JP7283223B2 - VEHICLE POSITIONING SYSTEM, VEHICLE POSITIONING METHOD AND VEHICLE POSITIONING PROGRAM - Google Patents

Description

TECHNICAL FIELD The present invention relates to a technique for specifying the position of a vehicle based on information detected around the vehicle.

Japanese Patent Application Laid-Open No. 2015-41348 describes a system that identifies the position of the vehicle in a parking lot by recognizing images of guide marks drawn on the road surface, and automatically drives the vehicle into and out of the parking space. A parking guidance system is disclosed that does. In particular, in places where GPS (Global Positioning System) signals do not reach, or when higher positional accuracy than GPS accuracy is required, image recognition of guidance signs in this way enables highly accurate vehicle positioning. can be located.

Japanese Unexamined Patent Application Publication No. 2015-41348

By the way, the image recognition process is generally a process with a high processing load, and there is a delay from when the guide sign is photographed by a camera or the like until the guide sign is recognized. During this time, if the vehicle is running, an error will occur between the position of the vehicle specified based on the guidance sign and the position where the vehicle actually exists. The faster the vehicle travels, the larger this error becomes, and the slower the vehicle travels, the smaller the error. is desired. In addition, this is not limited to general image recognition processing for visible light imaging, but recognition processing for infrared images using infrared light, distance images using ultrasonic waves, laser radar, etc. But it's the same.

In view of the above background, it is desired to provide a technique for accurately specifying the position of a vehicle based on information detected around the vehicle.

In view of the above, the vehicle position identification system has, as one aspect, a detection information acquisition unit that acquires detection information around the target vehicle, and a detection time information acquisition unit that acquires detection time information indicating the detection time of the detection information. a recognition processing unit that performs recognition processing of a recognition target object included in the detection information and calculates a relative position between the recognition target object and the target vehicle; and the target based on the position of the target vehicle at each time. a traveling locus storage unit for storing a traveling locus of a vehicle; an object information acquiring unit for acquiring object absolute position information indicating a position of the recognition object in a reference coordinate system; a first position identifying unit that identifies a first position, which is the position of the target vehicle at the detection time, based on object absolute position information; a second position specifying unit that specifies a second position, which is the position of the target vehicle at the target time, based on the target travel locus, which is the travel locus, and the first position;

According to this configuration, the position (first position) of the target vehicle is specified by going back to the detection time when the detection information was detected, and based on the target travel locus, which is the travel locus from the first position, and the first position. , the position (second position) of the target vehicle at the target time corresponding to the current time is specified. Therefore, even if the recognition processing for the detection information takes time, it is possible to suppress the error of the second position caused by the movement of the target vehicle during that time. That is, according to this configuration, it is possible to accurately identify the position of the vehicle based on the information detected around the vehicle.

The above technical characteristics of the vehicle position specifying system can also be applied to the vehicle position specifying direction and the vehicle position specifying program. A representative embodiment thereof is illustrated below. For example, a vehicle localization method has various steps with features of the vehicle localization system described above. Further, the vehicle position identification program causes the computer to implement various functions having the characteristics of the vehicle position identification system described above. Of course, these vehicle position specifying method and vehicle position specifying program can also produce the effects of the vehicle position specifying system described above. Furthermore, as preferred aspects of the vehicle position identification system, various additional features illustrated in the description of the embodiments below can also be incorporated into these vehicle position identification methods and vehicle position identification programs. The program can also perform effects corresponding to each additional feature.

As one preferred aspect, a vehicle position identification method includes a detection information acquisition step of acquiring detection information around a target vehicle; a detection time information acquisition step of acquiring detection time information indicating a detection time of the detection information; a recognition processing step of performing recognition processing of a recognition target contained in the detection information and calculating a relative position between the recognition target and the target vehicle; a running locus storing step of storing a running locus; an object information obtaining step of obtaining object absolute position information indicating a position of the recognition object in a reference coordinate system; a first position specifying step of specifying a first position, which is the position of the target vehicle at the detection time, based on the absolute position information; and the traveling of the target vehicle from the detection time to a target time corresponding to the current time. a second position specifying step of specifying a second position, which is the position of the target vehicle at the target time, based on the target travel trajectory, which is a trajectory, and the first position.

Further, as one preferred aspect, the vehicle position identification program includes a detection information acquisition function for acquiring detection information around the target vehicle, and a detection time information acquisition function for acquiring detection time information indicating the detection time of the detection information. a recognition processing function for performing recognition processing of a recognition target object included in the detection information, and calculating a relative position between the recognition target object and the target vehicle; a running track storage function for storing a running track of a vehicle; an object information acquisition function for acquiring object absolute position information indicating a position of the recognition object in a reference coordinate system; a first position identifying function that identifies a first position, which is the position of the target vehicle at the detection time, based on object absolute position information; A computer is caused to realize a second position specifying function of specifying a second position, which is the position of the target vehicle at the target time, based on the target travel trajectory, which is the travel trajectory, and the first position.

Further features and advantages of the vehicle localization system, vehicle localization method and vehicle localization program will become apparent from the following description of the embodiments described with reference to the drawings.

A diagram showing an example of a parking lot to which a vehicle positioning system is applied Block diagram schematically showing a configuration example of a vehicle position identification system The figure which shows an example which acquires the detection information around a vehicle including a recognition target. A diagram showing an example of a positional relationship between a recognition target object and a vehicle in a local coordinate system. The figure which shows an example which correct|amends the position of the vehicle in a reference coordinate system. Flowchart showing an example of a vehicle position identification procedure Flowchart showing another example of the procedure for identifying the vehicle position A diagram showing an example of the target travel locus based on the procedure shown in FIG. A diagram showing an example of the target travel locus based on the procedure shown in FIG. The figure which shows an example which correct|amends a driving|running|working locus. The figure which shows the other example which acquires the detection information around a vehicle containing a recognition target.

Hereinafter, embodiments of a vehicle position specifying system (vehicle position specifying method, vehicle position specifying program) will be described based on the drawings. Vehicle location identification systems include, for example, an automatic valet parking system that automatically enters and exits a vehicle in a parking lot, a driving support system that assists the driver in the parking lot (parking support system), a route to the destination that allows the passenger to It can be used for a navigation system or the like that guides the A form in which a vehicle position specifying system (vehicle position specifying method, vehicle position specifying program) is used in an automatic valet parking system will be described below as an example.

FIG. 1 shows an example of a parking lot 100 to which an automated valet parking system is applied. The parking lot 100 includes a parking space 103 that is partitioned by a parking frame line 105 and serves as a parking place for a vehicle 9, an alighting space 101 that an occupant 91 gets off from a vehicle 9 to be entered into the parking space 103, and a parking space. A boarding space 102 is provided for a passenger 91 to board the vehicle 9 to be delivered from the space 103 . Although FIG. 1 shows an example in which one alighting space 101 and one boarding space 102 are provided, a plurality of each of the alighting space 101 and boarding space 102 may be provided. Also, the same one space may be used as both the alighting space 101 and the boarding space 102 . Alternatively, a plurality of spaces may be provided without fixing the alighting space 101 and the boarding space 102, and may be used as the alighting space 101 or the boarding space 102 as appropriate.

At the time of parking, after the passenger 91 gets off from the vehicle 9 in the unloading space 101, the vehicle 9 automatically travels from the unloading space 101 to the empty parking space 103 in response to the parking instruction, and automatically moves in the parking space 103. stop at As a result, the vehicle 9 is automatically parked (entered) in the parking space 103 . When leaving the parking lot, the vehicle 9 automatically travels from the parking space 103 to the boarding space 102 and automatically stops in the boarding space 102 in response to the leaving instruction. Thereby, the vehicle 9 is automatically pulled out of the parking space 103 . A warehousing instruction and a warehousing instruction for the vehicle 9 are given to the vehicle 9 by an operation of a terminal device 93 by an occupant 91 of the vehicle 9 . The terminal device 93 may be a mobile terminal carried by the passenger 91 or may be a fixed terminal installed in the parking lot 100 .

That is, in the automatic valet parking system, the vehicle 9 automatically travels along the set travel routes R from the drop-off space 101 to the parking space 103 and from the parking space 103 to the boarding space 102. . In order to realize such automatic driving, it is desirable to specify the position of the vehicle 9 in the parking lot 100 with high accuracy.

As shown in FIG. 2 , the vehicle position specifying system 10 is configured as one function of a vehicle control device 50 that controls the vehicle 9 . The vehicle control device 50 is configured as an ECU (Electronic Control Unit) with a processor such as a microcomputer working in cooperation with peripheral circuits such as a memory as a core. Programs and various parameters are stored in the memory, and the vehicle position specifying system 10 is implemented by cooperation between software such as the program and hardware such as a processor. The vehicle position specifying system 10 is a system for specifying the position of the vehicle 9, and the vehicle position specifying method is a method for specifying the position of the vehicle 9 using hardware and software that constitute the vehicle position specifying system 10. The identification program is a program that is executed by hardware (computer) that constitutes the vehicle position identification system 10 and that implements a vehicle position identification function.

Although the details will be described later, as shown in FIG. 2, the vehicle position specifying system 10 includes a detection information acquisition unit 1, a detection time information acquisition unit 2, a recognition processing unit 3, a travel locus storage unit 4, an object An information acquisition unit 5 and a vehicle position specifying unit 6 are provided. The own vehicle position specifying section 6 includes a position information acquiring section 7 , a first position specifying section 11 and a second position specifying section 12 . In addition, the vehicle control device 50 includes a travel control unit 20 and an object information storage unit 30 (database: DB) in addition to the vehicle position specifying system 10 . A camera C, a GPS receiver 41, an orientation sensor 42, and a rotation sensor 43 are connected to the vehicle control device 50 (vehicle position specifying system 10). As shown in FIG. 3, in this embodiment, the vehicle 9 is equipped with four cameras C each having a photographing range of the front, the rear, the left side, and the right side. However, the configuration may be such that only one or more cameras C are provided.

The vehicle control device 50 is connected by wireless communication to a control device 200 that manages the parking lot 100 . The control device 200 includes a parking lot management section 201 , a data storage section 202 and a route setting section 203 . The parking lot management unit 201 manages the vacancy of the parking space 103 by means of a photographing device (not shown) that photographs the parking space 103 and a sensor (not shown) that detects the presence or absence of the vehicle 9 in the parking space 103 . In addition, the parking lot management unit 201 manages information specifying the vehicle 9 for which the parking instruction is to be entered and the vehicle 9 for which the parking instruction is to be issued, such as identification information and position information of each vehicle 9 . The parking lot management unit 201 is also configured with a processor such as a microcomputer as a core. The data storage unit 202 stores map information of the parking lot 100, information of features G (see FIG. 3) such as a mark M serving as a recognition object OT for specifying the position of the vehicle 9 (coordinates in a reference coordinate system to be described later). (absolute coordinates), type, etc.). The data storage unit 202 is composed of a memory, a disk device, and the like.

The route setting unit 203 sets a travel route R between the alighting space 101 and the parking space 103 and a travel route R between the parking space 103 and the boarding space 102 . When there is a parking instruction, the route setting unit 203 sets a traveling route R from the alighting space 101 where the vehicle 9 targeted for the parking instruction is parked to the parking space 103 designated as the parking destination. do. The route setting unit 203 sets a travel route R from the parking space 103 in which the vehicle 9 for which the leaving instruction is issued is parked to the boarding space 102 when the leaving instruction is given.

As described above, vehicle control device 50 wirelessly communicates with control device 200 . The vehicle control device 50 receives position information of the vehicle 9 (self-vehicle position information) provided by the vehicle position specifying system 10, information on the travel route R provided by the control device 200, and information provided by the azimuth sensor 42 and the rotation sensor 43. The vehicle 9 is autonomously controlled based on the behavior information of the vehicle 9 received.

As described above, the vehicle position identification system 10 is configured with various functional units that are realized through cooperation between hardware and software. The detection information acquisition unit 1 acquires detection information around a vehicle 9 (target vehicle) whose position is to be specified. In this embodiment, the detection device that provides the detection information is the camera C, and the detection information is the captured image captured by the camera C. FIG. The detection time information acquisition unit 2 acquires detection time information indicating the detection time (“ts” described later) of the detection information. Although the details will be described later, for example, the detection time ts corresponds to the shooting time of the camera C or the like. The recognition processing unit 3 performs recognition processing of the recognition target object OT included in the detection information, and also calculates the relative position between the recognition target object OT and the vehicle 9 (target vehicle). When the detection information is a photographed image, the recognition processing by the recognition processing unit 3 is image recognition processing. Further, although the details will be described later with reference to FIG. 4 and the like, the relative position is expressed as coordinate values of the recognition target object OT in a local coordinate system with the vehicle 9 as a reference. The travel locus storage unit 4 stores a travel locus K of the vehicle 9 based on the positions (P1 to P6) of the vehicle 9 at respective times (eg, t1 to t6), as will be described later with reference to FIG.

The object information acquisition unit 5 acquires object absolute position information (absolute coordinates described later) indicating the position in the reference coordinate system of the object to be recognized OT (feature G such as mark M and parking frame line 105). get. Here, the reference coordinate system may be a coordinate system set for the entire parking lot 100, a coordinate system set for a specific region, or a coordinate system set for the entire earth. In the present embodiment, any coordinate system may be used as long as it has coordinate values that do not overlap at least anywhere in the parking lot 100 .

Further, in the present embodiment, the object information acquisition unit 5 acquires object absolute position information from the object information storage unit 30 provided in the vehicle control device 50 . The target object information storage unit 30 stores a plurality of target object absolute position information indicating respective positions in the reference coordinate system of a plurality of target objects (features G such as the mark M and the parking frame line 105) to be the recognition target object OT. (absolute coordinates) are stored. However, the object information acquisition unit 5 may acquire the object absolute position information from the control device 200 . Further, through communication between the control device 200 and the vehicle control device 50, the object absolute position information is acquired from the data storage unit 202 of the control device 200 and stored in the object information storage unit 30, and the object information acquisition unit 5 Alternatively, the object absolute position information may be acquired from the object information storage unit 30 .

As will be described later with reference to FIGS. 5 to 10, the first position specifying unit 11 obtains the relative position (local coordinates (coordinates in the local coordinate system)) and the object absolute position information (absolute coordinates) of the recognition object OT. ), the first position Ps, which is the position of the vehicle 9 (target vehicle) at the detection time ts, is specified. The second position specifying unit 12 determines the target travel trajectory KT, which is the travel trajectory K of the vehicle 9 (target vehicle) from the detection time ts to the target time tc corresponding to the current time, and the first position Ps. A second position Pc (current position), which is the position of the vehicle 9 at time tc, is identified.

As described above, the vehicle position specifying section 6 includes the position information acquisition section 7 in addition to the first position specifying section 11 and the second position specifying section 12 . The position information acquisition unit 7 acquires the position of the vehicle based on GPS signals from the GPS receiver 41, for example. The detection results of the azimuth sensor 42 and the rotation sensor 43 are also used to update the own vehicle position that changes as the vehicle 9 moves. Further, when the vehicle position is specified with high accuracy by the first position specifying unit 11 and the second position specifying unit 12, the second position Pc is acquired and the vehicle position is updated.

The identification of the vehicle position by the vehicle position identification system 10 will be described in detail below. As described above, in the present embodiment, the detection information acquisition unit 1 acquires the peripheral image of the vehicle 9 photographed by the camera C mounted on the vehicle 9 as detection information. For example, when the vehicle 9 exists at a position as shown in FIG. 3, the detection information (captured image) includes the first recognition target object OT1 and the second recognition target object OT2 as the recognition target objects OT. . These recognition objects OT are marks M arranged on the road surface (parking space 103).

The recognition processing unit 3 performs recognition processing of the recognition target object OT (OT1, OT2) included in the detection information (captured image), and as shown in FIG. The coordinates of the recognition object OT2 are detected. The local coordinate system is a two-dimensional orthogonal coordinate system having the vehicle 9 (target vehicle) as a reference and having an LX axis (local X axis (local first axis)) and a LY axis (local Y axis (local second axis)). The LX axis is a direction along the longitudinal direction of the vehicle 9. The local coordinates (LX, LY) of the reference point "P" indicating the vehicle position in the vehicle 9 are (0, 0). The unit 3 detects the local coordinates (LX ₁ , LY ₁ ) of the first recognition target object OT1 and the local coordinates (LX ₂ , LY ₂ ) of the second recognition target object OT2.

The travel control unit 20 causes the vehicle 9 to travel along the travel route R set in the parking lot 100 based on the position information (absolute coordinates) of the vehicle 9 in the reference coordinate system in the parking lot 100 . Therefore, the position information acquiring section 7 acquires the vehicle position (absolute coordinates) in the reference coordinate system. As described above, in this embodiment, the reference coordinate system is at least a coordinate system having coordinate values that do not overlap at any location in the parking lot 100 .

A form in which the first recognition target object OT1 is recognized as the recognition target object OT will be described below as an example. As shown in FIG. 4, the reference point "P" of the vehicle 9 is the origin of the local coordinate system, and the values of the local coordinates (LXp, LYp) are (0, 0). As shown in FIG. 5, the position information acquisition unit 7 acquires the absolute coordinates of this reference point "P" as (WX'p, WY'p). The coordinate values (absolute coordinates) in the reference coordinate system of the first recognition object OT1 positioned at the local coordinates (LX ₁ , LY ₁ ) are (WX'p+LX ₁ , WY'p+LY ₁ ). This coordinate is called a temporary absolute position. Here, as described above, the object absolute position information (actual coordinates (WXp, WYp) in the reference coordinate system) of the first recognition object OT1 is stored in the object information storage unit 30 .

As shown in FIG. 5, when the coordinates of the temporary absolute position (WX'p+LX ₁ , WY'p+LY ₁ ) are different from the actual coordinates (WXp, WYp) of the first recognition object OT1, There is a high possibility that the absolute coordinates (WX'p, WY'p) of the vehicle 9 deviate from the coordinate values where the vehicle 9 is actually located. Therefore, the local coordinate system is moved so that these two coordinate values (WX'p+LX ₁ , WY'p+LY ₁ ) and (WXp, WYp) are the same. That is, the absolute coordinates of the vehicle 9 are corrected by moving the reference point "P" of the vehicle 9 to (WXp, WYp). Although the details will be described later, the first position Ps is specified by correcting the reference point “P” at the detection time ts in this way by the first position specifying unit 11 .

By the way, when the coordinate values of the vehicle 9 specified in the reference coordinate system deviate from the actual coordinate values in this way, the coordinate values in the object absolute position information are also used for the recognition object OT as described above. and the coordinate values based on the recognition result are different. For this reason, when a plurality of objects (features G) that can be the recognition target object OT are provided, after specifying which object (feature G) the recognized recognition target object OT is, , the absolute coordinates of the vehicle 9 need to be corrected. When a plurality of objects are candidates for the recognition target object OT, the coordinates (temporary absolute position) specified by the relative relationship (relationship in the local coordinate system) between the recognition target object OT and the vehicle 9 and the object absolute position An object having a small deviation D from the coordinates defined by the position information is selected as the recognition object OT.

For example, as shown in FIG. 5, consider a case where a first mark M1 and a second mark M2 as features G are objects that are candidates for the recognition object OT. In this case, the deviation D (first deviation D1) between the temporary absolute position and the position of the first mark M1 is larger than the deviation D (second deviation D2) between the temporary absolute position and the position of the second mark M2. Since it is small, the first mark M1 is selected as the recognition object OT, and the vehicle position is corrected. It should be noted that if the deviation D is too large, there is a possibility that the recognition processing unit 3 will misrecognize, so it is preferable not to select any target object as the recognition target object OT.

That is, an object having object absolute position information that minimizes the deviation D of the distance between the temporary absolute position and the position indicated by the object absolute position information in the reference coordinate system is selected as the recognition object OT. At this time, if the deviation D between the temporary absolute position and the object absolute position information of the recognition object OT is equal to or less than a predetermined standard deviation, the vehicle position information at the detection time ts is the first position Ps is corrected to On the other hand, if the deviation D is greater than the reference deviation, the vehicle position information is not corrected based on the first position Ps.

If the position of the vehicle 9 in the reference coordinate system deviates from the actual position, there is a possibility that this deviation is greater than the interval at which the multiple objects are distributed. In such a case, an object different from the recognition target object OT may be selected as the recognition target object OT, and the positional error of the vehicle 9 may rather increase. Therefore, when the deviation D is larger than the reference deviation, it is preferable not to correct the vehicle position information based on the first position Ps.

By recognizing the recognition object OT and correcting the position information of the vehicle 9 in this manner, the position of the vehicle can be specified with higher accuracy. However, recognition processing such as image recognition processing is processing with a relatively large load and tends to take a long time to execute. If the execution time is long, the vehicle 9 moves during that time, so the effect of the increased accuracy of identifying the position of the vehicle is limited. Therefore, the vehicle position specifying system 10 of the present embodiment is configured so that the position of the vehicle can be specified with high accuracy regardless of the execution time of the recognition process.

As shown in FIGS. 6 and 7, the vehicle position specifying processing by the vehicle position specifying system 10 includes recognition system processing $10 which is mainly executed by the detection information acquisition unit 1, the detection time information acquisition unit 2, and the recognition processing unit 3. , a position specifying system processing $20 mainly executed by the object information acquisition unit 5, the first position specifying unit 11, and the second position specifying unit 12, and the traveling locus storage mainly executed by the traveling locus storage unit 4 Processing $30 is executed in cooperation with this. The position information acquisition unit 7 cooperates with the GPS receiver 41, the direction sensor 42, the rotation sensor 43, etc., and sequentially executes the process of acquiring the latest vehicle position specified by the second position specifying unit 12. It partially cooperates with the recognition system processing $10 and the position specifying system processing $20 by using the second position Pc. In the vehicle position specifying process, recognition system processing $10 and position specifying system process $20 are the main processes, and running locus storage process $30 is linked to these main processes. That is, the traveling locus storage unit 4 cooperates with the GPS receiver 41, the direction sensor 42, the rotation sensor 43, etc., and sequentially stores the traveling locus K in the traveling locus storage processing $30 (running locus storage step/function #31). , and partially cooperates with the recognition system processing $10 and the position specifying system processing $20.

The vehicle position specifying process illustrated in FIG. 6 and the vehicle position specifying process illustrated in FIG. 7 differ mainly in the form of acquiring the detection time ts (or the time corresponding to the detection time ts). When distinguishing the recognition system processing $10 in FIGS. 6 and 7, FIG. 6 is called the first recognition system processing $11, and FIG. 7 is called the second recognition system processing $12. 6 is referred to as first position specifying system processing $21, and FIG. 7 is referred to as second position specifying system processing $22.

In the vehicle position specifying process illustrated in FIG. 6, the time is synchronized between the recognition system processing $10 and the position specifying system process $20. Furthermore, it is preferable that the time is synchronized with the travel locus storage processing $30. That is, in the vehicle position specifying process illustrated in FIG. 6, the time is synchronized in each functional unit in the vehicle position specifying system 10 . On the other hand, in the vehicle position identification processing illustrated in FIG. 7, the recognition system processing $10, the position identification system processing $20, and the traveling locus storage processing $30 do not necessarily have to be synchronized in time.

As shown in FIGS. 6 and 7, at the beginning of the vehicle position specifying process, the detection information acquisition unit 1 acquires detection information around the target vehicle (#1: detection information acquisition step/function). In the following description, "step/function" is appropriately abbreviated as "step", but "step" means "step/function".

Next, in the first recognition system processing $11 shown in FIG. 6, the detection time information acquisition unit 2 acquires the detection time ts of the detection information (#3). This step #3 corresponds to detection time information acquisition step/function #4. For example, when the detection information is an image taken by camera C, the time taken by camera C is included in the taken image. A time (for example, time “t2” shown in FIG. 8) is obtained. In this case, it is preferable that the time of the camera C is also synchronized with each functional unit of the vehicle position specifying system 10 . However, if the camera C is a camera that shoots moving images, the image is generally shot at a frame rate of 15 frames/second, 30 frames/second, 60 frames/second, or the like. Therefore, the time difference between the time when the image was actually captured and the time when the detection information acquisition unit 1 acquired the captured image can be defined from the frame rate and the transfer time from the camera C. FIG. Therefore, the detection time information acquisition unit 2 may set the time at which the detection information acquisition unit 1 acquires the captured image as the acquisition time (t2). That is, in the first recognition system processing $11, the detection time information is the acquisition time (t2) based on the time when the detection information was detected or the time when the detection information acquisition unit 1 acquired the detection information. include. The detection time information acquisition unit 2 sets this acquisition time (t2) as the detection time ts.

On the other hand, in the second recognition system processing $12 shown in FIG. 7, the detection time information acquisition unit 2 does not acquire the acquisition time (t2), but first activates the counter (#2). The function of the counter will be described later.

After step #3 in the first recognition system processing $11 and after counter start step #2 in the second recognition system processing $12, the recognition processing unit 3 performs recognition processing of the recognition target object OT included in the detection information. At the same time, the relative position (local coordinates of the recognition target object OT) between the recognition target object OT and the vehicle 9 (target vehicle) is calculated (#5: recognition processing step/function).

In the first recognition system processing $11, when the recognition processing step #5 is completed, the processing result of the first recognition system processing $11 is output, and the processing ends. The processing result of the first recognition system processing $11 includes the detection time ts acquired in step #3 (detection time information acquisition step #4), the recognition object OT and the vehicle 9 calculated in recognition processing step #5. It is a relative position (local coordinates (LX, LY) of the recognition target object OT) with respect to (target vehicle).

In the second recognition system processing $12, when the recognition processing step #5 is completed, the counter is stopped (#6), the processing result of the second recognition system processing $12 is output, and the processing ends. As described above, the counter is started at counter start step #2 immediately before recognition processing step #5 and stopped at counter stop step #6 immediately after recognition processing step #5. Therefore, the counter value in the counter stop step #6 indicates the elapsed time PT (target elapsed time) of the recognition processing step #5. The processing result of the second recognition system processing $12 includes the elapsed time PT of the recognition processing step #5 and the relative position (recognition local coordinates (LX, LY) of the object OT. The elapsed time PT (target elapsed time) corresponds to detection time information indicating the detection time ts of the detection information, and the counter start step #2 and counter stop step #6 correspond to the detection time information acquisition step #4.

As described above, the travel locus storage unit 4 stores the travel locus K of the vehicle 9 based on the positions (P1 to P6) of the vehicle 9 at each time (t1 to t6), as shown in FIGS. A travel locus storage step #31 to be stored is separately executed.

In the first position specifying system processing $21 following the first recognition system processing $11 and the second position specifying system processing $22 following the second recognition system processing $12, first, output from each recognition system processing $10 The result obtained is obtained (#11, #12). Note that the order of step #11 and step #12 may be reversed. In the first position specifying system processing $21, the local coordinates (LX, LY) of the recognition object OT are obtained (#11), and then the detection time ts is obtained (#12A). In the second position specifying system processing $22, the local coordinates (LX, LY) of the object to be recognized OT are obtained (#11), and then the elapsed time PT is obtained (#12B).

In the first position specifying system processing $21, after steps #11 and #12, the position of the vehicle 9 at the detection time ts (own vehicle position) is acquired (#13A), and the object to be recognized OT ( The object absolute position information indicating the position of the feature G) in the reference coordinate system is obtained (#14), the first position Ps is specified as described above with reference to FIG. 9 is corrected (#15: first position specifying step/function).

In the second position specifying system processing $22, after steps #11 and #12, the position of the vehicle 9 (self-vehicle position) at the time before the current time (target time tc) before the elapsed time PT is acquired (# 13B), the object absolute position information indicating the position of the object (feature G) to be the recognition object OT in the reference coordinate system is acquired (#14), and as described above with reference to FIG. One position Ps is identified and the absolute position of the vehicle 9 at the detection time ts is corrected (#15). Considering the time from the counter stop step #6 in the second recognition system processing $12, for example, the position of the vehicle 9 at the time "PT+α" before the current time may be acquired. However, compared to the time required for recognition processing step #5, the time from counter stop step #6 and the time required for position identification processing $20 are much shorter, so the elapsed time is shorter than the current time (target time tc). There is no problem even if the position of the vehicle 9 (self-vehicle position) at the time before the PT is acquired.

8 and 9, the travel locus storage unit 4 stores the travel locus K of the vehicle 9 based on the position (vehicle position: P1 to P6) of the vehicle 9 at each time (t1 to t6). As is clear from the fact that the vehicle position and time are stored discretely. Therefore, for the "time before the elapsed time PT from the current time (target time tc)", the time closest to the time before the elapsed time PT is selected from among the times accumulated as discrete values. Become. The time selected in this manner corresponds to the detection time ts.

Of the position specifying system processing $20, the above-mentioned step #14 is an object information acquisition step for acquiring object absolute position information indicating the position of the object (feature G) to be the recognition object OT in the reference coordinate system. Equivalent to the / function. In the first position specifying system processing $21, steps #13A and #15 described above are performed by determining the relative position (local coordinates (LX, LY) of the recognition target object OT) and the absolute object position of the recognition target object OT. This corresponds to a first position specifying step/function of specifying a first position Ps, which is the position of the vehicle 9 (target vehicle) at the detection time ts, based on information (for example, absolute coordinates (WX ₁ , WY ₁ )). In addition, the above-described steps #13B and #15 of the second position specifying system processing $22 are the relative position (local coordinates (LX, LY) of the recognition target object OT) and the object absolute position of the recognition target object OT. information (for example, absolute coordinates (WX ₁ , WY ₁ )) and the position of the vehicle 9 (target vehicle) at detection time ts (“time before elapsed time PT from current time (target time)). It corresponds to the first localization step/function of identifying one position Ps.

After the object information acquisition step #14 and the first position specifying step (#13, #15), the traveling locus K is corrected and the second position Pc is specified (#16: second position specifying step/function). . That is, the second position specifying unit 12 operates based on the target travel trajectory KT, which is the travel trajectory K of the vehicle 9 (subject vehicle) from the detection time ts to the target time tc corresponding to the current time, and the first position Ps. , a second position Pc, which is the position of the vehicle 9 at the target time tc.

As described above, the first position specifying unit 11 specifies the first position Ps, which is the absolute coordinates (own vehicle position) of the vehicle 9 at the detection time ts. As a result, as shown in FIG. 10, the position "P2" of the vehicle 9 at the detection time ts stored in the travel locus storage unit 4 is corrected to the first position Ps. After the detection time ts, the position of the vehicle 9 changes with the position "P2" as the starting point. Therefore, the target running trajectory KT from the detection time ts to the current time (target time tc) is also the corrected " The trajectory is corrected to a new trajectory K (second trajectory K2) starting from the first position Ps, which is P2″. That is, the running locus storage unit 4 corrects the running locus ahead of the first position Ps based on the first position Ps. As a result, the second position Pc, which is the position of the vehicle 9 at the current time (target time tc), is also corrected from the position "P6" on the first travel locus K1 to the position "P2" on the second travel locus K2. .

It should be noted that the target time tc is preferably the time when the execution of the process in the second position identifying section 12 is started (the time when the execution of the second position identifying step #16 is started). When the target time tc is set in this manner, the processing of the second position specifying unit 12 can be executed with the earliest time at which the actual time and the position of the vehicle 9 can be obtained as the target time tc. As a result, the second position Pc can be specified with high accuracy. As described above, the execution time of the position specifying process $20 is much shorter than the execution time of the recognition process (recognition step #5). may be the time of

In this way, the vehicle position specifying system 10 can accurately specify the position of the vehicle based on the information detected around the vehicle. In addition, as in the present embodiment, when the vehicle position identification system is used for automatic valet parking in which the vehicle is automatically driven at a relatively low speed, it is possible to identify a vehicle with a high degree of accuracy and avoid other obstacles during automatic driving. A vehicle position can be identified for non-interference.

[Other embodiments]
Other embodiments will be described below.

(1) In the above description, the mark M placed on the road surface was exemplified as the feature G as the object to be recognized OT. However, even if such a mark M is not installed, as illustrated in FIG. 11, the end point 105T of the parking frame line 105 as the feature G may be used as the recognition object OT. The end points of the parking frame line 105 can be easily detected using known edge extraction or the like, and can be detected appropriately. Also, although not shown in the drawings, various planar features such as crosswalks and stop lines as features G installed on the ground surface may be used as recognition target objects OT. Further, the object to be recognized OT is not limited to a two-dimensional feature set on the ground surface as described above, but may be a three-dimensional feature G such as a guardrail, a fence, or a guide sign.

(2) In the above embodiment, the example in which the detection device that provides detection information is the camera C (image capturing device) using visible light has been described. However, the detection device is not limited to such an example, and may be a sensor using invisible light such as infrared rays and ultraviolet rays, electromagnetic waves of various wavelengths, sound waves, and the like. For example, as the detection device, various sensors such as millimeter wave radar, sonar, and LIDAR (Light Detection and Ranging/Laser Imaging Detection and Ranging) that transmit search waves to a predetermined range around the vehicle 9 can be used. Even when a sensor other than the camera C is used as the detection device, the detection information acquisition unit 1 acquires detection information around the vehicle 9 regarding the detection target of the sensor. In this case, the recognition target object OT included in the detection information is an object that can be detected by the sensor.

(3) It should be noted that the configurations disclosed in the respective embodiments described above can be applied in combination with configurations disclosed in other embodiments as long as there is no contradiction. Regarding other configurations, the embodiments disclosed in this specification are merely examples in all respects. Therefore, various modifications can be made as appropriate without departing from the scope of the present disclosure.

[Overview of embodiment]
An outline of the vehicle position specifying system (10), the vehicle position specifying method, and the vehicle position specifying program described above will be briefly described below.

As one aspect, a vehicle position specifying system (10) includes a detection information acquisition unit (1) that acquires detection information around a target vehicle (9), and detection time information that indicates a detection time (ts) of the detection information. A detection time information acquisition unit (2) that acquires the recognition target object (OT) included in the detection information and performs recognition processing of the recognition target object (OT) and the relative position between the recognition target object (OT) and the target vehicle (9) and a recognition processing unit (3) that calculates the travel locus (K) of the target vehicle (9) based on the position (P1 to P6) at each time (t1 to t6) of the target vehicle (9). a trajectory storage unit (4); an object information acquisition unit (5) for acquiring object absolute position information indicating the position of the recognition object (OT) in a reference coordinate system; and the relative position and the recognition object. a first position specifying unit (11) that specifies a first position (Ps) that is the position of the target vehicle (9) at the detection time (ts) based on the target object absolute position information; Based on the target travel trajectory (KT), which is the travel trajectory (K) of the target vehicle (9) from ts) to the target time (tc) corresponding to the current time, and the first position (Ps), and a second position specifying unit (12) that specifies a second position (Pc), which is the position of the target vehicle (9) at the target time (tc).

According to this configuration, the position (first position (Ps)) of the target vehicle (9) is identified by going back to the detection time (ts) at which the detection information was detected, and the travel locus ( K), and the position (second position (Pc)) of the target vehicle (9) at the target time (tc) corresponding to the current time based on the target travel trajectory (KT) and the first position (Ps). be done. Therefore, even if the recognition process for the detection information takes time, it is possible to suppress the error of the second position (Pc) caused by the movement of the target vehicle (9) during that time. That is, according to this configuration, it is possible to accurately identify the position of the vehicle based on the information detected around the vehicle.

The technical features of the vehicle localization system (10) described above can also be applied to vehicle localization directions and vehicle localization programs. A representative embodiment thereof is illustrated below. For example, the vehicle localization method has various steps with features of the vehicle localization system (10) described above. Further, the vehicle position specifying program causes the computer to implement various functions having the features of the vehicle position specifying system (10) described above. Of course, these vehicle position specifying method and vehicle position specifying program can also produce the effects of the vehicle position specifying system described above. Furthermore, as preferred aspects of the vehicle position identification system (10), various additional features exemplified below can also be incorporated into these vehicle position identification methods and vehicle position identification programs, and the method and program can be Effects corresponding to each additional feature can also be exhibited.

As one preferred aspect, the vehicle position identification method includes a detection information acquisition step (#1) of acquiring detection information around a target vehicle (9), and detection time information indicating a detection time (ts) of the detection information. a detection time information acquisition step (#4) for acquiring the recognition target object (OT) included in the detection information; A recognition processing step (#5) for calculating a position, and storing the travel locus (K) of the target vehicle (9) based on the positions (P1 to P6) of the target vehicle (9) at each time (t1 to t6). an object information acquisition step (#14) for acquiring object absolute position information indicating the position of the recognition object (OT) in the reference coordinate system; and the relative position and the recognition object a first position specifying step (#13, #15) for specifying a first position, which is the position of the target vehicle at the detection time, based on the object absolute position information; Based on the target travel trajectory (KT), which is the travel trajectory (K) of the target vehicle (9) up to the target time (tc) corresponding to the target time (tc), and the first position (Ps), the target time (tc ), and a second position specifying step (#16) of specifying a second position (pc), which is the position of the target vehicle (9) in ).

Further, as one preferred aspect, the vehicle position specifying program includes a detection information acquisition function (#1) for acquiring detection information around the target vehicle (9), and a detection function indicating the detection time (ts) of the detection information. A detection time information acquisition function (#4) that acquires time information, performs recognition processing of a recognition target (OT) included in the detection information, and recognizes the recognition target (OT) and the target vehicle (9). A recognition processing function (#5) that calculates the relative position of the target vehicle (9), and the travel locus (K) of the target vehicle (9) based on the positions (P1 to P6) at each time (t1 to t6) of the target vehicle (9) an object information acquisition function (#14) for acquiring object absolute position information indicating the position of the object to be recognized (OT) in the reference coordinate system; the relative position and the object to be recognized; A first position identifying function (#13, #15), the target travel trajectory (KT), which is the travel trajectory (K) of the target vehicle (9) from the detection time (tc) to the target time (tc) corresponding to the current time, and the first a second position specifying function (#16) for specifying a second position (Pc), which is the position of the target vehicle (9) at the target time (tc), based on the position (Ps) and the position (Ps). Let

As one aspect, the detection time information includes acquisition time information based on the time when the detection information is detected or the time when the detection information acquisition unit acquires the detection information, and the running locus storage. A unit (4) is based on the position (P1 to P6) at each time (t1 to t6) defined by the same reference time as the detection device (C) that detects the detection information or the detection information acquisition unit (1). and the target travel locus (KT) is preferably the travel locus (K) from the acquisition time to the target time (tc).

According to this configuration, the travel locus (K) is determined based on the positions (P1 to P6) at each time (t1 to t6) defined by the same reference time as the detection device (C) or the detection information acquisition unit (1). Since it is stored, the position (first position (Ps)) of the target vehicle (9) is specified by going back to the detection time (ts) at which the detection information was detected appropriately based on the acquisition time, and the target time (tc ) of the target vehicle (9) (second position (Pc)).

Further, as another aspect, the detection time information is the elapsed time from the time when the detection information is detected or the time when the detection information acquisition unit acquires the detection information to the target time (tc). The target running trajectory (KT) includes time (PT) information, and the target running trajectory (KT) is the running trajectory (K ) is preferred.

According to this configuration, only the target elapsed time (PT), which is the elapsed time including the recognition process, from the time when the detection information is detected or the time when the detection information is acquired by the detection information acquisition unit to the target time (tc) , the position (first position (Ps)) of the target vehicle (9) can be retroactively specified, and the position (second position (Pc)) of the target vehicle (9) at the target time (tc) can be specified. Therefore, even if the recognition process takes time, it is possible to suppress the error of the second position (Pc) caused by the movement of the target vehicle (9) during that time.

Further, as described above, the detection time information is the elapsed time from the time when the detection information is detected or the time when the detection information acquisition unit acquires the detection information to the target time (tc). When time (PT) information is included, the target time (tc) is preferably the time when the second position specifying unit (12) starts executing the process.

According to this configuration, the processing of the second position specifying unit (12) can be executed with the target time (tc) set to the latest time at which the actual time and the position of the target vehicle (9) can be acquired. As a result, the second position (Pc) can be specified with high accuracy.

Further, the object information acquisition unit (5) can acquire a plurality of object absolute position information indicating respective positions of a plurality of objects to be the recognition object (OT) in the reference coordinate system. , the first position specifying unit (11), wherein the vehicle position information indicating the position of the target vehicle (9) at the detection time (tc) indicates the position of the target vehicle (9) in the reference coordinate system. , based on the relative position, set the temporary absolute position of the recognition object (OT) in the reference coordinate system, and the deviation of the distance between the temporary absolute position and the object absolute position information in the reference coordinate system ( D) is selected as the recognition target object (OT), the target object having the target absolute position information that minimizes D), and the relative position and the target absolute position information of the selected recognition target object (OT) are selected. It is preferable to specify the first position (Ps) based on the positional information.

The relative positions of the object to be recognized (OT) and the target vehicle (9) are considered to have relatively high accuracy when the object to be recognized (OT) is positioned near the target vehicle (9). . Here, if the position of the target vehicle (9) in the reference coordinate system deviates from the actual position, the position of the object to be recognized (OT) in the reference coordinate system will also deviate. When there are a plurality of objects (G) to be recognition objects (OT), it is necessary to determine which object (G) is the recognition object (OT). In many cases, the position of the object vehicle (9) in the reference coordinate system and the actual position are considered to be smaller than the intervals in which the objects (G) are distributed. Therefore, as described above, the object (G) having the object absolute position information that minimizes the deviation (D) of the distance between the temporary absolute position and the object absolute position information in the reference coordinate system is defined as the recognition object ( OT), the first position (Ps) can be appropriately specified.

Here, the object having the object absolute position information that minimizes the deviation (D) of the distance between the temporary absolute position and the object absolute position information in the reference coordinate system is defined as the recognition object (OT ), the detection time (ts ) is corrected to the first position (Ps), and if the deviation (D) is greater than the reference deviation, the correction of the vehicle position information based on the first position (Ps) is performed. It is preferable not to do so.

If the position of the target vehicle (9) in the reference coordinate system deviates from the actual position, there is a possibility that this deviation is larger than the interval over which the multiple targets (G) are distributed. In such a case, an object (G) different from the recognition object (OT) may be selected as the recognition object (OT), and the positional error of the target vehicle (9) may rather increase. have a nature. Therefore, when the deviation (D) is larger than the reference deviation, it is preferable not to correct the vehicle position information based on the first position (Ps).

Further, it is preferable that the running locus storage unit (4) corrects the running locus (K) ahead of the first position (Ps) based on the first position (Ps).

When the position of the target vehicle (9) at the detection time (ts) is corrected to the first position (Ps), the starting point of the traveling locus (K) ahead of the detection time (ts) is at the first position (Ps). It will change. When the travel locus (K) ahead of the first position (Ps) is corrected based on the first position (Ps), the position of the target vehicle (9) at the target time (tc) is also corrected accordingly. , and the second position specifying unit 12 can appropriately specify the second position (Pc).

1: Detection information acquisition unit 2: Detection time information acquisition unit 3: Recognition processing unit 4: Traveling locus storage unit 5: Object information acquisition unit 9: Vehicle (target vehicle)
10: Vehicle position specifying system 11: First position specifying unit 12: Second position specifying unit 105T: End point C of parking frame line: Camera (detection device)
D: deviation G: feature (object)
K: Traveling locus KT: Target traveling locus M: Mark (object)
OT: Recognition target object PT: Elapsed time (target elapsed time)
Pc: Second position Ps: First position Tc: Target time tc: Target time ts: Detection time #1: Detection information acquisition step/function #4: Detection time information acquisition step/function #5: Recognition processing step/function # 31: Traveling locus storage step/function #14: Object information acquisition step/function #13, #15: First position specifying step/function #16: Second position specifying step/function

Claims (8)

a detection information acquisition unit that acquires detection information around the target vehicle;
a detection time information acquisition unit that acquires detection time information indicating the detection time of the detection information;
a recognition processing unit that performs recognition processing of a recognition target included in the detection information and calculates a relative position between the recognition target and the target vehicle;
a travel locus storage unit that stores a travel locus of the target vehicle based on the position of the target vehicle at each time;
an object information acquisition unit that acquires object absolute position information indicating the position of the recognition object in the reference coordinate system;
a first position identifying unit that identifies a first position, which is the position of the target vehicle at the detection time, based on the relative position and the object absolute position information of the recognition object;
A second position, which is the position of the target vehicle at the target time, based on the target travel trajectory, which is the travel trajectory of the target vehicle from the detection time to the target time corresponding to the current time, and the first position. and a second position specifying unit that specifies the
The object information acquisition unit is capable of acquiring a plurality of object absolute position information indicating respective positions of a plurality of objects to be recognized objects in the reference coordinate system,
The first position specifying unit assumes that vehicle position information indicating the position of the target vehicle at the detection time indicates the position of the target vehicle in the reference coordinate system, and determines the position of the recognition target object based on the relative position. setting a temporary absolute position in the reference coordinate system;
selecting, as the recognition object, the object having the object absolute position information that minimizes the deviation of the distance between the temporary absolute position and the object absolute position information in the reference coordinate system;
A vehicle position identification system that identifies the first position based on the relative position and the absolute object position information of the selected recognition object. The detection time information includes acquisition time information based on the time when the detection information is detected or the time when the detection information acquisition unit acquires the detection information,
The running locus storage unit stores the running locus based on the position at each time defined by the same reference time as the detection device that detects the detection information or the detection information acquisition unit,
The vehicle position specifying system according to claim 1, wherein the target travel locus is the travel locus from the acquisition time to the target time. The detection time information includes information on a target elapsed time, which is the elapsed time from the time when the detection information is detected or the time when the detection information acquisition unit acquires the detection information to the target time,
2. The vehicle position specifying system according to claim 1, wherein the target travel locus is the travel locus from a time before the target elapsed time before the target time to the target time. The vehicle position specifying system according to any one of claims 1 to 3, wherein the target time is a time when execution of processing in the second position specifying unit is started. correcting the vehicle position information at the detection time to the first position when the deviation between the temporary absolute position and the object absolute position information of the recognition object is equal to or less than a predetermined reference deviation; 5. The vehicle position specifying system according to any one of claims 1 to 4 , wherein the vehicle position information based on the first position is not corrected when the deviation is greater than the reference deviation. The travel locus storage unit corrects the position of the target vehicle at the detection time stored in the travel locus storage unit to the first position, and changes the target travel locus to the position at the detection time before correction. 6. The traveling locus ahead of the first position is corrected by correcting the traveling locus starting from the position of the target vehicle to the traveling locus starting from the first position. A vehicle localization system according to any one of the preceding claims. a detection information acquisition step in which the detection information acquisition unit acquires detection information around the target vehicle;
A detection time information obtaining step in which a detection time information obtaining unit obtains detection time information indicating a detection time of the detection information;
a recognition processing step in which a recognition processing unit performs recognition processing of a recognition target object included in the detection information and calculates a relative position between the recognition target object and the target vehicle;
a travel locus storage step in which a travel locus storage unit stores a travel locus of the target vehicle based on the position of the target vehicle at each time;
A target object information acquiring unit capable of acquiring a plurality of target absolute position information indicating respective positions of a plurality of target objects to be the recognition target objects in the reference coordinate system, is capable of acquiring the positions of the recognition target objects in the reference coordinate system . an object information acquiring step of acquiring the object absolute position information indicating
a first position specifying step in which a first position specifying unit specifies a first position, which is the position of the target vehicle at the detection time, based on the relative position and the object absolute position information of the recognition object;
A second position specifying unit determines the position of the target vehicle at the target time based on the target travel trajectory, which is the travel trajectory of the target vehicle from the detection time to the target time corresponding to the current time, and the first position. a second position specifying step of specifying a second position, which is a position ;
In the first locating step,
Assuming that the vehicle position information indicating the position of the target vehicle at the detection time indicates the position of the target vehicle in the reference coordinate system, a provisional absolute position of the recognition object in the reference coordinate system is calculated based on the relative position. and set
selecting, as the recognition object, the object having the object absolute position information that minimizes the deviation of the distance between the temporary absolute position and the object absolute position information in the reference coordinate system;
A method of identifying a position of a vehicle, wherein the first position is identified based on the relative position and the absolute object position information of the selected recognition object. A detection information acquisition function that acquires detection information around the target vehicle,
A detection time information acquisition function for acquiring detection time information indicating the detection time of the detection information;
a recognition processing function for performing recognition processing of a recognition target included in the detection information and calculating a relative position between the recognition target and the target vehicle;
a running locus storage function for storing a running locus of the target vehicle based on the position of the target vehicle at each time;
an object information acquisition function for acquiring object absolute position information indicating the position of the recognition object in the reference coordinate system;
a first position specifying function for specifying a first position, which is the position of the target vehicle at the detection time, based on the relative position and the object absolute position information of the recognition object;
A second position, which is the position of the target vehicle at the target time, based on the target travel trajectory, which is the travel trajectory of the target vehicle from the detection time to the target time corresponding to the current time, and the first position. A second position specifying function that specifies the
The object information acquisition function is capable of acquiring a plurality of object absolute position information indicating respective positions of a plurality of objects to be recognized objects in the reference coordinate system,
In the first position specifying function,
Assuming that the vehicle position information indicating the position of the target vehicle at the detection time indicates the position of the target vehicle in the reference coordinate system, a provisional absolute position of the recognition target object in the reference coordinate system is calculated based on the relative position. and set
selecting, as the recognition object, the object having the object absolute position information that minimizes the deviation of the distance between the temporary absolute position and the object absolute position information in the reference coordinate system;
A vehicle position specifying program for specifying the first position based on the relative position and the object absolute position information of the selected recognition object.

Images