Object recognition device and object recognition method

Abstract

The object recognition device of the present invention includes a prediction processing unit, a temporary setting unit, and a correlation processing unit. The prediction processing unit predicts the position of the moving destination of the tracking object as a predicted position on the object model obtained after modeling the tracking object based on the trajectory formed by at least one object among multiple objects moving as the tracking object. The temporary setting unit sets the position of at least one candidate point on the object model based on the specifications of the sensor that detects the tracking object. The correlation processing unit determines a reference position on the object model based on the position of the candidate point and the predicted position. The correlation processing unit determines whether there is a correlation between the position of the detection point and the predicted position based on the positional relationship between the correlation range set based on the reference position on the object model and the detection point when the sensor detects at least one of the multiple objects.

Description

Technical Field

[0001] The invention relates to an object recognition device and an object recognition method. Background Art

[0002] In the past, an object recognition device is known that matches the position of a detection point when a sensor detects an object with a shape model of the object, and determines the position of a route point that forms the object's route based on the position of the detection point in the shape model of the object (for example, refer to patent document 1).

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent Application Publication No. 2017-215161 Summary of the invention

[0006] Technical problem to be solved by the invention

[0007] It is known that in the conventional object recognition device described in Patent Document 1, whether the position of the object's destination is correlated with the position of the detection point is determined based on whether the position of the detection point is included in a correlation range set around the position of the object's destination.

[0008] However, in the conventional object recognition device shown in Patent Document 1, the correlation range may not be set correctly depending on the resolution of the sensor. In this case, the determination of whether the position of the object's destination and the position of the detection point are correlated may be erroneous. Therefore, the accuracy of the object route data indicating the position of the object's route point is reduced.

[0009] The object of the present invention is to solve the above-mentioned problem and to obtain an object recognition device and an object recognition method that can improve the accuracy of object route data.

[0010] Technical means for solving technical problems

[0011] The object recognition device of the present invention includes: a prediction processing unit, which predicts the position of the moving destination of the tracking object as a predicted position on the object model obtained after modeling the tracking object based on a trajectory formed by at least one of the multiple objects moving as the tracking object; a temporary setting unit, which sets the position of at least one candidate point on the object model based on the specifications of the sensor that detects the tracking object; and a correlation processing unit, which determines a reference position on the object model based on the position of the candidate point and the predicted position, and determines whether there is a correlation between the position of the detection point and the predicted position based on the positional relationship between a correlation range set based on the reference position on the object model and a detection point when the sensor detects at least one of the multiple objects.

[0012] Effects of the Invention

[0013] According to the object recognition device of the present invention, the accuracy of object route data can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a block diagram showing a functional configuration example of a vehicle control system according to the present embodiment.

[0015] Figure 2 It is shown Figure 1 A diagram of an example of the relative position relationship between a sensor and an object.

[0016] Figure 3 It means to become Figure 2 A diagram of an example of a first candidate point of a candidate for a position of a detection point in a vehicle.

[0017] Figure 4 It means to become Figure 2 A diagram of an example of a candidate point of a second candidate for the position of a detection point in a vehicle.

[0018] Figure 5 FIG. 1 is a diagram showing an example of another candidate point of the position of the detection point in the vehicle.

[0019] Figure 6 It means when N is a natural number Figures 3 to 5 A diagram showing an example of setting the reliability of candidate points.

[0020] Figure 7 It is shown Figure 1 A plot of an example of the predicted data.

[0021] Figure 8 Is based on Figure 7A diagram showing an example of a reference position determined based on a predicted position of prediction data and a candidate point.

[0022] Fig. 9 It is shown that Figure 8 A diagram of a first setting example of a related range set based on a reference position of .

[0023] Fig.10 It is shown that Figure 8 A diagram of a second setting example of a related range set based on a reference position of .

[0024] Fig.11 It is shown that Figure 8 A diagram of a third setting example of a related range set based on a reference position of .

[0025] Fig.12 It means in Figure 7 An example diagram of the route data also containing directions.

[0026] Fig.13 It means in Figure 7 The route data also contains an example graph of altitude.

[0027] Fig.14 It is shown Figure 7 The route data also includes an example diagram of the upper end position and the lower end position.

[0028] Fig.15 This is a schematic illustration of Figure 2 The object model of the relevant judgment object centered at the position of the detection point is Figure 8 Overlapping graph of object models of tracked objects centered at the predicted position.

[0029] Fig.16 It is explained by Figure 1 Flowchart of the processing performed by the object recognition device.

[0030] Fig.17 Yes Description Fig.16 Flowchart of the relevant association processing in step S19.

[0031] Fig.18 Yes Description Fig.17 Flowchart of the relevant range setting process in step S38.

[0032] Fig.19 Yes Description Fig.16 Flowchart of the correlation determination process in step S20.

[0033] Fig. 20 Yes Description Fig.19 Flowchart of the rationality determination process in step S75.

[0034] Fig.21 It is a diagram for explaining an example of hardware configuration.

[0035] Fig. 22 It is a diagram for explaining another hardware configuration example. DETAILED DESCRIPTION

[0036] Figure 1 1 is a block diagram showing an example of the functional configuration of a vehicle control system according to the present embodiment. Figure 1 As shown, the vehicle control system includes a plurality of vehicle exterior information sensors 1 , a plurality of vehicle information sensors 2 , an object recognition device 3 , a notification control device 4 and a vehicle control device 5 .

[0037] Each of the plurality of exterior information sensors 1 is installed on the vehicle. For example, among the plurality of exterior information sensors 1, some of the exterior information sensors 1 are installed individually inside the front bumper, inside the rear bumper, and on the cabin side of the windshield. The exterior information sensor 1 installed inside the front bumper observes objects located in front of or on the side of the vehicle C. The exterior information sensor 1 installed inside the rear bumper observes objects located behind or on the side of the vehicle C.

[0038] The exterior information sensor 1 installed on the cabin side of the windshield is arranged beside the interior rearview mirror. The exterior information sensor 1 installed beside the interior rearview mirror on the cabin side of the front windshield observes objects in front of the vehicle C.

[0039] Thus, each of the plurality of vehicle exterior information sensors 1 mounted on the vehicle is a sensor capable of acquiring information about objects around the vehicle as detection data dd. The respective detection data dd about objects around the vehicle acquired by each of the plurality of vehicle exterior information sensors 1 are integrated and generated as detection data DD. The detection data DD is generated as a data structure that can be provided to the object recognition device 3. The detection data DD includes at least one information related to the position P of at least one detection point DP.

[0040] The vehicle exterior information sensor 1 observes an object by detecting any point on the surface of the object as a detection point DP. Each detection point DP represents a point in the object observed by the vehicle exterior information sensor 1 around the vehicle. For example, the vehicle exterior information sensor 1 irradiates the surroundings of the vehicle with light as irradiation light, and receives reflected light reflected at each reflection point of the object. Each reflection point corresponds to each detection point DP.

[0041] In addition, according to the measurement principle of the vehicle exterior information sensor 1 , the information related to the object observable at the detection point DP is different.

[0042] As the type of the vehicle exterior information sensor 1, a millimeter wave radar, a laser sensor, an ultrasonic sensor, an infrared sensor, a camera, etc. can be used. In addition, the description of the ultrasonic sensor and the infrared sensor is omitted.

[0043] For example, the millimeter-wave radar is installed on the front bumper and the rear bumper of the vehicle, respectively. The millimeter-wave radar has a transmitting antenna and multiple receiving antennas. The millimeter-wave radar can measure the distance and relative speed to an object. For example, the distance and relative speed to an object are measured by the FMCW (Frequency Modulation Continuous Wave) method. Therefore, the position P of the detection point DP and the speed V of the detection point DP can be observed based on the distance and relative speed to the object measured by the millimeter-wave radar.

[0044] In the following description, the speed V at the detection point DP may be a relative speed between the host vehicle and the object, or may be a speed based on an absolute position by further utilizing GPS.

[0045] The millimeter wave radar can measure the azimuth of an object. The azimuth of the object is measured based on the phase difference of each radio wave received by each of the multiple receiving antennas. Therefore, the direction θ of the object can be observed based on the azimuth of the object measured by the millimeter wave radar.

[0046] Thus, using the millimeter wave radar, in addition to the position P of the detection point DP, detection data DD including the speed V of the detection point DP and the direction θ of the object can be observed as information related to the object. Among the position P of the detection point DP, the speed V of the detection point DP, and the direction θ of the object, the speed V of the detection point DP and the direction θ of the object are dynamic elements that determine the state of the object. Each of these dynamic elements is an object determination element.

[0047] In addition, in the FMCW millimeter wave radar, when measuring the relative speed with an object, the frequency shift caused by the Doppler effect between the frequency of the transmitted signal and the frequency of the received signal, that is, the Doppler frequency, is detected. Since the detected Doppler frequency is proportional to the relative speed with the object, the relative speed can be derived from the Doppler frequency.

[0048] In addition, the velocity resolution of millimeter wave radar is determined by the resolution of Doppler frequency. The resolution of Doppler frequency is the inverse of the observation time of the received signal. Therefore, the longer the observation time, the higher the resolution of Doppler frequency. Therefore, the longer the observation time, the higher the velocity resolution of millimeter wave radar.

[0049] For example, when the vehicle is traveling on a highway, the observation time of the millimeter wave radar is set longer than when the vehicle is traveling on a normal road. Therefore, the speed resolution of the millimeter wave radar can be set higher. Therefore, when the vehicle is traveling on a highway, the change in speed can be observed more quickly than when the vehicle is traveling on a normal road. As a result, objects around the vehicle can be observed more quickly.

[0050] In addition, the distance resolution of millimeter wave radar is defined as the speed of light divided by the modulation bandwidth. Therefore, the wider the modulation bandwidth, the higher the distance resolution of millimeter wave radar.

[0051] For example, when the vehicle is driving in a parking lot, the modulation bandwidth is set wider than when the vehicle is driving on an ordinary road or a highway. As a result, the distance resolution of the millimeter-wave radar can be set higher. When the distance resolution of the millimeter-wave radar is set higher, the minimum unit distance that can be detected around the vehicle becomes finer, so that adjacent objects can be distinguished from each other.

[0052] For example, when pedestrians and vehicles C exist as objects around the vehicle, pedestrians with low reflection intensity and vehicles C with high reflection intensity are mixed with each other with respect to electromagnetic waves irradiated from the millimeter wave radar. Even in this state, the electromagnetic waves reflected from pedestrians are not absorbed by the electromagnetic waves reflected from vehicles C, so pedestrians can be detected.

[0053] The laser sensor is installed, for example, on the outside of the vehicle roof. As a laser sensor, for example, LIDAR (Light Detection And Ranging) is installed on the outside of the vehicle roof. LIDAR has a plurality of light-emitting units, a light-receiving unit, and a computing unit. The plurality of light-emitting units are arranged at a plurality of angles in a direction perpendicular to the front of the moving direction of the vehicle.

[0054] LIDAR uses TOF (Time Of Flight) method. Specifically, the multiple light-projecting units in LIDAR have the function of rotating in the horizontal direction and radially projecting laser light during a preset light-projecting time. The photosensitive unit in LIDAR has the function of receiving reflected light from an object during a preset photosensitive time. The calculation unit in LIDAR has the function of calculating the round-trip time that is the difference between the light-projecting moment in the multiple light-projecting units and the photosensitive moment in the photosensitive unit. The calculation unit in LIDAR has the function of calculating the distance to the object based on the round-trip time.

[0055] LIDAR also has the function of determining the direction to an object by finding the distance to the object. Therefore, the position P of the detection point DP, the speed V of the detection point DP, and the direction θ of the object can be observed based on the measurement results measured by LIDAR.

[0056] Thus, by using LIDAR, in addition to the position P of the detection point DP, the detection data DD including the speed V of the detection point DP and the direction θ of the object can be observed as information related to the object. As described above, among the position P of the detection point DP, the speed V of the detection point DP and the direction θ of the object, the speed V of the detection point DP and the direction θ of the object are respectively object identification elements.

[0057] In addition, the speed resolution of LIDAR is determined by the emission interval of the pulses constituting the laser. Therefore, the shorter the emission interval of the pulses constituting the laser, the higher the speed resolution of LIDAR.

[0058] For example, when the vehicle is traveling on a highway, the speed resolution of the LIDAR can be set higher by setting the light emission interval of the pulses of laser light emitted from the LIDAR to be shorter than when the vehicle is traveling on a normal road. Therefore, when the vehicle is traveling on a highway, the change in speed can be observed more quickly than when the vehicle is traveling on a normal road. As a result, objects around the vehicle can be observed more quickly.

[0059] In addition, the distance resolution of LIDAR is determined by the pulse width of the laser light. Therefore, the shorter the pulse width of the laser light, the higher the distance resolution of LIDAR.

[0060] For example, when the vehicle is traveling in a parking lot, the pulse width of the laser light emitted from the LIDAR is set shorter than when the vehicle is traveling on a normal road or a highway. As a result, the distance resolution of the LIDAR can be set higher. When the distance resolution of the LIDAR is set higher, the minimum unit distance that can be detected around the vehicle becomes finer, so that objects arranged adjacently can be distinguished from each other.

[0061] For example, when pedestrians and vehicles C exist as objects around the host vehicle, pedestrians with low reflection intensity and vehicles C with high reflection intensity are mixed with each other with respect to the laser irradiated from the LIDAR. Even in this state, the reflected light from the pedestrians is not absorbed by the reflected light from the vehicles C, so the pedestrians can be detected.

[0062] The camera is installed next to the interior rearview mirror on the cabin side of the front windshield. For example, a monocular camera is used as the camera. The monocular camera has a shooting element. The shooting element is, for example, a CCD (Charge Coupled Device) image sensor or a CMOS (Complementary Metal Oxide Semiconductor) image sensor. The monocular camera continuously detects the presence and distance of an object using a pixel level in a two-dimensional space orthogonal to the shooting direction of the shooting element as a minimum unit. For example, the monocular camera includes a structure in which primary color filters of red, green, and blue are added to a lens. With such a structure, the distance can be calculated based on the parallax of the light divided using the primary color filters. Therefore, the position P of the detection point DP, the width W, and the length L of the object are observed based on the measurement results measured by the camera.

[0063] As described above, according to the camera, in addition to the position P of the detection point DP, the detection data DD including the width W and length L of the object can be observed as information related to the object. Among the position P of the detection point DP, the width W and length L of the object, the width W and length L of the object are static elements for determining the size of the object. Each of these static elements is an object determining element.

[0064] In addition, in addition to single-lens cameras, TOF cameras, stereo cameras, infrared cameras, etc. are also used.

[0065] The plurality of vehicle information sensors 2 have a function of detecting vehicle information of the vehicle such as vehicle speed, steering angle, and yaw rate as own vehicle data cd. The own vehicle data cd is generated in a data structure that can be provided to the object recognition device 3 .

[0066] The object recognition device 3 includes a time measurement unit 31, a data receiving unit 32, a temporary setting unit 33, a prediction processing unit 34, a correlation processing unit 35, and an update processing unit 36. The time measurement unit 31, the data receiving unit 32, the temporary setting unit 33, the prediction processing unit 34, the correlation processing unit 35, and the update processing unit 36 ​​have functions implemented by a CPU that executes a program stored in a nonvolatile memory or a volatile memory.

[0067] The time measuring unit 31 has a function of measuring the time of the object recognition device 3 . The time measuring unit 31 generates the measured time as a common time CT. The common time CT is generated in a data structure that can be provided to the data receiving unit 32 .

[0068] The data receiving unit 32 has a function of an input interface.

[0069] Specifically, the data receiving unit 32 has a function of receiving detection data dd from each vehicle exterior information sensor 1. Each detection data dd is integrated into detection data DD by the data receiving unit 32. The data receiving unit 32 generates detection data DD by associating the common time CT generated by the time measuring unit 31 with the detection data DD as the associated time RT. RT Function. Detection data DD RT The data structure is generated so as to be able to be provided to the temporary setting unit 33 and the correlation processing unit 35 respectively.

[0070] When the detection data dd is received from the vehicle exterior information sensor 1, the data receiving unit 32 determines that the detection data dd can be acquired. The data receiving unit 32 sets the failure flag indicating that a failure has occurred in the corresponding vehicle exterior information sensor 1 to 0, and generates the detection data DD. RT .

[0071] Here, when the failure flag is set to 0, it indicates that the corresponding vehicle exterior information sensor 1 has no failure. In addition, when the failure flag is set to 1, it indicates that the corresponding vehicle exterior information sensor 1 has a failure.

[0072] On the other hand, when the detection data dd is not received from the vehicle exterior information sensor 1, the data receiving unit 32 determines that the detection data dd cannot be obtained, sets the failure flag to 1, and generates the detection data DD. RT .

[0073] When the detection data dd is received from the vehicle exterior information sensor 1, the data receiving unit 32 determines the rationality of the detection data dd. When it is determined that the detection data dd is not rational, the data receiving unit 32 determines that the detection data dd cannot be obtained, and sets the data rationality flag indicating that the detection data dd corresponding to the corresponding vehicle exterior information sensor 1 is not rational to 0. When the data receiving unit 32 determines that the detection data dd is rational, it determines that the detection data dd can be obtained, and sets the above data rationality flag to 1.

[0074] In this way, the determination result of whether the detection data dd is acquired by the data receiving unit 32 can be referred to by referring to at least one of the failure flag and the data rationality flag.

[0075] The data receiving unit 32 has a function of receiving the own vehicle data cd from the vehicle information sensor 2. The data receiving unit 32 generates the own vehicle data CD by associating the common time CT generated by the time measuring unit 31 with the own vehicle data cd as the associated time RT. RT This vehicle data CD RT A data structure is generated that can be provided to the prediction processing unit 34 .

[0076] The temporary setting unit 33 has a function of setting an object model C obtained by modeling the tracking object based on the resolution of the vehicle exterior information sensor 1 that detects at least one object among the plurality of objects as the tracking object. model1 The temporary setting unit 33 has a function of generating temporary setting data DH including the position HP of at least one candidate point DPH. The temporary setting data DH is generated into a data structure that can be provided to the correlation processing unit 35 by the temporary setting unit 33.

[0077] The resolution of the exterior information sensor 1 is included in the specification of the exterior information sensor 1. In addition, the resolution of the exterior information sensor 1 is different according to the specification of each exterior information sensor 1. According to the specification of the exterior information sensor 1, the attributes related to the action setting of the exterior information sensor 1 and the attributes related to the configuration state of the exterior information sensor 1 are determined. The attributes related to the action setting of the exterior information sensor 1 are the observable measurement range, the resolution of the measurement range, the sampling frequency, etc. The attributes related to the configuration state of the exterior information sensor 1 are the possible configuration angle of the exterior information sensor 1, the durable ambient temperature of the exterior information sensor 1, the measurable distance between the exterior information sensor 1 and the observed object, etc.

[0078] The prediction processing unit 34 receives the vehicle data CD from the data receiving unit 32. RT The prediction processing unit 34 has the function of receiving the route data TD from the update processing unit 36. RT-1 Function. Transmit route data TD RT-1 The prediction processing unit 34 has the vehicle data CD based on the relevant time RT. RT and the route data TD of the associated time RT-1 RT-1 , the route data TD associated with the time RT is generated by a known algorithm RT Prediction data TD RTpred A well-known algorithm is, for example, an algorithm such as a Kalman filter that can estimate the center point of an object that changes in time series based on observation values.

[0079] That is, the prediction processing unit 34 predicts the position of the movement destination of the tracking object as the object model C obtained by modeling the tracking object based on the trajectory formed by at least one object among the plurality of objects moving as the tracking object. model1 The predicted position PredP is included in the predicted data TD RTpredThe predicted position PredP is the position of the predicted point Pred. The predicted point Pred is set on the object model C model1 Therefore, the predicted position PredP is set at the center of the object model C model1 center.

[0080] The correlation processing unit 35 receives the detection data DD RT , temporary setting data DH including the position HP of the candidate point DPH, and route data TD RT Prediction data TD RTpred The correlation processing unit 35 has the function of determining the detection data DD RT and the forecast data TD of the route data TD RTpred Whether it is related. Use SNN (Simple Nearest Neighbor: simple nearest neighbor) algorithm, GNN (Global Nearest Neighbor: global nearest neighbor) algorithm, JPDA (Joint Probabilistic Data Association: joint probability data association) algorithm, etc. to determine the detection data DD RT and route data TD RT Prediction data TD RTpred Is there any correlation between them?

[0081] That is, the correlation processing unit 35 determines the object model C based on the position HP of the candidate point DPH and the predicted position PredP. model1 The correlation processing unit 35 is based on the object model C model1 The correlation processing unit 35 determines whether the position P of the detection point DP is correlated with the predicted position PredP based on the positional relationship between the correlation range RA and the detection point DP when the vehicle exterior information sensor 1 detects at least one of the plurality of objects.

[0082] Specifically, the detection data DD is determined based on whether the Mahalanobis distance dm exceeds the correlation range RA. RT With route data TD RT Prediction data TD RTpred Is there any correlation? Based on the test data DD RT The position P of the detection point DP and the route data TD contained in RT Prediction data TD RTpred The predicted position PredP included in the Mahalanobis distance dm is derived. When the derived Mahalanobis distance dm does not exceed the correlation range RA, it is determined that the detection data DD RT and route data TD RT Prediction data TD RTpredWhen the derived Mahalanobis distance dm exceeds the correlation range RA, it is determined to be the detection data DD RT and route data TD RT Prediction data TD Rtpred Not relevant.

[0083] That is, the correlation processing unit 35 determines whether or not there is a correlation between the position P of the detection point DP and the predicted position PredP.

[0084] In the above example, the Mahalanobis distance dm derived based on the position P of the detection point DP and the predicted position PredP is used as an index for comparison with the correlation range RA, but the present invention is not limited to this.

[0085] As described above, the correlation range RA is set based on the reference position BP. The reference position BP is determined based on the position HP of the candidate point DPH and the predicted position PredP. Therefore, the predicted position PredP is associated with the reference position BP. Therefore, the Mahalanobis distance dm can be derived based on the position P of the detection point DP and the reference position BP.

[0086] In addition, the index used for comparison with the relevant range RA may not be the Mahalanobis distance dm. The Euclidean distance du of the difference vector between the position P of the detection point DP and the reference position BP may be used. In this case, the detection data DD is determined based on whether the Euclidean distance du exceeds the relevant range RA. RT With route data TD RT Prediction data TD RTpred Is there any relevant relationship?

[0087] That is, the correlation processing unit 35 determines whether there is a correlation between the position P of the detection point DP and the predicted position PredP based on whether any one of the Euclidean distance du of the differential vector between the position P of the detection point DP and the reference position BP and the Mahalanobis distance dm obtained based on the position P of the detection point DP and the reference position BP exceeds the correlation range RA.

[0088] The relevant range RA is set as the observable range of the exterior information sensor 1. The observable range of the exterior information sensor 1 differs depending on the type of the exterior information sensor 1. Therefore, the relevant range RA differs depending on the type of the exterior information sensor 1.

[0089] The correlation processing unit 35 has the following features: RT and route data TD RT Prediction data TD RTpred Where relevant, determine the test data DD RT and route data TD RT Prediction data TD RTpredThe correlation processing unit 35 has a function of generating the correlation data RD RT function, the relevant data RD RT The detection data DD RT , temporary setting data DH including the position HP of the candidate point DPH, route data TD RT Prediction data TD RTpred Data obtained by integrating data related to the determined correspondence relationship. Related data RD RT The data structure is generated so as to be provided by the correlation processing unit 35 to the update processing unit 36 ​​.

[0090] The update processing unit 36 ​​receives the relevant data RD RT The update processing unit 36 ​​has the function of updating the route data TD based on the position P of the detection point DP and the position HP of the candidate point DPH. RT Route data TD RT Specifically, the update is performed through tracking processing such as the least squares method, Kalman filter, particle filter, etc.

[0091] The notification control device 4 has the function of receiving the route data TD RT The notification control device 4 has a function based on the route data TD RT Function of generating notification data. Notification data is data that determines the content of notification, and is generated in a form corresponding to the device of the output destination. The notification control device 4 outputs the notification data to a display (not shown), so that the display notifies the content of the notification data. Thus, the content of the notification data is visually notified to the driver in the vehicle. The notification control device 4 outputs the notification data to a speaker (not shown), so that the speaker notifies the content of the notification data. Thus, the content of the notification data is auditorily notified to the driver in the vehicle.

[0092] The vehicle control device 5 has a function of receiving the route data TD outputted by the update processing unit 36. RT The vehicle control device 5 has a function based on the route data TD RT The vehicle control device 5 controls the movement of the vehicle based on the route data TD. RT Control the vehicle's movements so that it avoids the object.

[0093] Figure 2 It is shown Figure 1 A diagram of an example of the relative position relationship between a sensor and an object.

[0094] Here, the point located in the center when the vehicle exterior information sensor 1 is viewed from the front is set as the origin O. The horizontal axis in the left-right direction passing through the origin O is set as the Ys axis. In the Ys axis, when the vehicle exterior information sensor 1 is viewed from the front, the right direction is set as the positive direction. The vertical axis in the up-down direction passing through the origin O is set as the Zs axis. In the Zs axis, when the vehicle exterior information sensor 1 is viewed from the front, the upward direction is set as the positive direction. The axis in the front-back direction passing through the origin O and orthogonal to the Ys axis and the Zs axis is set as the Xs axis. In the Xs axis, the front of the vehicle exterior information sensor 1 is set as the positive direction.

[0095] like Figure 2 As shown by the dotted lines in , the observable range of the exterior information sensor 1 is divided into a plurality of imaginary resolution units. The resolution unit is determined based on the resolution of the exterior information sensor 1. The resolution unit is obtained by dividing the observable range of the exterior information sensor 1 according to the angular resolution and the distance resolution of the exterior information sensor 1. As described above, the angular resolution and the distance resolution of the exterior information sensor 1 differ depending on the measurement principle of the exterior information sensor 1.

[0096] Each resolution unit is determined by the minimum detection range MR (i, j). Here, i determines the position of the resolution unit along the circumferential direction with the origin O as the reference. j determines the position of the resolution unit along the radial direction of the concentric circle with the origin O as the reference. Therefore, the number of i varies according to the angular resolution of the external vehicle information sensor 1. Therefore, the higher the angular resolution of the external vehicle information sensor 1, the greater the maximum number of i. On the other hand, the number of j varies according to the distance resolution of the external vehicle information sensor 1. Therefore, the higher the distance resolution of the external vehicle information sensor 1, the greater the maximum number of j. Regarding the positive and negative signs of i, the clockwise direction with the Xs axis as the reference is set as the positive circumferential direction, and the counterclockwise direction with the Xs axis as the reference is set as the negative circumferential direction.

[0097] When the vehicle Ca is detected by the vehicle exterior information sensor 1, the detection point DP (Ca) is included in the minimum detection range MR (3, 3). The minimum detection range MR (3, 3) is set to a size that includes only the rear left side of the vehicle Ca. Therefore, since the positional relationship between the position P of the detection point DP (Ca) and the vehicle Ca is determined, the position P of the detection point DP (Ca) on the vehicle Ca is determined to be on the rear left side of the vehicle Ca. In addition, since the detection point DP (Ca) is included in the minimum detection range MR (3, 3), the position P of the detection point DP (Ca) relative to the vehicle exterior information sensor 1 is determined to be the position P of the closest point with the shortest distance from the vehicle exterior information sensor 1 to the vehicle Ca.

[0098] On the other hand, when the vehicle exterior information sensor 1 detects the vehicle Cb, the detection point DP(Cb) is included in the minimum detection range MR(2,7). If the comparison is made along the radial direction of the concentric circle with the origin O as the reference, the minimum detection range MR(2,7) is farther from the origin O than the minimum detection range MR(3,3). The farther the minimum detection range MR(i,j), that is, the resolution unit, is from the origin O along the radial direction of the concentric circle, the lower the angular resolution of the vehicle exterior information sensor 1. Therefore, the angular resolution of the vehicle exterior information sensor 1 in the minimum detection range MR(2,7) is lower than the angular resolution of the vehicle exterior information sensor 1 in the minimum detection range MR(3,3).

[0099] In addition, the minimum detection range MR (2, 7) is set to include the entire rear of the vehicle Cb. Therefore, it is impossible to determine at which position P the position P of the detection point DP (Cb) is in the entire rear of the vehicle Cb. Therefore, it is impossible to determine the positional relationship between the position P of the detection point DP (Cb) and the vehicle Cb. Therefore, it is impossible to determine the position P of the detection point DP (Cb) on the vehicle Cb.

[0100] Therefore, a description will be given of a process for specifying the position P of the detection point DP (Ca) on the vehicle Ca and the position P of the detection point DP (Cb) on the vehicle Cb.

[0101] Figure 3 It means to become Figure 2 A diagram showing an example of a candidate point DPH(1) which is the first candidate for the position P of the detection point DP(Ca) on the vehicle Ca. When the vehicle exterior information sensor 1 detects the vehicle Ca as an object, the detection point DP(Ca) is included in the minimum detection range MR(3, 3). The minimum detection range MR(3, 3) is set to a size that includes only the rear left side of the vehicle Ca. Therefore, as described above, the nearest point is assumed to be the position P of the detection point DP(Ca) on the vehicle Ca. When the nearest point is assumed to be the position P of the detection point DP(Ca) on the vehicle Ca, the position HP of the candidate point DPH(1) becomes the first candidate for the position P of the detection point DP(Ca) on the vehicle Ca.

[0102] In other words, Figure 3 In one example, the position HP of the candidate point DPH(1) is the first candidate for the position P of the detection point DP(Ca) on the vehicle Ca.

[0103] Figure 4 It means to become Figure 2A diagram showing an example of a candidate point DPH(2) which is the second candidate for the position P of the detection point DP(Cb) on the vehicle Cb. When the vehicle exterior information sensor 1 detects the vehicle Cb as an object, the detection point DP(Cb) is included in the minimum detection range MR(2, 7). The minimum detection range MR(2, 7) is set to a size that includes the entire rear of the vehicle Cb. Therefore, as described above, it is impossible to determine which position P in the entire rear of the vehicle Cb the position P of the detection point DP(Cb) is. When it is impossible to determine which position P in the entire rear of the vehicle Cb the position P of the detection point DP(Cb) is, the position HP of the candidate point DPH(2) becomes the second candidate for the position P of the detection point DP(Cb) on the vehicle Cb. The position HP of the candidate point DPH(2) is assumed to be the rear center point of the rear of the vehicle Cb. The rear center point is a point located in the center when the rear of the vehicle Cb is viewed from the front.

[0104] In other words, Figure 4 In one example, the position HP of the candidate point DPH(2) is a second candidate for the position P of the detection point DP(Cb) on the vehicle Cb.

[0105] Figure 5 : is a diagram showing an example of a candidate point DPH(3) which is another candidate for the position P of the detection point DP(Cc) in the vehicle Cc. When the vehicle exterior information sensor 1 detects the vehicle Cc as an object, the detection point DP(Cc) is included in the minimum detection range MR(-1, 7). For example, the minimum detection range MR(-1, 7) is farther from the origin O than the minimum detection range MR(-1, 3). Therefore, the angular resolution of the vehicle exterior information sensor 1 in the minimum detection range MR(-1, 7) is lower than the angular resolution of the vehicle exterior information sensor 1 in the minimum detection range MR(-1, 3).

[0106] Specifically, the minimum detection range MR (-1, 7) is set to a size that includes the entire front of the vehicle Cc. Therefore, it is impossible to determine at which position P in the entire front of the vehicle Cc the position P of the detection point DP (Cc) is. When it is impossible to determine at which position P in the entire front of the vehicle Cc the position P of the detection point DP (Cc) is, the position HP of the candidate point DPH (3) becomes another candidate for the position P of the detection point DP (Cc) on the vehicle Cc. The position HP of the candidate point DPH (3) is assumed to be the front center point of the front of the vehicle Cc. The front center point is a point located in the center when the front of the vehicle Cc is viewed from the front.

[0107] In other words, Figure 5 In one example, the position HP of the candidate point DPH(3) is another candidate for the position P of the detection point DP(Cc) on the vehicle Cc.

[0108] Reference Figure 3 and Figure 4 , when the vehicle exterior information sensor 1 is a millimeter wave radar monitoring the front of the vehicle, the position HP of the candidate point DPH(1) becomes a candidate for the position P of the detection point DP(Ca) on the vehicle Ca. In addition, the position HP of the candidate point DPH(2) becomes a candidate for the position P of the detection point DP(Cb) on the vehicle Cb.

[0109] In addition, refer to Figure 4 If the vehicle exterior information sensor 1 is a camera that monitors the front of the vehicle, the position HP of the candidate point DPH(2) becomes a candidate for the position P of the detection point DP(Cb) on the vehicle Cb.

[0110] Reference Figure 3 and Figure 5 When the vehicle exterior information sensor 1 is a millimeter wave radar monitoring the rear of the vehicle, the position HP of the candidate point DPH(1) becomes a candidate for the position P of the detection point DP(Ca) on the vehicle Ca. In addition, the position HP of the candidate point DPH(3) becomes a candidate for the position P of the detection point DP(Cc) on the vehicle Cc.

[0111] In this way, when there are multiple candidate points DPH for the position P of the detection point DP, it is impossible to determine the respective positions P of the detection point DP (Ca) on the vehicle Ca, the detection point DP (Cb) on the vehicle Cb, and the detection point DP (Cc) on the vehicle Cc.

[0112] Therefore, the process of selecting and adopting one candidate point DPH from a plurality of candidate points DPH (N) is described. In the following description, when the vehicle Ca, the vehicle Cb, and the vehicle Cc are collectively referred to, they are referred to as the vehicle C. When the detection point DP (Ca), the detection point DP (Cb), and the detection point DP (Cc) are collectively referred to, they are referred to as the detection point DP (C).

[0113] Figure 6 It means that when N is set to a natural number, Figures 3 to 5 This figure shows an example of setting the reliability DOR(N) of the candidate point DPH(N). Figure 6 In one example, for the reliability DOR(N), a real number greater than 0 and less than 1 is set. As described above, if the vehicle exterior information sensor 1 is a millimeter wave radar that monitors the front of the vehicle, the candidate point DPH(1) and the candidate point DPH(2) become candidates for the detection point DP(C) on the vehicle C.

[0114] Therefore, by comparing the reliability DOR(1) for the candidate point DPH(1) and the reliability DOR(2) for the candidate point DPH(2), one of the candidate point DPH(1) and the candidate point DPH(2) is selected and set as a candidate for the position P of the detection point DP(C) on the vehicle C. Thus, one of the candidate point DPH(1) and the candidate point DPH(2) is adopted.

[0115] Specifically, as described above, the farther the resolution unit is from the origin O along the radial direction of the concentric circle, the lower the angular resolution of the exterior vehicle information sensor 1. In other words, the closer the resolution unit is to the origin O along the radial direction of the concentric circle, the higher the angular resolution of the exterior vehicle information sensor 1.

[0116] Therefore, if the distance from the exterior information sensor 1 to the detection point DP (C) is short, the rear of the vehicle C is not buried in the resolution unit. Therefore, if the distance from the exterior information sensor 1 to the detection point DP (C) is short, the reliability DOR is high.

[0117] In other words, the reliability DOR of the candidate point DPH is obtained based on the distance from the vehicle exterior information sensor 1 to the position P of the detection point DP. In addition, the reliability DOR of the candidate point DPH is obtained based on the distance from the vehicle exterior information sensor 1 to the reference position BP. That is, the correlation processing unit 35 obtains each reliability DOR based on the distance from the vehicle exterior information sensor 1 to at least one of the position P of the detection point DP and the reference position BP.

[0118] Therefore, when the distance from the vehicle exterior information sensor 1 to the detection point DP (C) is less than Figure 6 The judgment threshold distance D TH1 , the reliability DOR(1) for the candidate point DPH(1) is set to 1, and the reliability DOR(2) for the candidate point DPH(2) is set to 0. In this case, the reliability DOR(1) is higher than the reliability DOR(2), so the reliability DOR(1) is selected. When the reliability DOR(1) is selected and set, the candidate point DPH(1) corresponding to the reliability DOR(1) is adopted. The position HP of the candidate point DPH(1) on the vehicle C is the position P of the closest point on the vehicle C.

[0119] Therefore, based on the position HP of the adopted candidate point DPH(1), it is assumed that the position P of the detection point DP(C) on the vehicle C is located at the position P of the closest point on the vehicle C.

[0120] In other words, when the distance from the vehicle exterior information sensor 1 to the detection point DP(C) is less than Figure 6 The judgment threshold distance D TH1In the case of , the position HP of the candidate point DPH(1) is adopted as a candidate for the position P of the detection point DP(C) on the vehicle C among the plurality of candidate points DPH(N). Thus, it is assumed that the position P of the detection point DP(C) on the vehicle C is located at the position P of the nearest point on the vehicle C.

[0121] On the other hand, if the distance from the exterior information sensor 1 to the detection point DP (C) is long, the rear of the vehicle C is also buried in the resolution unit. Therefore, if the distance from the exterior information sensor 1 to the detection point DP (C) is long, the reliability DOR is low.

[0122] Therefore, when the distance from the vehicle exterior information sensor 1 to the detection point DP (C) is Figure 6 The judgment threshold distance D TH2 In the above case, the reliability DOR(1) for the candidate point DPH(1) is set to 0, and the reliability DOR(2) for the candidate point DPH(2) is set to 1. In this case, the reliability DOR(2) is higher than the reliability DOR(1), so the reliability DOR(2) is selected. When the reliability DOR(2) is selected and set, the candidate point DPH(2) corresponding to the reliability DOR(2) is adopted. The position HP of the candidate point DPH(2) on the vehicle C is the position P of the rear center point on the vehicle C.

[0123] Therefore, based on the position HP of the adopted candidate point DPH(2), it is assumed that the position P of the detection point DP(C) on the vehicle C is located at the position P of the rear center point on the vehicle C.

[0124] In other words, when the distance from the vehicle exterior information sensor 1 to the detection point DP(C) is Figure 6 The judgment threshold distance D TH2 In the above case, among the plurality of candidate points DPH(N), the position HP of the candidate point DPH(2) is adopted as a candidate for the position P of the detection point DP(C) on the vehicle C. Thus, it is assumed that the position P of the detection point DP(C) on the vehicle C is located at the position P of the rear center point on the vehicle C.

[0125] According to the above description, the correlation processing unit 35 adopts the candidate point DPH(N) having the highest reliability DOR(N) among the positions HP of the plurality of candidate points DPH(N) on the vehicle C.

[0126] in addition, Figure 6 The judgment threshold distance D TH1 is set to include the distance from the origin O along the radial direction of the concentric circle, Figure 3 or Figure 5 The minimum detection range MR (3, 3) of the distance. In other words, Figure 6 The judgment threshold distance DTH1 is set to include the distance from the origin O along the radial direction of the concentric circle, Figure 3 , Figure 4 as well as Figure 5 The distance of the minimum detection range MR(i,3).

[0127] on the other hand, Figure 6 The judgment threshold distance D TH2 is set to include the distance from the origin O along the radial direction of the concentric circle, Figure 4 The minimum detection range MR (2, 7) of the distance. In other words, Figure 6 The judgment threshold distance D TH2 is set to include the distance from the origin O along the radial direction of the concentric circle, Figure 3 , Figure 4 as well as Figure 5 The distance of the minimum detection range MR(i,7).

[0128] In other words, the threshold distance D is determined TH2 is set to be greater than the judgment threshold distance D TH1 The further away from the origin O.

[0129] Specifically, the reliability DOR(1) is less than the judgment threshold distance D TH1 The reliability DOR(1) is set to 1 when the distance D is greater than the threshold value. TH1 When the reliability DOR(1) is at the judgment threshold distance DT, it starts to decrease. H2 Above this value, it is set to 0.

[0130] On the other hand, the reliability DOR(2) is less than the judgment threshold distance D TH1 The reliability DOR(2) is set to 0 at the decision threshold distance D TH1 The reliability DOR(2) starts to rise when the distance D is above the threshold. TH2 Above is set to 1.

[0131] Therefore, the reliability DOR(1) and the reliability DOR(2) are set to be smaller than the determination threshold distance D. TH1 and the judgment threshold distance D TH2 The above trends are opposite to each other.

[0132] The reliability DOR(1) and the reliability DOR(2) are determined when the distance is greater than or equal to the determination threshold distance DTH1 and when the distance is less than or equal to the determination threshold distance DTH2, respectively, based on the ratio of the distance resolution to the angle resolution of the vehicle exterior information sensor 1 .

[0133] Figure 7 It is shown Figure 1 Prediction data TDRTpred An example diagram of .

[0134] Prediction data TD RTpred The object model C is obtained by modeling the vehicle C as the object to be tracked. model1 The predicted position PredP of the predicted point Pred, the speed PredV of the predicted point Pred, and the object model C model1 The width W and length L are 4.

[0135] Object Model C model1 The predicted position PredP of the predicted point Pred, the speed PredV of the predicted point Pred, and the object model C model1 Of the four, width W and length L, the speed PredV of the predicted point Pred, the object model C model1 The width W and length L are the three elements that determine the object.

[0136] Object determination elements determine object model C model1 At least one of a state and a size.

[0137] Object Model C modell The predicted point Pred in is set as the object model C model1 Therefore, the predicted position PredP of the predicted point Pred is located at the center of the object model C model1 center.

[0138] Object Model C model1 The predicted position PredP of the predicted point Pred and the speed PredV of the predicted point Pred in represent the state of the object observable by the millimeter wave radar or LIDAR. model1 The width W and length L represent the size of the object that can be observed by the camera.

[0139] Therefore, the predicted data TD RTpred The data is formed by integrating the observation results of multiple different types of vehicle exterior information sensors 1. For example, the prediction data TD RTpred Constructed as TD RTpred (PredP, PredV, L, W) such vector data.

[0140] Figure 8 It is shown based on Figure 7 Prediction data TD RTpred A diagram showing an example of a reference position BP determined by a predicted position PredP and a candidate point DPH(1).

[0141] As described above, the predicted position PredP is the position of the predicted point Pred. The predicted point Pred is set at the object model C model1 In addition, the position HP of the candidate point DPH(1) is the center point of the object model C model1 The position of the nearest point in .

[0142] In addition, as described above, the prediction data TD RTpred Contains object model C model1 The predicted position PredP of the predicted point Pred, the speed PredV of the predicted point Pred, and the object model C model1 The four data are width W and length L.

[0143] When candidate point DPH(1) is used as candidate point DPH(N) with the highest reliability DOR(N), object model C is used. model1 The closest point in is taken as the candidate point DPH. The position of the closest point is determined as the reference position BP of the reference point B.

[0144] Therefore, the object model C model1 The reference position BP is set to s The position obtained by adding 1 / 2 of the width W to the predicted position PredP in the axial direction. model1 The reference position BP in the s The position obtained by subtracting 1 / 2 of the length L from the predicted position PredP in the axial direction.

[0145] That is, the correlation processing unit 35 determines the object model C based on model1 The object model C is determined by using at least one of the state and size of the object to determine the object model C. model1 The reference position BP in .

[0146] Specifically, the correlation processing unit 35 obtains the object model C from the vehicle exterior information sensor 1 model1 When at least one of the width W and the length L is used as the object identification element, the object identification element can be obtained in addition to the predicted position PredP and the candidate point DPH.

[0147] In this case, the correlation processing unit 35 specifies the reference position BP of the correlation range RA at the current correlation time RT based on the predicted position PredP, the candidate point DPH, and the acquired object identification element.

[0148] The correlation processing unit 35 fails to obtain the object model C from the vehicle exterior information sensor 1. model1When at least one of the width W and the length L is used as the object identification factor, although the predicted position PredP and the candidate point DPH can be obtained, the object identification factor cannot be obtained.

[0149] In this case, the correlation processing unit 35 is associated with the object model C model1 From the set values ​​individually set in advance corresponding to the width W and the length L of the vehicle, the set value corresponding to the object identification element that cannot be acquired from the vehicle exterior information sensor 1 is determined.

[0150] The correlation processing unit 35 determines the value of the object identification element that cannot be obtained from the vehicle exterior information sensor 1 based on the determined setting value. That is, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the predicted position PredP, the candidate point DPH and the setting value.

[0151] Sometimes the correlation processing unit 35 is unable to obtain the object model C from the vehicle exterior information sensor 1. model1 At least one of the width W and the length L is used as an object identification factor, and is not related to the object model C model1 The width W and length L of the device are individually preset with corresponding set values.

[0152] In this case, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the predicted position PredP and the candidate point DPH. Specifically, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the differential vector between the predicted position PredP and the candidate point DPH.

[0153] Next, the object model C is included in the object identification element model1 The case where at least one of the width W, length L and direction θ is used is described.

[0154] In addition, when the object model C can be obtained from the vehicle exterior information sensor 1 model1 When at least one of the width W, length L and direction θ is used as the object identification factor, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current association time RT based on the predicted position PredP, the candidate point DPH and the acquired object identification factor.

[0155] The correlation processing unit 35 fails to obtain the object model C from the vehicle exterior information sensor 1. model1 When at least one of the width W, length L, and direction θ is used as the object identification element, although the predicted position PredP and the candidate point DPH can be obtained, the object identification element cannot be obtained.

[0156] In this case, the correlation processing unit 35 is associated with the object model Cmodel1 The set value corresponding to the object identification element that cannot be acquired from the vehicle exterior information sensor 1 is determined from the set values ​​that are individually preset corresponding to the width W, the length L, and the direction θ.

[0157] The correlation processing unit 35 determines the value of the object identification element that cannot be obtained from the vehicle exterior information sensor 1 based on the determined setting value. That is, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the predicted position PredP, the candidate point DPH and the setting value.

[0158] Sometimes the correlation processing unit 35 is unable to obtain the object model C from the vehicle exterior information sensor 1. model1 At least one of the width W, length L, and direction θ is used as an object identification factor, and is not related to the object model C model1 The width W, length L and direction θ are individually preset with corresponding set values.

[0159] In this case, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the predicted position PredP and the candidate point DPH. Specifically, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the differential vector between the predicted position PredP and the candidate point DPH.

[0160] Next, the object model C is included in the object identification element model1 The case where at least one of the width W, length L, direction θ and height H is used is described.

[0161] The correlation processing unit 35 fails to obtain the object model C from the vehicle exterior information sensor 1. model1 When at least one of the width W, length L, direction θ, and height H is used as the object identification element, although the predicted position PredP and the candidate point DPH can be obtained, the object identification element cannot be obtained.

[0162] In this case, the correlation processing unit 35 is associated with the object model C model1 The set value corresponding to the object identification element that cannot be obtained from the vehicle exterior information sensor 1 is determined from the set values ​​that are individually preset in correspondence with the width W, length L, direction θ, and height H of the vehicle exterior information sensor 1 .

[0163] The correlation processing unit 35 determines the value of the object identification element that cannot be obtained from the vehicle exterior information sensor 1 based on the determined setting value. That is, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the predicted position PredP, the candidate point DPH and the setting value.

[0164] Sometimes the correlation processing unit 35 is unable to obtain the object model C from the vehicle exterior information sensor 1. model1 At least one of the width W, length L, direction θ, and height H of the object is used as an object identification factor, and is not related to the object model C model1 The width W, length L, direction θ and height H of the lens are individually preset with corresponding set values.

[0165] In this case, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the predicted position PredP and the candidate point DPH. Specifically, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the differential vector between the predicted position PredP and the candidate point DPH.

[0166] Next, the object model C is included in the object identification element model1 Width W, length L, direction θ, top Z H The position and lower end Z L The case of at least one of the positions is described.

[0167] The correlation processing unit 35 fails to obtain the object model C from the vehicle exterior information sensor 1. model1 Width W, length L, direction θ, top Z H The position and lower end Z L When at least one of the positions is used as an object identification element, although the predicted position PredP and the candidate point DPH can be obtained, the object identification element cannot be obtained.

[0168] In this case, the correlation processing unit 35 is associated with the object model C model1 Width W, length L, direction θ, top Z H The position and lower end Z L Among the set values ​​individually preset corresponding to the respective positions, the set value corresponding to the object identification element that cannot be acquired from the vehicle exterior information sensor 1 is determined.

[0169] The correlation processing unit 35 determines the value of the object identification element that cannot be obtained from the vehicle exterior information sensor 1 based on the determined setting value. That is, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the predicted position PredP, the candidate point DPH and the setting value.

[0170] Sometimes the correlation processing unit 35 is unable to obtain the object model C from the vehicle exterior information sensor 1. model1 Width W, length L, direction θ, top Z H The position and lower end Z LAt least one of the positions is used as an object identification factor and does not correspond to the object model C model1 Width W, length L, direction θ, top Z H The position and lower end Z L The set values ​​are individually preset corresponding to the positions.

[0171] In this case, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the predicted position PredP and the candidate point DPH. Specifically, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the differential vector between the predicted position PredP and the candidate point DPH.

[0172] In the above, the case where the candidate point DPH(1) is used as the candidate point DPH(N) with the highest reliability DOR(N) is described. However, instead of using the position HP of the candidate point DPH(N) with the highest reliability DOR(N), the position HP of the candidate point DPH(N) obtained by weightedly averaging the positions P of a plurality of candidate points DPH(N) using the reliability DOR(N) may be used.

[0173] Specifically, the correlation processing unit 35 determines the object model C by weighted averaging the positions HP of the plurality of candidate points DPH in the object according to the respective reliabilities DOR. model1 The reference position BP in.

[0174] In summary, when the object model C model1 When there are plural candidate points DPH(N), the correlation processing unit 35 determines the reference position BP based on the reliability DOR(N) of each of the plurality of candidate points DPH(N) and the position HP of each of the plurality of candidate points DPH(N).

[0175] As described above, the correlation range RA is set based on the determined reference position BP.

[0176] For example, the positions along the Xs-axis direction are set to +1 [m] and -1 [m] with respect to the reference position BP, respectively.

[0177] The positions along the Ys-axis direction are set to +1 [m] and -1 [m] with respect to the reference position BP, respectively.

[0178] Furthermore, the speeds in the Xs-axis direction are set to +3 [km / h] and -3 [km / h], respectively, with respect to a reference speed BV of a reference point B located at a reference position BP.

[0179] Furthermore, the speeds in the Ys-axis direction are set to +3 [km / h] and -3 [km / h], respectively, with respect to a reference speed BV of a reference point B located at a reference position BP.

[0180] In the following description, the position along the Xs-axis direction is referred to as the Xs-axis position, the position along the Ys-axis direction is referred to as the Ys-axis position, the speed along the Xs-axis direction is referred to as the Xs-axis speed, and the speed along the Ys-axis direction is referred to as the Ys-axis speed.

[0181] Fig. 9 It is shown that Figure 8 A diagram of a first setting example of the correlation range RA set based on the reference position BP.

[0182] The size of the correlation range RA varies depending on the candidate point DPH used. When the candidate point DPH (1) is used, the reference position BP is set to the position of the closest point. Here, at the reference point B located at the reference position BP, the Xs axis position is set to pnx, the Ys axis position is set to pny, the Xs axis speed is set to vnx, and the Ys axis speed is set to vny.

[0183] In addition, the standard deviation of the detection error of the vehicle exterior information sensor 1 measured in advance is calculated. Here, the standard deviation of the detection error of the Xs axis position is σx, the standard deviation of the detection error of the Ys axis position is σy, the standard deviation of the detection error of the Xs axis speed is σvx, and the standard deviation of the detection error of the Ys axis speed is σvy.

[0184] Therefore, if Fig. 9 As shown, the related range RA is set as follows.

[0185] Xs axis position: interval (pnx-σx, pnx+σx)

[0186] Ys axis position: interval (pny-σy, pny+σy)

[0187] Xs axis speed: interval (vnx-σvx, vnx+σvx)

[0188] Ys axis speed: interval (vny-σvy, vny+σvy)

[0189] Fig.10 It is shown that Figure 8 A diagram of a second setting example of the correlation range RA set based on the reference position of .

[0190] Using forecast data TD RTpred The object model C contained in model1 The width W and object model C model1 The length L.

[0191] like Fig.10 As shown, the related range RA is set as follows.

[0192] Xs axis position: interval (pnx-σx, pnx+σx+L)

[0193] Ys axis position: interval (pny-σy, pny+σy+W)

[0194] Xs axis speed: interval (vnx-σvx, vnx+σvx)

[0195] Ys axis speed: interval (vny-σvy, vny+σvy)

[0196] Fig.11 It is shown that Figure 8 A diagram of a third setting example of the correlation range RA set based on the reference position BP.

[0197] When the candidate point DPH (2) is adopted, the reference position BP is set as the predicted position PredP. That is, the reference point B is set as the predicted point Pred. Here, at the reference point B located at the reference position BP, the Xs axis position is set to pcx, the Ys axis position is set to pcy, the Xs axis speed is set to vcx, and the Ys axis speed is set to vcy.

[0198] The standard deviation of the detection error of the vehicle exterior information sensor 1 statistically measured in advance is the same as the above case.

[0199] Therefore, if Fig.11 As shown, the related range RA is set as follows.

[0200] Xs axis position: interval (pcx-σx-L / 2, pcx+σx+L / 2)

[0201] Ys axis position: interval (pcy-σy-W / 2, pcy+σy+W / 2)

[0202] Xs axis speed: interval (vcx-σvx, vcx+σvx)

[0203] Ys axis speed: interval (vcy-σvy, vcy+σvy)

[0204] Here, the standard deviation of the detection error of the vehicle exterior information sensor 1 that has been statistically measured in advance may also be reflected in the prediction data TD RTpred Object model C in model1 The width W and length L are.

[0205] Specifically, for the predicted data TD RTpred The object model C contained in model1The width W of the vehicle exterior information sensor 1 is set to σ. W For the predicted data TD RTpred The object model C contained in model1 The length L of the vehicle exterior information sensor 1 is set to σL, and the standard deviation of the detection error of the vehicle exterior information sensor 1 is set to σL.

[0206] Therefore, in the correlation range RA, the prediction data TD is set as follows RTpred The object model C contained in model1 The width W and length L.

[0207] Width W: interval (W-σ W , W+σ W )

[0208] Length L: interval (L-σ L , L+σ L )

[0209] In addition, in the prediction data TD RTpred When the direction θ is included in , the direction in the correlation range RA may be set as follows.

[0210] Direction: Difference from θ is within 45[deg]

[0211] In addition, the size of the correlation range RA may be adjusted according to the reliability DOR of the candidate point DPH.

[0212] Specifically, the standard deviation of the detection error of the vehicle exterior information sensor 1 is multiplied by (1-DOR) to obtain a coefficient corresponding to the reliability DOR.

[0213] Therefore, the relevant range RA is set as follows.

[0214] Xs axis position: interval (pnx-(2-DOR)σx, pnx+(2-DOR)σx)

[0215] Ys axis position: interval (pny-(2-DOR)σy, pny+(2-DOR)σy)

[0216] X s Axis speed: interval (vnx-(2-DOR)σvx, vnx+(2-DOR)σvx)

[0217] Ys axis speed: interval (vny-(2-DOR)σvy, vny+(2-DOR)σvy)

[0218] Therefore, the lower the reliability DOR is, the more the influence of the standard deviation of the detection error of the vehicle exterior information sensor 1 is reflected. Therefore, the lower the reliability DOR is, the larger the size of the correlation range RA is.

[0219] In other words, the correlation processing unit 35 is based on the object model C centered at the predicted position PredP. model1 The size and object model C model1 The correlation range RA is set based on the statistic of the detection error caused by the vehicle exterior information sensor 1 that is related to the size of .

[0220] Furthermore, the correlation processing unit 35 adjusts the size of the set correlation range RA according to the plurality of reliabilities DOR(N).

[0221] Fig.12 It is shown in Figure 7 The route data TD also includes an example of the direction θ. Object model C model1 The width W is related to the object model C model1 The direction θ is perpendicular to the object model C model1 The size of the object model C model1 The length L is the same as the object model C model1 The direction θ is parallel to the object model C model1 size.

[0222] According to the measurement principle of the vehicle exterior information sensor 1, when the object model C can be obtained model1 When the direction θ is reached, add the object model C model1 The direction θ is used as the object determination factor of the detection data DD. According to the measurement principle of the vehicle exterior information sensor 1, when the object model C cannot be obtained, model1 When the direction θ is model1 , that is, the object’s speed relative to the ground to change the setting of direction θ.

[0223] When the object's velocity relative to the ground is not zero, since the object model C can be observed model1 On the other hand, when the ground velocity of the object is zero, that is, the object is a stationary object, the initial angle 0 [deg] is included in the temporary setting data DH as a preset setting value.

[0224] Fig.13 It is shown in Figure 7 The route data TD also includes an example of the height H. Object model C model1 The direction θ is parallel to the road surface RS and the object model C model1 The height H is vertical.

[0225] According to the measurement principle of the vehicle exterior information sensor 1, when the object model C can be obtained model1 When the height H is reached, add the object model C model1The height H is used as the object determination factor of the detection data DD. According to the measurement principle of the vehicle exterior information sensor 1, when the object model C cannot be obtained, model1 When the height H is set, the initial height 1.5 [m] is included in the temporary setting data DH as a preset setting value.

[0226] Fig.14 It is shown Figure 7 The route data TD also includes the upper Z H Position and lower end Z L Here, the upper end Z is set to H Position ≥ lower end Z L The position of the lower end Z L When the position of is greater than 0 [m], it is determined that the object is an object existing above the object such as a signboard or a road sign.

[0227] When the upper end Z can be obtained through the measurement principle of the vehicle exterior information sensor 1 H Position and lower end Z L When the position is H Position and lower end Z L The position of the upper end Z is used as the detection element of the detection data DD. H Position and lower end Z L When the position is set to the initial upper end Z, the temporary setting data DH includes HDEF =1.5[m] and initial lower end Z LDEF =0[m] as a preset setting value.

[0228] Fig.15 This is a schematic illustration of Figure 2 The determination target object model C of the relevant determination target is centered at the position P of the detection point DP. model2 With Figure 8 The object model C of the tracked object is centered at the predicted position PredP model1 Overlapping graph.

[0229] like Fig.15 As shown, the object model C model2 The area S O With object model C model1 The overlap ratio SO / ST of the area ST is defined as the overlap ratio R. By using the overlap ratio R and a plurality of reliability DOR(N), the detection point DP(C model2 ) and whether the result of the determination of whether there is a correlation between the position P and the predicted position PredP is appropriate.

[0230] Determination target object model C model2is generated by modeling an object centered at the position P of the detection point DP. On the other hand, as described above, the object model C model1 It is generated by modeling an object centered at the predicted position PredP.

[0231] Specifically, when α and β are coefficients represented by real numbers greater than or equal to 0, and the evaluation value is γ1, the following evaluation function (1) is expressed.

[0232] α×(1-R)+β×(1-DOR)=γ1 (1)

[0233] Therefore, the higher the overlap rate R, the smaller the term including α. In addition, the higher the reliability DOR, the smaller the term including β. Therefore, the smaller the evaluation value γ1, the smaller the detection point DP (C model2 ) and the predicted position PredP, the more appropriate the result of the determination is.

[0234] In this case, for example, the relevant validity flag is set to 1.

[0235] On the other hand, the larger the evaluation value γ1 is, the higher the detection point DP (C model2 ) is more inappropriate in determining whether there is a correlation between the position P and the predicted position PredP.

[0236] In this case, for example, the relevant validity flag is set to 0.

[0237] Furthermore, by setting a threshold value TH1 for the evaluation value γ1, the relevant validity flag may be set to either 1 or 0.

[0238] For example, if the evaluation value γ1 is smaller than the threshold value TH1, the relevant validity flag is set to 1. On the other hand, if the evaluation value γ1 is greater than or equal to the threshold value TH1, the relevant validity flag is set to 0.

[0239] In other words, the correlation processing unit 35 obtains the determination target object model C obtained by modeling the object centered at the position P of the detection point DP. model2 and the object model C centered at the predicted position PredP model1 The correlation processing unit 35 evaluates whether the result of the determination of whether there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate based on the overlap ratio R and the plurality of reliabilities DOR(N).

[0240] In the above description, an example of the evaluation function using the overlap ratio R has been described, but the present invention is not limited to this.

[0241] For example, the candidate point DPH with the highest reliability DOR is used, and the Euclidean distance du is used in comparison with the correlation range RA. Here, if α and β are coefficients represented by real numbers greater than 0, and the evaluation value is γ2, the evaluation function shown in the following (2) is expressed.

[0242] α×du+β×(1-DOR)=γ2 (2)

[0243] Therefore, the smaller the Euclidean distance du is, the smaller the term including α is. In addition, the higher the reliability DOR is, the smaller the term including β is. Therefore, the smaller the evaluation value γ2 is, the more likely it is that the detection point DP (C model2 ) and the predicted position PredP, the more appropriate the judgment result is.

[0244] In this case, for example, the relevant validity flag is set to 1.

[0245] On the other hand, the larger the evaluation value γ2 is, the more likely it is that the detection point DP (C model2 ) is more inappropriate in determining whether there is a correlation between the position P and the predicted position PredP.

[0246] In this case, for example, the relevant validity flag is set to 0.

[0247] Furthermore, by setting a threshold value TH2 for the evaluation value γ2, the relevant validity flag may be set to either 1 or 0.

[0248] For example, if the evaluation value γ2 is smaller than the threshold value TH2, the relevant validity flag is set to 1. On the other hand, if the evaluation value γ2 is greater than or equal to the threshold value TH2, the relevant validity flag is set to 0.

[0249] For example, the candidate point DPH with the highest reliability DOR is used, and the Mahalanobis distance dm is used for comparison with the correlation range RA. Here, if α and β are coefficients represented by real numbers greater than 0, and the evaluation value is γ3, the evaluation function shown in the following (3) is expressed.

[0250] α×dm+β×(1-DOR)=γ3 (3)

[0251] Therefore, the smaller the Mahalanobis distance dm is, the smaller the term including α is. In addition, the higher the reliability DOR is, the smaller the term including β is. Therefore, the smaller the evaluation value γ3 is, the more likely it is that the detection point DP (C model2 ) and the predicted position PredP, the more appropriate the judgment result is.

[0252] In this case, for example, the relevant validity flag is set to 1.

[0253] On the other hand, the larger the evaluation value γ3 is, the more likely it is that the detection point DP (C model2 ) is more inappropriate in determining whether there is a correlation between the position P and the predicted position PredP.

[0254] In this case, for example, the relevant validity flag is set to 0.

[0255] In addition, by setting a threshold value TH3 for the evaluation value γ3, the relevant validity flag can be set to either 1 or 0. For example, if the evaluation value γ3 is less than the threshold value TH3, the relevant validity flag is set to 1. On the other hand, if the evaluation value γ3 is greater than the threshold value TH3, the relevant validity flag is set to 0.

[0256] In addition, for example, according to each reliability DOR(N), the determination target object model C is adopted. model2 The candidate point DPH after weighted average of each position HP of multiple candidate points DPH(N) in the correlation range RA is compared with the Euclidean distance du. Here, if α and β are set as coefficients represented by real numbers greater than 0, the reliability average value is set to DOR avr , and setting the evaluation value to γ4, the evaluation function shown in the following (4) is expressed.

[0257] α×du+β×(1-DOR avr )=γ4 (4)

[0258] Therefore, the smaller the Euclidean distance du is, the smaller the term containing α is. In addition, the reliability average DOR avr The higher the value, the smaller the term including β. Therefore, the smaller the evaluation value γ4 is, the more likely it is that the detection point DP (C model2 ) and the predicted position PredP, the more appropriate the judgment result is.

[0259] In this case, for example, the relevant validity flag is set to 1.

[0260] On the other hand, the larger the evaluation value γ4 is, the more likely it is that the detection point DP (C model2 ) is more inappropriate in determining whether there is a correlation between the position P and the predicted position PredP.

[0261] In this case, for example, the relevant validity flag is set to 0.

[0262] In addition, by setting a threshold value TH4 for the evaluation value γ4, the relevant validity flag can be set to either 1 or 0. For example, if the evaluation value γ4 is less than the threshold value TH4, the relevant validity flag is set to 1. On the other hand, if the evaluation value γ4 is greater than the threshold value TH4, the relevant validity flag is set to 0.

[0263] In addition, for example, according to each reliability DOR(N), the determination target object model C is adopted. model2 The candidate point DPH after weighted average of each position HP of multiple candidate points DPH(N) in the correlation range RA is compared with the Mahalanobis distance dm. Here, if α and β are set as coefficients represented by real numbers greater than 0, the reliability average value is set to DOR avr , and setting the evaluation value to γ5, the evaluation function shown in the following (5) is expressed.

[0264] α×dm+β×(1-DOR avr )=γ5 (5)

[0265] Therefore, the smaller the Mahalanobis distance dm is, the smaller the term including α is. In addition, the reliability average DOR avr The higher the value, the smaller the term including β. Therefore, the smaller the evaluation value γ5 is, the more likely it is that the detection point DP (C model2 ) and the predicted position PredP, the more appropriate the judgment result is.

[0266] In this case, for example, the relevant validity flag is set to 1.

[0267] On the other hand, the larger the evaluation value γ5 is, the more likely it is that the detection point DP (C model2 ) is more inappropriate in determining whether there is a correlation between the position P and the predicted position PredP.

[0268] In this case, for example, the relevant validity flag is set to 0.

[0269] In addition, by setting a threshold value TH5 for the evaluation value γ5, the relevant validity flag can be set to either 1 or 0. For example, if the evaluation value γ5 is less than the threshold value TH5, the relevant validity flag is set to 1. On the other hand, if the evaluation value γ5 is greater than the threshold value TH5, the relevant validity flag is set to 0.

[0270] In other words, the correlation processing unit 35 evaluates whether the result of determining whether there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate based on either the Euclidean distance du or the Mahalanobis distance dm and a plurality of reliabilities DOR(N).

[0271] Here, the Euclidean distance du is obtained from the difference vector between the position P of the detection point DP and the reference position BP, while the Mahalanobis distance dm is obtained from the position P of the detection point DP and the reference position BP.

[0272] Furthermore, for example, the correlation processing unit 35 obtains the object model C centered at the predicted position PredP. model1 The vertices of the object are modeled with the object centered at the position P of the detection point DP to obtain the judgment object model C model2 The minimum sum of distances between the vertices of .

[0273] The correlation processing unit 35 evaluates whether the result of determination as to whether there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate based on the obtained minimum value and the plurality of reliabilities DOR(N).

[0274] Specifically, α and β are coefficients expressed by real numbers greater than 0. In addition, the object model C centered at the predicted position PredP is model1 The judgment object model C is obtained by modeling the object with the position P of the detection point DP as the center. model2 The minimum value of the sum of the distances of the vertices of is set to Rm. Here, if the evaluation value is set to γ6, the evaluation function shown in the following (6) is expressed.

[0275] α×Rm+β×(1-DOR)=γ6 (6)

[0276] Therefore, the smaller the minimum value Rm is, the smaller the term including α is. In addition, the higher the reliability DOR is, the smaller the term including β is. Therefore, the smaller the evaluation value γ6 is, the more likely it is that the detection point DP (C model2 ) and the predicted position PredP, the more appropriate the judgment result is.

[0277] In this case, for example, the relevant validity flag is set to 1.

[0278] On the other hand, the larger the evaluation value γ6 is, the more likely it is that the detection point DP (C model2 ) is more inappropriate in determining whether there is a correlation between the position P and the predicted position PredP.

[0279] In this case, for example, the relevant validity flag is set to 0.

[0280] In addition, by setting a threshold value TH6 for the evaluation value γ6, the relevant validity flag may be set to either 1 or 0. For example, if the evaluation value γ6 is less than the threshold value TH6, the relevant validity flag is set to 1. On the other hand, if the evaluation value γ6 is greater than or equal to the threshold value TH6, the relevant validity flag is set to 0.

[0281] In addition, the object model C centered at the predicted position PredP will be obtained model1 The judgment object model C is obtained by modeling the object with the position P of the detection point DP as the center. model2 The minimum value of the sum of distances of the vertices is attributed to solving the minimum Steiner tree problem. The minimum Steiner tree problem is the shortest network problem.

[0282] Therefore, the correlation processing unit 35 evaluates the detection point DP (C model2 ) is appropriate to determine whether there is a correlation between the position P and the predicted position PredP.

[0283] Next, explain Figure 1 Processing of the object recognition device 3 in.

[0284] Fig.16 It is explained by Figure 1 Flowchart of the processing performed by the object recognition device 3.

[0285] In step S11, the time measuring unit 31 determines whether the current time has reached the processing time tk. If the time measuring unit 31 determines that the current time has reached the processing time tk, the process in step S11 is transferred to the process in step S12. If the time measuring unit 31 determines that the current time has not reached the processing time tk, the process in step S11 is continued.

[0286] In step S12, the data receiving unit 32 receives the detection data dd from each vehicle exterior information sensor 1. Next, the process of step S12 is transferred to the process of step S13.

[0287] In step S13 , the data receiving unit 32 associates the time at which the detection data dd is received from each vehicle exterior information sensor 1 with the detection data DD, and sets it as the current association time RT. Next, the process of step S13 moves to the process of step S14 .

[0288] In step S14, the data receiving unit 32 marks all the vehicle exterior information sensors 1 as unused. Next, the process of step S14 is transferred to the process of step S15.

[0289] In step S15, the data receiving unit 32 determines whether there is an unused vehicle exterior information sensor 1. When the data receiving unit 32 determines that there is an unused vehicle exterior information sensor 1, the process of step S15 is transferred to the process of step S16. When the data receiving unit 32 determines that there is no unused vehicle exterior information sensor 1, the process of step S15 is not transferred to other processes, and the process of the object recognition device 3 ends.

[0290] In step S16, the prediction processing unit 34 calculates the predicted data TD of the route data TD at the current correlation time RT based on the route data TD at the previous correlation time RT. RTpred Next, the process of step S16 is transferred to the process of step S17.

[0291] In step S17, the temporary setting unit 33 selects the vehicle exterior information sensor 1 to be used. Next, the process of step S17 is transferred to the process of step S18.

[0292] In step S18, the temporary setting unit 33 sets the position HP of at least one candidate point DPH on the object model Cmodel1 obtained by modeling the object detected by the selected exterior information sensor 1 based on the resolution of the selected exterior information sensor 1. Next, the process of step S18 is transferred to the process of step S19.

[0293] In step S19, the correlation processing unit 35 executes the Fig.17 The related association process described later. Next, the process of step S19 is transferred to the process of step S20.

[0294] In step S20, the correlation processing unit 35 executes the Fig.19 Correlation determination processing described later Next, the process of step S20 is transferred to the process of step S21.

[0295] In step S21, the correlation processing unit 35 determines whether the correlation validity flag is set to 1. When the correlation processing unit 35 determines that the correlation validity flag is set to 1, the process of step S21 shifts to the process of step S22. When the correlation processing unit 35 determines that the correlation validity flag is set to 1, the process of step S21 shifts to the process of step S23.

[0296] In step S22, the update processing unit 36 ​​updates the route data TD at the current correlation time RT based on the corrected position P of the detection point DP relative to the vehicle exterior information sensor 1 at the current correlation time RT. Next, the process of step S22 moves to the process of step S23.

[0297] In step S23, the data receiving unit 32 marks the selected vehicle exterior information sensor 1 as being used. Next, the process of step S23 is transferred to the process of step S15.

[0298] Next, explain Fig.16 The relevant association processing in step S19.

[0299] Fig.17 Yes Description Fig.16 Flowchart of the relevant association processing in step S19.

[0300] In step S31, the correlation processing unit 35 determines whether the number of candidate points DPH is plural. When the correlation processing unit 35 determines that the number of candidate points DPH is plural, the process of step S31 is transferred to the process of step S32.

[0301] On the other hand, when the correlation processing unit 35 determines that the number of candidate points DPH is not plural, the process of step S31 shifts to the process of step S42.

[0302] In step S42, the correlation processing unit 35 adopts the set candidate point DPH. Next, the process of step S42 is transferred to the process of step S35.

[0303] Returning to step S32, the correlation processing unit 35 obtains the reliability DOR of each of the plurality of candidate points DPH based on the distance from the selected vehicle exterior information sensor 1 to at least one of the position P of the detection point DP and the reference position BP. Next, the process of step S32 is transferred to the process of step S33.

[0304] In step S33, the correlation processing unit 35 determines whether to execute weighted averaging. When the correlation processing unit 35 determines to execute weighted averaging, the process of step S33 is transferred to the process of step S34.

[0305] In step S34, the correlation processing unit 35 uses the candidate point DPH obtained by weighted averaging the positions HP of the plurality of candidate points DPH in the object according to the respective degrees of reliability DOR. Next, the process of step S34 is transferred to the process of step S35.

[0306] On the other hand, when the correlation processing unit 35 determines in step S33 that weighted averaging is not to be performed, the process of step S33 is transferred to the process of step S39.

[0307] In step S39, the correlation processing unit 35 adopts the candidate point DPH having the highest reliability DOR among the positions HP of the plurality of candidate points DPH. Then, the process of step S39 shifts to the process of step S35.

[0308] In step S35, the correlation processing unit 35 determines whether the object identifying element has been acquired. When the correlation processing unit 35 determines that the object identifying element has been acquired, the process of step S35 is transferred to the process of step S41.

[0309] In step S41, the correlation processing unit 35 specifies the reference position BP of the correlation range RA at the current correlation time RT based on the predicted position PredP, the candidate point DPH, and the object identification element acquired from the vehicle exterior information sensor 1. Next, the process of step S41 shifts to the process of step S38.

[0310] On the other hand, when the correlation processing unit 35 determines in step S35 that the object specifying element cannot be acquired, the process of step S35 shifts to the process of step S36.

[0311] In step S36, the correlation processing unit 35 determines whether a set value is individually preset corresponding to the object identification element that cannot be obtained from the vehicle exterior information sensor 1. When the correlation processing unit 35 determines that a set value is individually preset corresponding to the object identification element that cannot be obtained from the vehicle exterior information sensor 1, the process of step S36 is transferred to the process of step S40.

[0312] In step S40, the correlation processing unit 35 determines the reference position BP of the correlation range RA at the current correlation time RT based on the predicted position PredP, the candidate point DPH, and the set value individually set in advance corresponding to the object identification element that cannot be obtained from the vehicle exterior information sensor 1. Then, the process of step S40 is transferred to the process of step S38.

[0313] On the other hand, if the correlation processing unit 35 determines in step S36 that a set value has not been individually set in advance corresponding to the object identifying element that cannot be acquired from the vehicle exterior information sensor 1 , the process of step S36 shifts to the process of step S37 .

[0314] In step S37, based on the predicted position PredP and the candidate point DPH, the reference position BP of the correlation range RA at the current correlation time RT is determined. Next, the process of step S37 is transferred to the process of step S38.

[0315] In step S38, the correlation processing unit 35 executes the Fig.18 The related range setting process will be described later.

[0316] Next, the process of step S38 is not transferred to other processes, and the relevant association process ends.

[0317] Next, explain Fig.17 The relevant range setting process in step S38.

[0318] Fig.18 Yes Description Fig.17 Flowchart of the relevant range setting process in step S38.

[0319] In step S51, the correlation processing unit 35 determines whether the reliability DOR has been obtained. When the correlation processing unit 35 determines that the reliability DOR has been obtained, the process of step S51 is transferred to the process of step S52.

[0320] In step S52 , the correlation processing unit 35 sets the reliability flag to 1. Next, the process of step S52 is transferred to the process of step S54 .

[0321] On the other hand, when the correlation processing unit 35 determines in step S51 that the degree of reliability DOR has not been obtained, the process of step S51 shifts to the process of step S53.

[0322] In step S53 , the correlation processing unit 35 sets the reliability flag to 0. Next, the process of step S53 shifts to the process of step S54 .

[0323] In step S54, the correlation processing unit 35 obtains the object model C of the tracking target centered at the predicted position PredP. model1 Next, the process of step S54 is transferred to the process of step S55.

[0324] In step S55, the correlation processing unit 35 determines whether the detection error caused by the vehicle exterior information sensor 1 is reflected in the correlation range RA. When the correlation processing unit 35 determines that the detection error caused by the vehicle exterior information sensor 1 is reflected in the correlation range RA, the process of step S55 shifts to the process of step S56.

[0325] On the other hand, when the correlation processing unit 35 determines in step S55 that the detection error caused by the vehicle exterior information sensor 1 is not reflected in the correlation range RA, the process of step S55 shifts to the process of step S60 .

[0326] In step S60, the correlation processing unit 35 sets the correlation range RA based on the reference position BP. Next, the process of step S60 is transferred to the process of step S58.

[0327] Returning to step S56, the correlation processing unit 35 obtains the model1 The statistic of the detection error caused by the vehicle exterior information sensor 1 related to the size of . Next, the process of step S56 is transferred to the process of step S57.

[0328] In step S57, the correlation processing unit 35 generates an object model C centered at the predicted position PredP based on the object model C centered at the predicted position PredP. model1 The relevant range RA is set based on the size and statistics of . Next, the process of step S57 is transferred to the process of step S58.

[0329] In step S58 , the correlation processing unit 35 determines whether the reliability flag is set to 1. When the correlation processing unit 35 determines that the reliability flag is set to 1, the process of step S58 is transferred to the process of step S59 .

[0330] In step S59, the correlation processing unit 35 adjusts the size of the set correlation range RA according to the reliability DOR. Then, the process of step S59 is not transferred to other processes, and the correlation range setting process ends.

[0331] On the other hand, when the correlation processing unit 35 determines in step S58 that the reliability flag is not set to 1, the process of step S58 is not transferred to other processes, and the correlation range setting process ends.

[0332] Next, explain Fig.16 The relevant determination process in step S20.

[0333] Fig.19 Yes Description Fig.16 Flowchart of the correlation determination process in step S20.

[0334] In step S71, the correlation processing unit 35 determines whether any one of the Euclidean distance du of the difference vector between the position P of the detection point DP and the reference position BP and the Mahalanobis distance dm obtained based on the position P of the detection point DP and the reference position BP exceeds the correlation range RA.

[0335] When the correlation processing unit 35 determines that either the Euclidean distance du or the Mahalanobis distance dm exceeds the correlation range RA, the process of step S71 is transferred to the process of step S73.

[0336] In step S73 , the correlation processing unit 35 sets the correlation flag to 0. Next, the process of step S73 is transferred to the process of step S74 .

[0337] On the other hand, when the correlation processing unit 35 determines in step S71 that either the Euclidean distance du or the Mahalanobis distance dm does not exceed the correlation range RA, the process of step S71 is transferred to the process of step S72.

[0338] In step S72 , the correlation processing unit 35 sets the correlation flag to 1. Next, the process of step S72 is transferred to the process of step S74 .

[0339] In step S74 , the correlation processing unit 35 determines whether the correlation flag is set to 1. When the correlation processing unit 35 determines that the correlation flag is set to 1, the process of step S74 shifts to the process of step S75 .

[0340] In step S75, the correlation processing unit 35 executes the Fig. 20 The validity determination process will be described later. Next, step S75 does not transfer to other processes, and the relevant determination process ends.

[0341] On the other hand, when the correlation processing unit 35 determines in step S74 that the correlation flag is not set to 1, the process of step S74 is not transferred to other processes, and the correlation determination process ends.

[0342] Next, explain Fig.19 The validity determination process in step S76.

[0343] Fig. 20 Yes Description Fig.19 Flowchart of the validity determination process in step S76 of FIG.

[0344] In step S91, the correlation processing unit 35 determines whether to use any one of the Euclidean distance du of the difference vector between the position P of the detection point DP and the reference position BP and the Mahalanobis distance dm obtained based on the position P of the detection point DP and the reference position BP.

[0345] When the correlation processing unit 35 determines that either the Euclidean distance du or the Mahalanobis distance dm is used, the process of step S91 is transferred to the process of step S92.

[0346] In step S92, the correlation processing unit 35 determines whether the reliability DOR has been obtained. When the correlation processing unit 35 determines that the reliability DOR has been obtained, the process of step S92 is transferred to the process of step S93.

[0347] In step S93, the correlation processing unit 35 evaluates whether the result of the determination of whether there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate based on any one of the Euclidean distance du and the Mahalanobis distance dm and the plurality of reliabilities DOR(N). Then, the process of step S93 is transferred to the process of step S97.

[0348] On the other hand, in step S91 , when it is determined that neither the Euclidean distance du nor the Mahalanobis distance dm is used, the process of step S91 is transferred to the process of step S94 .

[0349] In step S94, the correlation processing unit 35 determines whether to use the determination target object model C obtained by modeling the object centered at the position P of the detection point DP. model2 and the object model C centered at the predicted position PredP model1 The overlap ratio R.

[0350] When the correlation processing unit 35 determines that the determination target object model C is used model2 With object model C model1 When the overlap ratio R is , the processing of step S94 is transferred to the processing of step S95.

[0351] In step S95, the correlation processing unit 35 determines the object model C based on the object model C. model2 With object model C model1 The overlap ratio R and the plurality of reliabilities DOR(N) are used to evaluate whether the result of the determination of whether there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate. Next, the process of step S95 is transferred to the process of step S97.

[0352] On the other hand, in step S94, the correlation processing unit 35 determines that the determination target object model C is not used. model2 With object model C model1 When the overlap ratio R is , the processing of step S94 is transferred to the processing of step S96.

[0353] In step S96, the correlation processing unit 35 performs the following operations based on the object model C model1 Each vertex and the judgment object model C model2 The minimum value of the sum of distances between the vertices and the plurality of reliabilities DOR(N) are used to evaluate whether the result of the determination of whether there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate. Next, the process of step S96 is transferred to the process of step S97.

[0354] On the other hand, when the correlation processing unit 35 determines in step S92 that the reliability DOR has not been obtained, the process of step S92 shifts to the process of step S97.

[0355] In step S97 , the correlation processing unit 35 determines whether the result of determination as to whether there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate.

[0356] When the correlation processing unit 35 determines that the result of the determination as to whether or not there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate, the process of step S97 is transferred to the process of step S98.

[0357] In step S98 , the correlation processing unit 35 sets the correlation validity flag to 1. Then, the process of step S98 is not transferred to other processes, and the validity determination process ends.

[0358] On the other hand, when the correlation processing unit 35 determines in step S97 that the result of the determination as to whether or not there is a correlation between the position P of the detection point DP and the predicted position PredP is inappropriate, the process of step S97 is transferred to the process of step S99.

[0359] In step S99 , the correlation processing unit 35 sets the correlation validity flag to 0. Then, the process of step S98 is not transferred to other processes, and the validity determination process ends.

[0360] According to the above description, the correlation processing unit 35 determines the object model C based on the position HP of the candidate point DPH and the predicted position PredP. model1 The reference position BP in.

[0361] The correlation processing unit 35 determines whether there is a correlation between the position P of the detection point DP and the predicted position PredP based on the positional relationship between the correlation range RA and the detection point DP. Here, the correlation range RA is set based on the reference position BP. The detection point DP is obtained when the vehicle exterior information sensor 1 detects at least one of the plurality of objects.

[0362] Specifically, the object model C model1 The predicted position PredP is the position of the object's destination as the object model C. model1 Therefore, the resolution of the vehicle exterior information sensor 1 is not reflected in the predicted position PredP.

[0363] As described above, the exterior information sensor 1 has different resolutions depending on the measurement principle of the exterior information sensor 1. Therefore, the temporary setting unit 33 sets the object model C based on the resolution of the exterior information sensor 1. model1 Thus, the resolution of the vehicle exterior information sensor 1 is reflected in the position HP of the candidate point DPH.

[0364] Furthermore, the correlation processing unit 35 determines the object model C based on the position HP of the candidate point DPH and the predicted position PredP. model1 Therefore, the resolution of the vehicle exterior information sensor 1 is also reflected in the object model C model1 Therefore, the resolution of the vehicle exterior information sensor 1 is also reflected in the correlation range RA set based on the reference position BP.

[0365] Furthermore, the correlation processing unit 35 uses the correlation range RA for a process of determining whether or not there is a correlation between the position P of the detection point DP and the predicted position PredP.

[0366] Thus, the result of the determination process of whether or not there is a correlation between the position P of the detection point DP and the predicted position PredP is a result that takes into account the deviation caused by the resolution of the vehicle exterior information sensor 1. As a result, for the correlation range RA set with the reference position BP of the correlation range RA as a reference, the deviation caused by the resolution of the vehicle exterior information sensor 1 is taken into account. Therefore, if it is determined whether or not there is a correlation between the correlation range RA and the position P of the detection point DP, the deviation caused by the resolution of the vehicle exterior information sensor 1 is taken into account in the determination result. Thus, it is possible to suppress the occurrence of erroneous determination of the correlation between the correlation range RA and the position P of the detection point DP, thereby improving the accuracy of the object route data TD.

[0367] The correlation processing unit 35 determines the object model C based on model1 The object model C is determined by using at least one of the state and size of the object identification element model1 By using the object identification element, the position HP of the candidate point DPH, the predicted position PredP, and the object model C model1 The positional relationship between them becomes clear. Therefore, the object model C mode1l The positional relationship between the reference position BP and the object model C can be determined accurately. model1 The reference position BP in .

[0368] In addition, the correlation processing unit 35 may not be able to obtain the object model C from the vehicle exterior information sensor 1. model1 In this case, the correlation processing unit 35 uses at least one of the width W and the length L of the object model C as the object identification factor. model1 The correlation processing unit 35 determines the value of the object identification factor that cannot be obtained from the vehicle exterior information sensor 1 based on the determined setting value.

[0369] Therefore, even if the object model C cannot be obtained from the vehicle exterior information sensor 1 model1 The route data TD can be updated while suppressing the error by adjusting at least one of the width W and the length L. Thus, the relative positional relationship between the host vehicle and the object will not deviate greatly, thereby minimizing the reduction in the automatic driving accuracy of the host vehicle.

[0370] In addition, the correlation processing unit 35 may not be able to obtain the object model C from the vehicle exterior information sensor 1. model1 In this case, the correlation processing unit 35 uses at least one of the width W, length L and direction θ of the object model C as the object identification factor. model1The correlation processing unit 35 determines the value of the object identification factor that cannot be obtained from the vehicle exterior information sensor 1 based on the determined set value.

[0371] Therefore, even if the object model C cannot be obtained from the vehicle exterior information sensor 1 model1 The route data TD can be updated while suppressing the error by adjusting at least one of the width W, length L and direction θ of the vehicle. As a result, the relative positional relationship between the vehicle and the object will not deviate significantly, thereby minimizing the reduction in the automatic driving accuracy of the vehicle.

[0372] In addition, the correlation processing unit 35 may not be able to obtain the object model C from the vehicle exterior information sensor 1. model1 In this case, the correlation processing unit 35 uses at least one of the width W, length L, direction θ, and height H of the object model C as the object identification factor. model1 The correlation processing unit 35 determines the value of the object identification element that cannot be obtained from the vehicle exterior information sensor 1 based on the determined setting value.

[0373] Therefore, even if at least one of the width W, length L, direction θ, and height H of the object cannot be obtained from the vehicle exterior information sensor 1, the error can be suppressed and the route data TD can be updated. As a result, the relative positional relationship between the host vehicle and the object will not deviate significantly, and the reduction in the automatic driving accuracy of the host vehicle can be suppressed to a minimum.

[0374] In addition, the correlation processing unit 35 may not be able to obtain the object model C from the vehicle exterior information sensor 1. model1 Width W, length L, direction θ, top Z H Position and lower end Z L In this case, the correlation processing unit 35 determines the position of the object model C as the object identification factor. model1 Width W, length L, direction θ, top Z H The position and lower end Z L The correlation processing unit 35 determines the value of the object identification factor that cannot be obtained from the vehicle exterior information sensor 1 based on the determined set value.

[0375] Therefore, even if the object model C cannot be obtained from the vehicle exterior information sensor 1model1 Width W, length L, direction θ, top Z H The position and lower end Z L The course data TD can be updated while suppressing the error by checking at least one of the positions of the host vehicle. Thus, the relative positional relationship between the host vehicle and the object will not deviate greatly, thereby minimizing the reduction in the accuracy of the autonomous driving of the host vehicle.

[0376] In addition, if in addition to the object model C model1 In addition to the width W, length L, and direction θ, the upper end Z is also corrected. H The position and lower end Z L If the position of the object is determined, it can be determined whether the object is a stationary object. A stationary object is, for example, a signboard. A stationary object can be a road sign. Therefore, the type of the object can be determined. As a result, the automatic driving accuracy of the vehicle can be further improved.

[0377] In addition, in the correlation processing unit 35, there may be a plurality of candidate points DPH corresponding to one detection point DP. In this case, the correlation processing unit 35 selects the candidate points DPH based on the reliability DOR of each of the plurality of candidate points DPH and the object model C. model1 The object model C is determined by the positions P of multiple candidate points DPH in model1 The reference position BP in .

[0378] Therefore, the object model C is determined by also considering the reliability DOR of the position HP of the candidate point DPH. model1 Thus, each of the plurality of candidate points DPH can be effectively utilized.

[0379] In addition, the correlation processing unit 35 is based on the object model C model1 The object model C is determined by selecting the position HP of the candidate point DPH with the highest reliability DOR among the positions HP of the candidate points DPH in the object model C. model1 The reference position BP in .

[0380] If an object model C model1 If there are multiple candidate points DPH in , then one object model C model1 The setting accuracy of each of the positions HP of the plurality of candidate points DPH in the object model C may be different. model1 The object model C is determined by selecting the position HP of the candidate point DPH with the largest reliability DOR among the positions HP of the candidate points DPH in the object model C. model1 Therefore, an object model C can be used to model1 Thus, it is possible to use an object model C set based on the resolution of the same vehicle exterior information sensor 1.model1 The position HP of the candidate point DPH with the highest accuracy is set among the positions HP of the plurality of candidate points DPH.

[0381] In addition, the correlation processing unit 35 performs a correlation analysis on an object model C according to each reliability DOR. model1 The weighted average of the positions P of multiple candidate points DPH in the object model C is determined. model1 The reference position BP in.

[0382] If an object model C model1 If there are multiple candidate points DPH in , then one object model C model1 The positions HP of the plurality of candidate points DPH in the object model C are set with different precision. model1 The positions HP of multiple candidate points DPH in the object model C are determined by weighted average model1 Therefore, in one object model C model1 Among the multiple candidate points DPH in the object model C, the influence of the candidate point DPH with lower reliability DOR is weakened, and the influence of the candidate point DPH with higher reliability DOR is enhanced. On this basis, the object model C is determined. model1 Thus, the object model C can be determined based on the respective reliability DORs of the positions HP of the plurality of candidate points DPH in one object set based on the resolution of the same vehicle exterior information sensor 1. model1 The reference position BP in.

[0383] Furthermore, the correlation processing unit 35 obtains each reliability DOR based on the distance from the vehicle exterior information sensor 1 to at least one of the position P of the detection point DP and the reference position BP.

[0384] The resolution of the outside vehicle information sensor 1 becomes a different resolution according to the distance from the outside vehicle information sensor 1 to at least one of the position P of the detection point DP and the reference position BP. For example, when the outside vehicle information sensor 1 is composed of a millimeter wave radar, if the distance to the position P of the detection point DP is short, the possibility that the detection point DP is the closest point is high. On the other hand, if the distance to the position P of the detection point DP is far, the detection point DP will be buried in the resolution unit. Thus, the detection point DP is assumed to be a reflection point reflected from the center of the object. The reference position BP is also the same as the detection point DP. Therefore, the correlation processing unit 35 calculates each reliability based on the distance from the outside vehicle information sensor 1 to at least one of the position P of the detection point DP and the reference position BP. Thus, the reliability DOR can be calculated based on the performance of the outside vehicle information sensor 1.

[0385] In addition, the correlation processing unit 35 is based on the object model C centered at the predicted position PredP. modell The size and object model C model1 The correlation range RA is set based on the statistic of the detection error caused by the vehicle exterior information sensor 1 that is related to the size of .

[0386] Therefore, the correlation range RA reflects the correlation with the object model C. model1 This eliminates the possibility of false correlation between objects of different sizes.

[0387] In addition, the correlation processing unit 35 is based on the object model C centered at the predicted position PredP. model1 The size and object model C model1 The size of the correlation range RA is set based on the statistic of the detection error caused by the vehicle exterior information sensor 1 that is related to the size of the correlation range RA, and the correlation processing unit 35 adjusts the size of the set correlation range RA based on the plurality of reliabilities DOR(N).

[0388] For example, if you use Figure 6 As explained above, when the reliability DOR(1) is within the decision threshold distance D TH2 When the detection error is greater than 0, it is set to 0. In this case, since the reliability DOR(1) is low, the detection error is large. When the detection error is large, the position P of the detection point DP originally included in the correlation range RA deviates from the correlation range RA. Therefore, when considering the detection error, it is necessary to expand the correlation range RA. In this way, the position P of the detection point DP that deviates from the correlation range RA due to the detection error can be included in the correlation range RA.

[0389] On the other hand, the reliability DOR(1) is less than the judgment threshold distance D TH1 In this case, since the reliability DOR(1) is low, the detection error is small. When the detection error is small, it is assumed that the position P of the detection point DP included in the correlation range RA does not deviate from the correlation range RA. Therefore, the correlation range RA can be more or less narrowed in consideration of the detection error. Thus, it is possible to more accurately determine whether there is a correlation between the position P of the detection point DP and the predicted position PredP.

[0390] In addition, the correlation processing unit 35 uses the Euclidean distance du and the Mahalanobis distance d m Whether any one of them exceeds the correlation range RA, it is determined whether there is a correlation between the position P of the detection point DP and the predicted position PredP. Here, the Euclidean distance du is obtained by the difference vector between the position P of the detection point DP and the reference position BP. In addition, the Mahalanobis distance dm is obtained by the position P of the detection point DP and the reference position BP.

[0391] Therefore, the correlation between the position P of the detection point DP and the predicted position PredP can be determined based on simple indices such as the Euclidean distance du and the Mahalanobis distance dm. Therefore, the accuracy of determining the correlation between the position P of the detection point DP and the predicted position PredP can be improved.

[0392] In addition, the correlation processing unit 35 evaluates whether the result of the determination of whether there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate based on any one of the Euclidean distance du and the Mahalanobis distance dm and the plurality of reliabilities DOR(N). Here, the Euclidean distance du is obtained by the difference vector between the position P of the detection point DP and the reference position BP. In addition, the Mahalanobis distance dm is obtained by the position P of the detection point DP and the reference position BP.

[0393] Therefore, the validity of the result of determining whether there is a correlation between the position P of the detection point DP and the predicted position PredP is evaluated on the basis of including the reliability of the determination result which cannot be determined only by the indicators such as the Euclidean distance du and the Mahalanobis distance dm. Therefore, errors in the determination of the correlation between the position P of the detection point DP and the predicted position PredP can be eliminated.

[0394] The correlation processing unit 35 determines the target object model C based on the model2 With object model C model1 The overlap ratio R and multiple reliability DOR(N) are used to evaluate whether the determination result of whether there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate. model1 The predicted position PredP is the center. In addition, the judgment object model C model2 It is obtained by modeling the object centered at the position P of the detection point DP.

[0395] The overlap rate R is higher when the host vehicle and the object move in the same direction than when the host vehicle and the object move in different directions. Therefore, by evaluating whether the result of the determination of whether there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate based on the overlap rate R, objects that may not need to be determined for correlation in the future can be excluded.

[0396] The correlation processing unit 35 is based on the object model C model1 Each vertex and the judgment object model C model2 The minimum value of the sum of the distances between the vertices and the multiple reliability DOR(N) are used to evaluate whether the result of the determination of whether there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate. model1 The predicted position PredP is the center. In addition, the judgment object model Cmodel2 It is obtained by modeling the object centered at the position P of the detection point DP.

[0397] The object model C centered at the predicted position PredP will be obtained model1 The vertices of the object and the judgment object model C centered at the position P of the detection point DP model2 The minimum value of the sum of the distances of the vertices of the detection point DP is attributed to solving the minimum Steiner tree problem. The minimum Steiner tree problem is the shortest network problem. Therefore, the correlation processing unit 35 solves the shortest network problem, and also uses the reliability DOR to evaluate whether the determination result of whether there is a correlation between the position P of the detection point DP and the predicted position PredP is appropriate. As a result, the appropriateness of the determination result of whether there is a correlation between the position P of the detection point DP and the predicted position PredP can be more accurately determined.

[0398] In addition, each embodiment includes a processing circuit for executing the object recognition device 3. Here, the processing circuit can be dedicated hardware or a CPU (also called Central Processing Unit: central processing unit, central processing device, processing device, computing device, microprocessor, microcomputer, processor, DSP) that executes a program stored in a memory.

[0399] Fig.21 FIG. 2 is a diagram illustrating an example of a hardware configuration. Fig.21 In the embodiment, the processing circuit 201 is connected to the bus 202. When the processing circuit 201 is dedicated hardware, for example, a single circuit, a composite circuit, a programmed processor, an ASIC, an FPGA, or a combination thereof is equivalent to the processing circuit 201. The functions of each part of the object recognition device 3 can be implemented by the processing circuit 201, or the functions of each part can be realized by the processing circuit 201 in a combined manner.

[0400] Fig. 22 FIG. 2 is a diagram illustrating another hardware configuration example. Fig. 22In the embodiment, the processor 203 and the memory 204 are connected to the bus 202. When the processing circuit is a CPU, the functions of each part of the object recognition device 3 are realized by software, firmware, or a combination of software and firmware. The software or firmware is expressed in the form of a program and stored in the memory 204. The processing circuit reads the program stored in the memory 204 and executes it, thereby realizing the functions of each part. That is, the object recognition device 3 has a memory 204 for storing a program, and when the program is executed by the processing circuit, the program finally executes the steps of controlling the time measurement unit 31, the data receiving unit 32, the temporary setting unit 33, the prediction processing unit 34, the correlation processing unit 35, and the update processing unit 36. In addition, it can be said that these programs cause the computer to execute the steps or methods for executing the time measurement unit 31, the data receiving unit 32, the temporary setting unit 33, the prediction processing unit 34, the correlation processing unit 35, and the update processing unit 36. Here, the memory 204 corresponds to a nonvolatile or volatile semiconductor memory such as RAM, ROM, flash memory, EPROM, EEPROM, or a magnetic disk, a floppy disk, an optical disk, a compact disk, a mini optical disk, or a DVD.

[0401] In addition, part of the functions of each part of the object recognition device 3 may be implemented by dedicated hardware, while the other part may be implemented by software or firmware. For example, the function of the temporary setting unit 33 may be implemented by a processing circuit as dedicated hardware. In addition, the processing circuit reads and executes the program stored in the memory 204, thereby being able to implement the function of the relevant processing unit 35.

[0402] Therefore, the processing circuit can implement the above functions using hardware, software, firmware or a combination thereof.

[0403] As described above, in the first embodiment, the detection data DD is obtained by the SNN algorithm, the GNN algorithm, the JPDA algorithm, etc. RT and route data TD RT Prediction data TD RTpred An example of the process of whether or not there is a correlation has been described, but the present invention is not particularly limited to this.

[0404] For example, the detection data DD may be determined based on whether the difference between each detection element and each route element is within a predetermined error value e. RT With the predicted data TD RTpred Here, each detection element is included in the detection data DD RT In addition, each route element is included in the prediction data TD RTpred middle.

[0405] Specifically, the correlation processing unit 35 derives the detection data DD RT The position P relative to the vehicle exterior information sensor 1 and the route data TD contained inRT Prediction data TD RTpred The distance difference between the positions P contained in .

[0406] The correlation processing unit 35 derives the detection data DD RT The speed V and the route data TD contained in RT Prediction data TD RTpred The speed difference between the speed V contained in

[0407] The correlation processing unit 35 derives the detection data DD RT The azimuth and course data contained in TD RT Prediction data TD RTpred The azimuth difference between the azimuths contained in .

[0408] The correlation processing unit 35 obtains the square root of the sum of the squares of the distance difference, the speed difference, and the azimuth difference. When the obtained square root exceeds the error amount e, the correlation processing unit 35 determines that there is no correlation. When the obtained square root is less than the error amount e, the correlation processing unit 35 determines that there is a correlation. It is also possible to determine that the detection data DD RT With route data TD RT Prediction data TD RTpred Is there any relevant relationship?

[0409] In addition, for example, the speed over the ground of the detection point DP may be obtained based on the speed V of the detection point DP. The speed over the ground of the detection point DP may be obtained.

[0410] In this case, when it is determined based on the ground speed at the detection point DP that the object detected by the vehicle exterior information sensor 1 is the vehicle C, the width W and length L of the vehicle C may not be included in the object identification elements of the detection data DD.

[0411] In this case, the width W of the vehicle C is set to 2 [m], and the length L of the vehicle C is set to 4.5 [m]. The width W and length L of the vehicle C set in this way are also set values ​​individually set in advance corresponding to the object identification factor that cannot be obtained from the vehicle exterior information sensor 1.

[0412] In addition, the update processing unit 36 ​​may update the route data TD based on the speed V of the detection point DP when the object is detected by the vehicle exterior information sensor 1. Thus, the route data TD can be updated based on the speed V of the detection point DP in consideration of the observation result observed by the vehicle exterior information sensor 1. Thus, the relative positional relationship between the host vehicle and the object can be accurately captured, and thus the automatic driving accuracy of the host vehicle can be further improved.

[0413] Description of symbols

[0414] 1 vehicle exterior information sensor, 2 vehicle information sensor, 3 object recognition device, 4 notification control device, 5 vehicle control device, 31 time measurement unit, 32 data receiving unit, 33 temporary setting unit, 34 prediction processing unit, 35 correlation processing unit, 36 update processing unit.

Claims

1. An object recognition device, characterized in that: include: a prediction processing unit that predicts a position of a movement destination of the tracking object as a predicted position on an object model obtained by modeling the tracking object, based on a trajectory formed by at least one of the plurality of objects moving as the tracking object; a temporary setting unit that sets a position of at least one candidate point on the object model based on a specification of a sensor that detects the tracked object; as well as A correlation processing unit determines a reference position on the object model based on the position of the candidate point and the predicted position, and determines whether there is a correlation between the position of the detection point and the predicted position based on the positional relationship between a correlation range set based on the reference position on the object model and a detection point when the sensor detects at least one of the multiple objects.

2. The object recognition device according to claim 1, characterized in that: The correlation processing unit determines the reference position on the object model based on an object determination element for determining at least one of a state and a size of the object model.

3. The object recognition device according to claim 2, characterized in that: When at least one of the width and length of the object model cannot be obtained from the sensor as the object determination element, the relevant processing unit determines the value of the object determination element that cannot be obtained from the sensor based on the set value corresponding to the object determination element that cannot be obtained from the sensor, which is one of the set values ​​separately pre-set corresponding to the width and length of the object model.

4. The object recognition device according to claim 2, characterized in that: When at least one of the width, length and direction of the object model cannot be obtained from the sensor as the object determination element, the relevant processing unit determines the value of the object determination element that cannot be obtained from the sensor based on the set value corresponding to the object determination element that cannot be obtained from the sensor, which is one of the set values ​​separately pre-set corresponding to the width, length and direction of the object model.

5. The object recognition device according to claim 2, characterized in that: When at least one of the width, length, direction and height of the object model cannot be obtained from the sensor as the object determination element, the relevant processing unit determines the value of the object determination element that cannot be obtained from the sensor based on the set value corresponding to the object determination element that cannot be obtained from the sensor, which is one of the set values ​​separately pre-set corresponding to the width, length, direction and height of the object model.

6. The object recognition device according to claim 2, characterized in that: When at least one of the width, length, direction, upper end position and lower end position of the object model cannot be obtained from the sensor as the object determination element, the relevant processing unit determines the value of the object determination element that cannot be obtained from the sensor based on the set value corresponding to the object determination element that cannot be obtained from the sensor, which is one of the set values ​​separately pre-set corresponding to the width, length, direction, upper end position and lower end position of the object model.

7. The object recognition device according to any one of claims 1 to 6, characterized in that: When the number of the candidate points is plural, the correlation processing unit determines the reference position on the object model based on the reliability of each of the plurality of candidate points and the position of each of the plurality of candidate points on the object model.

8. The object recognition device according to claim 7, characterized in that: The correlation processing unit determines the reference position on the object model based on the position of the candidate point having the highest reliability among the positions of the plurality of candidate points on the object model.

9. The object recognition device according to claim 7, characterized in that: The correlation processing unit determines the reference position on the object model by performing weighted averaging on the positions of the plurality of candidate points on the object model according to the respective reliabilities.

10. The object recognition device according to any one of claims 7 to 9, characterized in that: The correlation processing unit obtains each of the reliabilities based on a distance from the sensor to at least one of the position of the detection point and the reference position.

11. The object recognition device according to any one of claims 7 to 10, characterized in that: The correlation processing unit sets the correlation range based on the size of the object model centered at the predicted position and a statistic of a detection error caused by the sensor and correlated with the size of the object model.

12. The object recognition device according to any one of claims 7 to 10, characterized in that: The correlation processing unit adjusts the size of the correlation range set based on the size of the object model centered at the predicted position and the statistic of the detection error caused by the sensor and related to the size of the object model, based on the plurality of reliabilities.

13. The object recognition device according to claim 11 or 12, characterized in that: The correlation processing unit determines whether there is a correlation between the position of the detection point and the predicted position based on whether the Euclidean distance of the differential vector between the position of the detection point and the reference position and any one of the Mahalanobis distances obtained based on the position of the detection point and the reference position exceeds the correlation range.

14. The object recognition device according to claim 13, characterized in that: The correlation processing unit evaluates whether a result of determining whether there is a correlation between the position of the detection point and the predicted position is appropriate based on either one of the Euclidean distance and the Mahalanobis distance and the plurality of reliabilities.

15. The object recognition device according to claim 13, characterized in that: The correlation processing unit evaluates whether the determination result of whether there is a correlation between the position of the detection point and the predicted position is appropriate based on the overlap rate between the determination object model obtained by modeling the object centered on the position of the detection point and the object model centered on the predicted position, and multiple reliabilities.

16. The object recognition device according to claim 13, characterized in that: The correlation processing unit evaluates whether the judgment result of whether there is a correlation between the position of the detection point and the predicted position is appropriate based on the minimum value of the sum of distances between each vertices of the object model centered on the predicted position and each vertices of the judgment object object model obtained after modeling the object centered on the position of the detection point, and multiple reliabilities.

17. An object recognition method, characterized in that: include: A step of predicting a position of a movement destination of at least one of a plurality of objects as a tracking object as a predicted position on an object model obtained by modeling the tracking object, based on a trajectory formed by the movement of at least one of the plurality of objects as the tracking object; a step of setting a position of at least one candidate point on the object model based on the specification of a sensor detecting the tracked object; as well as A process of determining a reference position on the object model based on the position of the candidate point and the predicted position, and determining whether there is a correlation between the position of the detection point and the predicted position based on the positional relationship between a correlation range set based on the reference position on the object model and a detection point when the sensor detects at least one of the multiple objects.

Images