deceleration. Of the timing teeth 10a are sequentially n = 1, 2, 4, 8, ... (Modes 1, 2, 3, 4, ... Shown in FIG. 21.
Instead of), the period difference between adjacent sections is obtained.

Then, when the cycle difference is equal to or greater than the reference value S, the acceleration α is calculated based on the following equation.

α = (1 / T _{n + 1} −1 / T _n ) / {(T _{n + 1} + T _n ) / 2} When the cycle difference is smaller than S, the acceleration α = 0 is calculated.

However, in this conventional device, the period deviation ΔT smaller than S is
However, the effect of the mechanical deviation on the cycle deviation ΔT larger than S could hardly be exhibited.

Especially, the number of teeth is around 100, and the number of bolts for mounting the wheel is 5
Assuming a typical timing gear 10 having six gears, the period of the waviness W generated corresponds to 16 to 20 timing teeth 10a, which is significantly larger than the period deviation S of the pulse signal.

Therefore, when the period of the waviness component W is close to a multiple of n, the device has a problem that it is almost impossible to calculate the acceleration without being influenced by the component.

[Object of the Invention] The present invention has been made in view of such conventional problems, and its object is to be affected by a swell component and a noise component included in an AC signal output from a sensor, An object of the present invention is to provide a speed detecting device capable of accurately measuring speed or acceleration in real time.

[Means and Actions for Solving Problems] In order to achieve the above object, the present invention provides a rectangular processing circuit that performs rectangular processing on the AC signal and outputs a mountain side rectangular pulse and a valley side rectangular pulse, and each rectangular pulse. A pulse time measuring circuit for measuring the pulse time of, a memory circuit for sequentially updating and storing a pulse time sequence of at least three or more latest rectangular pulses that are continuously output, and a half cycle from the pulse time sequence. An average value calculation circuit that selects and sets two or more calculation pulse time sequences corresponding to one cycle of the AC signal, and calculates and outputs the average value of the calculation pulse time sequences based on a predetermined average value calculation formula. And a speed calculation circuit that calculates the vehicle speed or acceleration based on the average value of the calculation pulse time series, and is affected by the low frequency swell component or high frequency noise component included in the AC signal. It is characterized by detecting vehicle speed or acceleration.

[Embodiment] Next, a preferred embodiment of the present invention will be described with reference to the drawings.

ESC System FIG. 1 shows a preferred embodiment when the speed detecting device according to the present invention is used in an ESC system. In addition,
In the present embodiment, members corresponding to those of the device shown in FIG. 14 are designated by the same reference numerals and the description thereof will be omitted.

In the speed detecting device of the present embodiment, the AC signal Y1 detected by the sensor 12 is input to the waveform processing calculation circuit 30, where the vehicle speed v and the acceleration α are calculated and output to the ESC main body 32.

A feature of the present invention is that the velocity v and the acceleration α can be accurately calculated and output without being affected by the low frequency swell component W and the high frequency noise component N included in the AC signal Y1 of the sensor 12. It is in.

Therefore, by inputting the vehicle speed v and the acceleration α into the ESC main body 32, the ESC main body 32 accurately determines that this is a skid state when the deceleration of the corresponding wheel becomes a certain value or more,
It is possible to control so as to loosen the corresponding brake cylinder 34, and it is possible to output a control command to the brake cylinder 34 so that the brake works again when the vehicle speed v is restored.

In particular, according to the speed detecting device of the present invention, the vehicle speed v and the acceleration α can be detected in real time with almost no influence of the low frequency swell component and the high frequency noise component, which have been problems in the past, and the detected data can be detected. Since the error component is hardly included, the ESC main body 32 hardly misunderstands whether the corresponding wheel is in the skid state or not, and even during the sudden braking, the corresponding wheel is passed through the brake cylinder 34. It is possible to control the brakes satisfactorily.

Waveform Processing Operation Circuit Next, a preferred embodiment of the waveform processing operation circuit will be described.

FIG. 1 shows a schematic block diagram of the waveform processing arithmetic circuit 30, FIG. 2 shows its concrete circuit configuration, and FIG. 3 shows waveforms at various parts of the circuit. .

First, when the sensor 12 outputs the AC signal Y1, the AC signal Y1 is input to the rectangular processing circuit 40, where it is subjected to predetermined rectangularization processing and output as a rectangular pulse Y2. In the embodiment, the rectangular processing circuit 40 is formed by using a Schmitt trigger circuit whose reference value is 0 level.
As shown in FIG. 3, the rectangular pulse Y2 has an AC signal Y1 of 0V.
When it becomes the above, it is output as mountain side rectangular pulse 100B, and 0V
In the following cases, it will be output as a valley-side rectangular pulse 100A.

Then, the rectangular pulse Y2 is input to the first trigger circuit 42, and as shown in FIG. 3, a positive voltage trigger signal τ1 is output in synchronization with the rising of the mountain-side rectangular pulse 100B.

Further, the rectangular pulse Y2 is inverted through the inversion circuit 44, and this inversion output (1) is input to the second trigger circuit 46.

The second trigger circuit 46 outputs the trigger signal τ2 of a positive voltage in synchronization with the rising of the inverted output ▲ ▼.

Then, the OR circuit 48 outputs the positive voltage trigger signals τ1, τ output from the first and second trigger circuits 42, 46.
2 is output as a signal Y3 to the pulse time measuring circuit 50.

In the embodiment, these first and second trigger circuits 42,4
The 6 and the OR circuit 48 are formed by connecting in parallel two sets of differentiating circuits formed by using a capacitor C, a resistor R and a diode RE, as shown in FIG. A spike-like pulse is output as the trigger signals τ1 and τ2 in synchronization with the rising edge of.

Then, the pulse time measuring circuit 50, based on the trigger signals τ1 and τ2 thus output, the pulse times TB and TA of the mountain side rectangular pulse 100B and the valley side rectangular pulse 100A of the rectangular wave Y2.
Are measured respectively.

In the embodiment, the pulse time measuring circuit 50 is composed of a counter 52 and a standard clock generating circuit 54,
The trigger signal Y3 output from the OR circuit 48 is formed so as to be synchronized with the standard clock CLK output from the standard clock generation circuit 14.

The standard clock CLK is set so as to be output at a time interval sufficiently shorter than the trigger signal Y3, and the counter 52 outputs the first trigger signal τ1 to the second trigger signal τ2. During this period, the standard clock CLK is counted, and the mountain side rectangular pulse 100B output from the rectangular processing circuit 40.
Measure N pulse time TBN.

Similarly, the counter 52 counts the standard clock CLK from the output of the second trigger signal τ2 to the output of the first trigger signal τ1 and outputs it from the rectangular processing circuit 40. Valley-side rectangular pulse 100AN pulse time TAN
To measure.

The respective pulse times TAN and TBN measured in this way are sequentially written and stored in the memory circuit 58 via the average value calculation circuit 56 as a pulse time train.

In the present invention, the storage circuit 58 sequentially adds the latest pulse times TAN-1, TBN-1, TAN, TBN, ... Of at least three rectangular pulses which are continuously output as described above. It is configured to update and store.

Then, the average value calculation circuit 56 selects and sets a plurality of predetermined calculation pulse time sequences from the pulse time sequences written in the storage circuit 58.

In the present invention, each of these operation pulse time trains is selected and set so as to correspond to one cycle of the AC signal Y1 and to overlap each other by a half cycle of the AC signal.

Then, each of the selected and set arithmetic pulse time series is substituted into a predetermined average value arithmetic expression, and the average value Tn of each arithmetic pulse time series is arithmetically output.

In this way, when an average value of a plurality of operation pulse time trains that overlap each other by a half cycle is obtained, an error component due to the influence of the low frequency swell component W and the high frequency noise component N included in the AC signal Y1. Cancel each other out, and the average value of the calculated operation pulse time series accurately represents the time interval in which each timing tooth 10a passes the front surface of the sensor 12.

Therefore, the speed calculation circuit 60 calculates the vehicle speed V and the acceleration α in real time based on the average value Tn of the calculation pulse time series thus obtained without being affected by the swell component W and the high frequency noise component N. Moreover, it can be detected accurately.

First Specific Example of Calculation of Average Value TN FIG. 4 shows a preferred example of calculating the Nth output cycle TN of the AC signal Y1. The average value calculation circuit of this example is shown. 56 reads the latest three rectangular pulse output times TBN-1, TAN, and TBN from the memory circuit 58, respectively.

Then, from the plurality of calling pulse times, a calculation pulse time train (TBN-1 + TAN) for one cycle of the AC signal that overlaps each other by a half cycle (in the embodiment, overlaps by TAN),
By selectively setting (TAN + TBN) and substituting this into the following equation to perform quadratic binomial expansion, TN is calculated from the average value of the calculated pulse time sequence.

TN = {(TBN-1 + TAN) + (TAN + TBN)} / 2 = {TBN-1 + 2
TAN + TBN} / 2 (1) This average value TN will be obtained as the output time interval between the (N-1th peak position) and (Nth peak position) of the AC signal Y1. I will explain in detail (a),
For the reason of (b), it is possible to remove the influence of the swell component W and the noise component N contained in the AC signal Y1.

(B) Reason why the undulation component W is removed That is, as shown in FIG. 5, even though the original position of the AC signal output from the sensor 12 is Y1, that position is affected by the undulation component W. As a result, assuming that the position has been shifted to the upper Y1 ′ position as indicated by the dotted line in the figure, the timing at which the mountain side rectangular pulse is output from the rectangular processing circuit 40 is advanced by Δt, and the valley side rectangular pulse is Is delayed by Δt.

On the contrary, when the AC signal shifts from the original position Y1 to the position Y1 ″ below due to the influence of the waviness component W, the peak-side rectangular pulse output is delayed by Δt, and the valley-side rectangular pulse output. Is faster than the main unit by Δt.

On the other hand, like the present invention, the average value of a plurality of operation pulse time trains (TBN-1 + TAN), (TAN + TBN), which correspond to the time of one cycle of the AC signal Y1 and overlap each other by a half cycle.
By obtaining Tn, it becomes possible to offset the time difference Δt at which the rectangular pulse Y1 is output, and obtain the output time TN for one cycle of the AC signal Y1 in a state where it is not affected by the waviness component W.

(B) Reason why the noise component N is removed Further, in the present invention, it is possible to statistically reduce the influence of the high frequency noise component N included in the AC signal Y1.

That is, the mountain side rectangular pulse 100B based on the AC signal Y1,
When the valley-side rectangular pulse 100A is output, the timing of its output time is influenced by the high frequency noise component N included in the AC signal Y1.

However, as in the present embodiment, the effect of the noise component N is statistically calculated by obtaining the average value of the two operation pulse time sequences based on the first equation. It is possible to significantly reduce the influence of the error by compressing into

As described in (a) and (b) above, according to the present invention, the average value TN of a plurality of operation pulse time sequences corresponding to one cycle of the AC signal Y1 and overlapping each other by a half cycle is obtained. By so doing, the effects of the low-frequency swell component W and the high-frequency noise component N contained in the AC signal Y1 can be effectively reduced, and the time interval at which each timing tooth 10a passes the front surface of the sensor 12 can be accurately obtained. Is possible.

Therefore, it will be understood that the vehicle speed V and the acceleration α can be accurately calculated in real time based on the average value TN thus obtained without being affected by the swell component W and the high frequency noise component N.

Second Specific Example of Calculation of Average Value TN Further, in the first specific example, the case where the average value of two calculation pulse time series is obtained has been described as an example, but the present invention is not limited to this, and may be performed as necessary. It is also possible to form an average value of three or more calculation pulse time series and calculate and output the vehicle speed v and the acceleration α based on the average value.
Specific examples will be described below.

In FIG. 6, the average value calculation circuit 56 reads out the pulse times TAN-1, TBN-1, TAN and TBN of four rectangular pulses from the memory circuit 58, and the pulse times TAN-1, TBN-1, TAN and TBN shown in FIG. A specific example in the case of selectively setting a set of calculation pulse time trains is shown.

First, as the first operation pulse time train, (TAN-1 + TBN-
1) is set, and (TBN-1 +
Select (TAN) and set (T
Select AN + TBN).

Then, the first and second operation pulse time series are substituted into the first equation and the average value TN ′ thereof is calculated, and similarly, the second and third operation pulse time series are calculated as The arithmetic processing is performed in the same manner as in Formula 1, and the average value TN ″ is calculated,
If the average value of the two average values TN ′ and Tn2 ″ is used as the average value TN of the final operation pulse time train, the waviness component W
The influence of the noise component N can be further reduced, and the vehicle speed v and the acceleration α can be calculated more accurately.

That is, the average value calculation circuit 56 of the embodiment performs the cubic binomial expansion as shown in the following expression, weights the central calculation pulse time train three times, and calculates the average of each calculation pulse time train. value
TN will be calculated, and by doing this,
The effect of the swell component W and the noise component N can be reduced more effectively than in the above embodiment, and the time interval for each timing tooth 10a to pass the front surface of the sensor 12 can be accurately determined as TN.

TN ′ = (TAM-1 + 2 ・ TBN-1 + TAN) / 2 (2A) TN ″ = (TBM-1 + 2 · TAN + TBN) / 2… (2B) TN = (TM ′ + TN ″) / 2 = (TAN-1 + 3 ( TBN-1 + TAN) + T
BN) / 4 (2C) Therefore, by using the average time TN thus obtained, the vehicle speed v and the acceleration α can be calculated more accurately without being affected by the swell component W and the high frequency noise component N. It becomes possible to ask.

As described above, in the first specific example, as shown in the first equation, the two operation pulse time sequences are selectively set so as to overlap in the TAN section, and in the second specific example, , As shown in the second equation,
The case has been described as an example where selection and setting are made so that they overlap in the sections of TBN-1 and TAN respectively.

However, the present invention is not limited to this, and it suffices to set a plurality of calculation pulse time trains so that a part of them overlaps for a half cycle. It may be formed by any method of selecting and setting the average value TN, and by doing so, the effects of the swell component W and the noise component N can be effectively reduced in the same manner as in each of the specific examples. , Vehicle speed v and acceleration α can be calculated and output.

Further, in the present invention, for example, two or more average values based on the first or second equation are calculated as TN ′, TN ″ ... It is also possible to obtain the average value TN of the calculation pulse time series of.

For example, in the formula 2C, the case where only the central operation pulse train (TBN-1 + TAN) is weighted three times to obtain the average value of each operation pulse time train has been described as an example, but the present invention is not limited to this. However, it is also possible to give higher order binomial expansion, that is, to give more weight to the central operation pulse time train, if necessary, and to even out each operation pulse time train without giving such weighting. It is also possible to add to and obtain the average value.

As described above, when two or more average values obtained by the first or second equation are used to further obtain the average value,
Rather than adding the pulse time of each rectangular pulse individually, the average value TN ′, T obtained using the first or second equation
It is more efficient to perform the TN operation using N ″ ..., By doing so, it becomes possible to simplify the circuit design.

In each of the above embodiments, three pulse times TBN-1,
An example is given for the case where the second-order binomial expansion is performed using TAN and TBN and the case where the third-order binomial expansion is performed using the four pulse times TAN-1, TBN-1, TAN, TBN. However, it is also possible to use five or more pulse times and perform binomial expansion of the fourth or higher order, if necessary.

However, according to the experiment, it was confirmed that even if the pulse time series determined by the binomial expansion of the fourth or higher order was taken and the average value was obtained, only the accuracy similar to the calculation data shown in the second formula could be obtained.

Another Embodiment of Rectangular Processing Circuit FIG. 7 shows a second embodiment of the rectangular processing circuit 40 shown in FIG. 1, and FIG. 8 shows its concrete circuit configuration. . The members corresponding to those in the above embodiment are designated by the same reference numerals, and the description thereof will be omitted.

The characteristic of the present embodiment is that the reference value Yref for rectangular processing is fixed to zero level in the above-described embodiment, whereas the reference value Yref of the rectangular processing circuit 40 is made to follow the waviness component W, This is to set an intermediate value between the peak value and the valley peak value for each output cycle of the AC signal Y1 each time.

By doing so, for example, as shown in FIG. 9, even if the AC output Y1 of the sensor 12 largely moves up and down due to the influence of the swell component W and does not cross the zero volt line, the old reference voltage Since a new reference voltage Yref is set between Yref and the peak value on the mountain side or the valley side, the rectangular pulse 100 is always applied regardless of the magnitude of the swell component W.
It is possible to reliably output A and 100B and perform the measurement more accurately.

The specific circuit configuration of the rectangular processing circuit 40 of this embodiment will be described below.

First, when the sensor 12 outputs an AC signal Y1 as shown in FIG. 9, the AC signal Y1 is input to an amplification circuit 61 with AGC (automatic gain adjustment), and has a constant width according to an AGC signal described later. Is amplified into the impedance signal Y1 'of the above and is input to the peak hold circuits 62a and 62b.

The peak hold circuit 62a holds the peak value Ymax on the mountain side of the AC signal Y1 ', and is specifically formed by using an operational amplifier OP2, a diode RE and a capacitor C1 as shown in FIG.

In addition, the other peak hold circuit 62b receives the AC signal Y1 '.
It holds the peak value Ymin on the valley side. Specifically, as shown in FIG. 8, operational amplifier OP3, diode
It is formed using RE and capacitor C2.

Then, the reference value setting circuit 64 sets the peak hold value Ym
Ax and the lower hold value Ymin are added and averaged, and the average value is set as the reference value Yref.

Actually, the peak hold value of the previous two times is reset to the reference value each time by the reset operation described later.

Such setting of the reference value Yref is performed by using a voltage dividing circuit composed of a variable resistor VR2, as shown in FIG. 8, and may be set to an appropriate value as necessary. However, in this embodiment, it is set to an intermediate value between the peak hold value and the lower hold value.

Then, the comparison circuit 66 compares the reference value Yref output in this way with the AC signal Y1 ′, and when Y1 ′ exceeds Yref, outputs the mountain side rectangular pulse 100B, and Y1 ′ is the reference value Yr.
When it goes below ef, the valley-side rectangular pulse 100A is output.

Further, the rectangular processing circuit 40 of the present embodiment uses the trigger signal Y3 output from the OR circuit 48 as a reset pulse for each peak so as to set the optimum reference value Yref that follows the waviness component W for each half cycle of the AC signal Y1. Feedback is input to the hold circuits 62a and 62b.

By doing so, each peak hold circuit 62a
The hold values of 62b and 62b are reset and updated to new values every half cycle of the AC signal Y1.

Therefore, the reference value setting circuit 40 sets the peak hold value Ymax and the lower hold value Ym that are newly set every half cycle.
Since the optimum reference voltage Yref can be set each time based on in, as described in detail in (a) and (b) below,
The comparison circuit 66 can output the rectangular pulse 100A or 100B at an optimum timing without being affected by the low-frequency swell component W and the high-frequency noise component N included in the AC signal Y1.

(A) Reason why the swell component W is removed As shown in FIG. 18, the swell component W should be output regardless of the fact that the AC signal should be output from the magnetic sensor 12 to the position of Y1.
It is assumed that an AC signal is output at the position of Y1 'due to the influence of.

In this case, in the above-mentioned zero-cross method, the rectangular pulse Y2 will be output shifted by ε from the original position,
This error component ε increases as the swell increases.

On the other hand, in the device of the present embodiment, based on the peak hold value Ymax newly set for each cycle of the AC signal Y1,
Since the optimum reference voltage Yref that follows the change of the swell component W, that is, the so-called second zero level Yref shown by the broken line in the figure can be set for each cycle, regardless of the swell component W, the optimum phase is always obtained. It becomes possible to accurately output the rectangular pulse Y2.

(B) Reason why the noise component is removed Further, the influence of the noise component N contained in the AC signal Y1 is determined by the magnitude of the fundamental wave cut out at zero level, that is, the sine wave component, and the fundamental wave component with respect to the noise component is determined. Is relatively large, the influence of the noise component is small.

That is, as shown in FIG. 17, it is assumed that a noise component is superimposed on the fundamental wave (sine wave).

In this case, if the reference voltage Yref is set near the peak value of the sine wave, the setting error due to the noise component N will be as large as ε. It is understood that the error due to the noise component N is significantly reduced to ε'when the positions are set to 2π ,.

Therefore, when the rectangular pulse Y2 is output using the zero-cross method, when the deviation amount between the zero potential, which is the reference voltage, and the original zero-cross position of the sine wave becomes large due to the influence of the waviness component W, the rectangular pulse Y2 is output. It will be understood that the output phase of ∘ will deviate significantly from the normal position.

On the other hand, according to the device of the present embodiment, the reference voltage Yref
Can be set to a position where the inclination of the sine wave component becomes maximum for each cycle of the AC signal Y1, each time,
Even if the AC signal Y1 contains a noise component N, or if the swell component W is contained in addition to the noise component N, the rectangular pulse Y2 is not affected by them.
Can be output with an accurate phase.

As described in (a) and (b) above, according to the present embodiment, the low frequency swell component W included in the AC signal Y1 is influenced by the high frequency noise component N from the comparison circuit 66. Without this, the rectangular pulses 100A and 100B are output in the correct phase each time each timing tooth 10a passes the front surface of the sensor 12.

Therefore, by performing the arithmetic processing of the present invention as shown in the first equation or the second equation on the pulse subjected to such rectangularization processing, as compared with the embodiment using the zero-cross method described above. It will be understood that the effect of the swell component W and the noise component N can be further greatly reduced and the vehicle speed v and the acceleration α can be detected.

Further, the rectangular processing circuit 40 of the present embodiment inputs the trigger signal Y3 output from the OR circuit 48 to the differential amplifier 68 and inputs the differential output to the AGC control voltage setting circuit 70.

Based on the output timing of the trigger signal Y3, the differential amplifier 68 and the AGC control voltage setting circuit 70 direct a predetermined AGC signal to the amplification circuit 61 with AGC so that the amplitude of the AC signal Y1 output from the sensor 12 becomes constant. It is outputting.

Specifically, as shown in FIG. 8, the peak hold value Ymax is inverted and output by the inverting amplification operational amplifier OP4, and the intermediate voltage YD between the inverted output ▲ ▼ and the lower peak hold value Ymin is output via the analog switch. Formed to capture.

The analog switch is formed by connecting two diodes RE in antiparallel so that a predetermined dead band width is set, the transmission of the circuit itself is prevented, and its high-speed followability is improved.

The intermediate voltage YD thus taken in is compared with the comparison voltage YE set by the volume VR1, and the amplitude is limited by the operational amplifier OP5, the capacitor c and the limiter Zener diode TD based on the differential voltage. It is integrated and output using an integrating circuit.

Here, as described above, the trigger signal Y3 output from the OR circuit 42 is input to the FET1 to control the integration value and the integration period of the intermediate voltage YD, and the integration value is used as the AGC signal AGC. It outputs to the attached amplification circuit 61.

As shown in FIG. 10, the signal Y3 ′ input from the OR circuit 42 to each peak hold circuit 62a, 62b uses the delay circuit 72 as compared with the signal Y3 input to the differential amplifier circuit 68. The output is delayed by a predetermined short time, and the timings of the AGC control operation and the reference voltage Yref setting operation described above are adjusted.

In the concrete circuit shown in FIG. 8, the first trigger circuit 42, the inverting circuit 44, the second trigger circuit 46, the OR circuit 48, the counter 52, the flat value calculating circuit 56, and the standard clock generating circuit 58. The delay circuit 72 and the speed calculation circuit 60 are formed as a one-chip element.

Examination of Measured Data Next, the data actually measured using the speed detecting device according to the present invention will be examined in comparison with the conventional device or other devices.

In FIG. 11, a timing gear having 100 timing teeth 10a projecting around it is used as the detected body 10, and when the gear 10 is mounted on the axle using five mounting bolts, A waveform diagram of the AC signal Y1 output from the sensor 12 is shown.

As is clear from the figure, this AC signal Y1 exhibits five undulations or five amplitude changes in relation to the fact that the number of mounting bolts is five.

Of these, the first 1.75 cycles were undulations, the last 1.75 cycles were amplitude changes, and the middle 1.5 cycles were gradual changes.
By doing so, the AC signal Y1 included in the first 1.75 cycles is expressed by the following equation, and Y1 = −0.5 · cos (0.1π · t) −cos (2π · t) ...
(3) The AC signal Y1 output in the latter half 1.75 cycles is expressed by the following equation.

Y1 =-(1-0.5 ・ cos (0.1π ・ t)) ・ cos (2π ・
t) (4) Under these conditions, the device shown in FIG. 1 (zero-cross method) is used to rectangularize the AC signal Y1, and the time TN for one cycle of the AC signal Y1 is simply When TN = TAN + TBN is obtained (conventional method), when it is obtained as shown in the first equation, and when it is obtained as shown in the second equation. And how much error is included in the acceleration α was experimentally determined.

In FIG. 11, the experimental data are shown in order from the left as the cumulative error from the normal, the relative error before and after, and the 8-wavelength moving average error.

Here, the error from the normal represents the deviation from the normal position of the timing gear during one cycle.

The front-to-back relative error represents the relative error of one tooth crest.

The 8-wavelength moving average error represents an average value of eight consecutive pitch errors or a value obtained by dividing the change amount of the regular positional deviation amount by 8.

Further, in FIG. 12, the respective comparison data thus obtained are converted into numerical data and displayed.

As is clear from this experimental result, according to the apparatus of the present embodiment, the error caused by the swell component W is at least about compared to the case where the vehicle speed v and the acceleration α are simply obtained from the pulse time of the rectangular wave. It is confirmed that it is possible to exhibit an excellent effect that the error can be reduced to about 1/10 and the error caused by the noise component N can be reduced to a negligible level.

Further, FIG. 13 shows data obtained when the same experiment was performed using the circuit shown in FIG. 7 as the rectangular processing circuit. In FIG. 11, the comparison data was converted into numerical data. The displayed value is displayed.

As shown in the figure, by setting the reference voltage Yref for each half cycle from the peak-side peak value and the valley-side peak value of the AC signal Y1, as is apparent from the experimental results, the error due to the swell component is clear. Is significantly reduced, and highly accurate vehicle speed v
It will be appreciated that the acceleration α can be detected.

In each of the above embodiments, as shown in FIG.
The case where the average value calculation circuit 56 and the speed calculation circuit 60 are separately formed is described as an example. However, when the speed detection device like the present invention is used in an ESC system, each of the calculation circuits 56 and The CPU 60 is often formed as a software using a CPU, and the CPU is often also used as the CPU of the ESC main body 32.

Therefore, in such a case, the CPU often executes the calculation of the vehicle speed v and the acceleration α and other calculation operations in parallel almost at the same time, and when the load of the other calculation processing operation is large,
The pulse time T output from the counter 52 is not always accepted.

Therefore, in such a case, after the pulse time output from the counter 52 is read and the average value TN is calculated based on the first expression or the second expression, the reading from the counter 52 is stopped for a predetermined time, The reading from the counter 52 may be started again when the load of other arithmetic processing operations is lightened, and the average value may be calculated based on the first or second equation.

In this case, the vehicle speed v can be calculated only based on the read average value TN, but the acceleration α can be calculated.
Since it is performed by detecting the change in TN per unit time, when detecting such an acceleration α, the number of teeth counted from the time when the previous average value is calculated to the time when the next average value is calculated. If the information about how many 10a's are skipped is input at the same time, the acceleration α can be calculated at the same time.

Further, in the above-described embodiment, the case where the Schmitt trigger circuit is used as the rectangular processing circuit 40 and the case where the circuit for holding each peak value and performing the rectangular processing is used as shown in FIG. 7 have been described as an example. The present invention is not limited to this,
The AC signal Y1 is compared with a predetermined reference value Yref, and the AC signal Y1 is
It is sufficient to be able to detect whether Y1> Yref or Y1 <Yref. Therefore, for example, the rectangular processing circuit 40 does not output the rectangular wave pulse itself, but the rectangular processing circuit 40 can be formed so as to output a signal representing only the rising and falling points of the pulse.

Further, when the circuit itself is configured at the TTL level, it is sufficient that the rectangular processing circuit 40 is configured to output a signal at the L level or the H level as the rectangular wave pulse Y2. AC signal Y1 is Y1 <Yr
It is sufficient to form it so as to detect whether ef or Y1> Yref.

Further, in each of the embodiments, the case where the count value of the standard clock CLK is written and set in the memory circuit 58 as the data representing the pulse time of each rectangular pulse is described as an example, but the present invention is not limited to this. Instead of the output time TAN, TBN ... Of each rectangular pulse, it is also possible to write the cumulative count value before these rectangular pulses are output.
In this case, the calculation shown in the following equation may be performed instead of the first equation.

TN = {(total count value up to TBN-total count value up to TBN-1) + (total count value up to TAN-total count value up to TAN-1)} / 2 (1) '[Effect of the invention As described above, according to the present invention, even if a low-frequency swell component or a high-frequency noise component is mixed in the AC signal output from the sensor, these are affected. It is possible to accurately detect the vehicle speed V or the acceleration α without using it, and therefore, it is extremely suitable as various vehicle-mounted devices that require extremely high detection accuracy, for example, a speed detection device for an ESC system.

In particular, according to the present invention, since it is difficult to be affected by the swell component contained in the AC signal, even if a timing gear or the like of an inexpensive pressed product is used as the detected object arranged facing the sensor, its speed Alternatively, the acceleration can be detected in real time and with high accuracy, and it is possible to provide an inexpensive and high-performance speed detecting device.

[Brief description of drawings]

1 and 2 are circuit diagrams showing a preferred embodiment of the vehicle speed detecting device according to the present invention, FIG. 3 is a waveform explanatory diagram in each part of the circuit shown in FIGS. 1 and 2, and FIG. ~ Fig. 6 is an explanatory diagram of the arithmetic operation performed by using the circuits shown in Figs. 1 and 2, and Figs. 7 and 8 are rectangular processing circuits used in the circuits shown in Figs. 1 and 2. 9 is a circuit diagram showing another embodiment of the present invention, FIG. 9 and FIG. 10 are explanatory diagrams of waveforms in each part of the arithmetic operation performed by the circuit shown in FIG. 7 and FIG. FIG. 13 is an explanatory view showing data obtained by experiments for the effect of the device of the present invention, FIG. 14 is a block diagram of a general speed detecting device, and FIGS. 15 to 18 each include a low frequency waviness component. Waveforms of AC signals or AC signals containing high frequency noise components, Figures 19 to 23 show conventional speed It is an explanatory view of the detection device. 10 …… Rotating object 12 …… Sensor 32 …… ESC main body 34 …… Brake cylinder 40 …… Rectangular processing circuit 42 …… First trigger circuit 44 …… Inversion circuit 46 …… Second trigger circuit 48… … Or circuit 50 …… Pulse time measurement circuit 56 …… Average value calculation circuit 58 …… Memory circuit 60 …… Speed calculation circuit

Claims (5)

[Claims] 1. A rotary detection object provided on a vehicle rotation side, and a rotation detection object provided on a vehicle fixing portion side so as to face the detection object.
A sensor for outputting an alternating current signal each time a detected body passes, and a vehicle speed detecting device for calculating a vehicle speed or acceleration based on the alternating current signal, wherein the alternating current signal is subjected to a rectangular process to form a mountain side rectangular pulse and a valley. A rectangular processing circuit that outputs side rectangular pulses, a pulse time measurement circuit that measures the pulse time of each rectangular pulse, and a pulse time train of the latest at least three rectangular pulses that are output consecutively are sequentially updated and stored. From the memory circuit and the pulse time train, two or more calculation pulse time trains that overlap each other by a half cycle and correspond to one cycle of the AC signal are selected and set, and the calculation is performed based on a predetermined average value calculation formula. An average value calculation circuit that calculates and outputs the average value of the pulse time series, and a speed calculation circuit that calculates the vehicle speed or acceleration based on the average value of the calculated pulse time series A vehicle speed detecting device for detecting a vehicle speed or acceleration without being affected by a low frequency swell component or a high frequency noise component. 2. The apparatus according to claim 1, wherein the rectangular processing circuit has a peak hold circuit that peak-holds both a peak-side peak value and a valley-side peak value of the AC signal, and both peak-hold circuits. A reference value setting circuit that sets a reference signal based on the average value of the values, and a comparison circuit that compares the AC signal with the reference value and outputs a ridge-side rectangular pulse or a valley-side rectangular pulse. A vehicle speed detection device, wherein the peak hold circuit is reset to output a rectangular pulse in which the influence of a low frequency swell component or a high frequency noise component included in an AC signal is reduced. 3. The apparatus according to any one of claims (1) and (2), wherein the memory circuit has at least three latest pulse time trains TBN-1 that are continuously output. , TAN, TBN ... Are sequentially updated and stored, and the average value calculation circuit calculates and outputs the average value TN of the calculation pulse time sequence based on the following equation. TN = {(TBN-1 + TAN) + (TAN + TBN)} / 2 However, TBN and TAN represent the peak pulse time and the valley pulse time, respectively, and the subscript N represents the output order of each rectangular pulse. 4. The apparatus according to claim 1, wherein the storage circuit outputs the latest pulse time of at least four rectangular pulses that are continuously output. Row TAN-1, TBN-1,
A vehicle speed detecting device characterized in that TAN and TBN are sequentially updated and stored, and the average value calculating circuit calculates and outputs an average value TN of a calculation pulse time sequence based on the following equation. TN = {(TAN-1 + TBN-1) +3 (TBN-1 + TAN) + (TAN + T
BN)} / 4 where TAN and TBN represent the rectangular pulse time on the mountain side or the valley side, respectively, and the subscript N represents the output order of the rectangular pulses. 5. The apparatus according to claim 3 or 4, wherein the average value calculation circuit further includes a plurality of average values TN of operation pulse time sequences obtained consecutively. A vehicle speed detection device characterized by averaging the individual values and outputting the average value as an average value of a calculation pulse time sequence.