Vehicle tracking device

Abstract

Vehicle tracking device (10) for tracking a target vehicle (2) driving in the vicinity of a self-propelled vehicle (1), comprising: an estimation part (121) that extrapolates a rectangular frame (7) based on previous data, which is an approximation of the outer shape of the target vehicle (2), an adaptation part (122) which, based on the position of the rectangular frame (7), adapts outline data (4) which are a plurality of detection point data (3) that indicate an outline of the target vehicle (2) detected by a sensor (20) mounted in the own vehicle (1) and which detects the relative position of the object to be detected relative to the own vehicle (1) as detection point data (3), and a determination part (123) that determines the direction of travel of the target vehicle (2) based on the adapted outline data (4).

Description

[Technical field]

[0001] The invention relates to a vehicle tracking device. [State of the art]

[0002] Technologies for tracking target vehicles traveling in the vicinity of a controlled vehicle have been conventionally proposed (see, e.g., JP 2017-129446A). The information obtained by tracking a vehicle (e.g., the target vehicle's position, direction of travel, and speed) is used for control purposes such as collision avoidance.

[0003] Furthermore, techniques for detecting target vehicles using image analysis of camera images are known from DE 10 2013 022 076 A1 and DE 10 2011 119 767 A1. [Summary of the invention][Problem to be solved by the invention]

[0004] During vehicle tracking, the target vehicle's direction of travel is determined at each detection point, and the target vehicle is identified using this determined direction. For example, the target vehicle's current position is estimated based on a previous direction of travel determined at an earlier detection point, and a vehicle detected near this estimated position is identified as the target vehicle.

[0005] When determining the direction of travel of a target vehicle, a sensor mounted on the vehicle itself is sometimes used. This sensor captures the relative position of the target vehicle relative to the vehicle as data points. Specifically, the target vehicle's direction of travel is determined based on outline data, which consists of multiple data points representing the outline of the target vehicle as captured by the sensor. In such cases, the accuracy of determining the target vehicle's direction of travel may decrease, making it difficult to track the target vehicle effectively. For example, determining the direction of travel using outline data generally assumes that the outline data contains multiple data points from the front or rear of the target vehicle.However, it is possible that the outline data may also contain data from a side surface of the target vehicle, in which case there is a risk that the specified direction of travel deviates significantly from the actual direction of travel and it becomes difficult to track the target vehicle appropriately.

[0006] With regard to this problem, the present invention aims to provide a vehicle tracking device with which a vehicle can be tracked in a suitable manner. [Means of solving the problem]

[0007] To solve this problem, the vehicle tracking device according to the invention comprises: an estimation component that extrapolates a rectangular frame based on previous data, representing an approximation of the target vehicle's outer shape, an adaptation part that, based on the position of the rectangular frame, adapts outline data, which consists of a plurality of capture point data, indicating an outline of the target vehicle captured by a sensor mounted in the vehicle and capturing the relative position of the object to be captured relative to the vehicle as capture point data, and a determination component that determines the direction of travel of the target vehicle based on the adapted outline data. [Advantages of the invention]

[0008] The present invention makes it possible to track a vehicle in a suitable manner. [Brief description of the drawings] Fig. Figure 1 is a block diagram showing a simplified arrangement of a self-propelled vehicle in which a vehicle tracking device according to an embodiment of the present invention is mounted. Fig. Figure 2 is a schematic sketch that provides an overview of a processing operation for tracking an oncoming vehicle in the vehicle tracking device according to the embodiment of the present invention. Fig. Figure 3 shows how the oncoming vehicle turns. Fig. Figure 4 is a schematic sketch showing how a vehicle tracking device determines the direction of travel of the oncoming vehicle according to a comparative example in a state where the oncoming vehicle is turning. Fig. Figure 5 is a flowchart showing an exemplary first processing step for determining the direction of travel of the oncoming vehicle in the vehicle tracking device according to the embodiment of the present invention. Fig. Figure 6 is a schematic sketch showing an exemplary rectangular frame extrapolated from an estimating part according to the embodiment of the present invention. Fig. Figure 7 is a schematic sketch showing how outline data is adapted by the first adaptation processing performed by an adaptation part according to the embodiment of the present invention. Fig. Figure 8 is a schematic sketch showing how outline data is adapted by a second adaptation processing performed by an adaptation part according to the embodiment of the present invention. Fig. Figure 9 is a schematic sketch showing how outline data is adapted by a third adaptation processing carried out by the adaptation part according to the embodiment of the present invention. Fig. Figure 10 is a schematic sketch showing how the direction of travel of the oncoming vehicle is determined by a determining part according to the embodiment of the present invention. Fig. Figure 11 is a flowchart showing an exemplary second processing step for determining the direction of travel of the oncoming vehicle in the vehicle tracking device according to the embodiment of the present invention. Fig. Figure 12 is a schematic sketch showing how the position of the rectangular frame according to the embodiment of the present invention differs from the position of the oncoming vehicle. Fig. Figure 13 is a schematic sketch showing an exemplary arrangement of outline data with respect to a base edge of the rectangular frame according to the embodiment of the present invention. Fig. 14 is a schematic sketch showing how in the Fig. In the example shown in Figure 13, according to the embodiment of the present invention, the normal direction of an approximation line segment is corrected by rotation. [Emphasis of the invention]

[0009] A preferred embodiment of the present invention is explained in detail below with reference to the drawings. The dimensions, materials, other specific numerical values, etc., mentioned in the embodiment are merely examples intended to facilitate understanding of the invention. These examples do not limit the present invention unless expressly stated otherwise. In the description and the drawings, components with essentially the same function and arrangement are designated with the same reference numerals to avoid redundant explanations. Furthermore, components that do not directly relate to the present invention have been omitted from the drawings.

[0010] The following example explains how the target vehicle traveling in the vicinity of the user's vehicle, which serves as the object of tracking, is an oncoming vehicle (i.e., a vehicle traveling in the opposite direction to the user's vehicle). However, the target vehicle serving as the object of tracking can also be a vehicle traveling ahead (i.e., a vehicle traveling in the same direction ahead of the user's vehicle). In this case, the oncoming vehicle is replaced by the vehicle traveling ahead in the processing described below for tracking the oncoming vehicle. <Anordnung des Eigenfahrzeugs>

[0011] Based on the Fig. Figures 1 to 4 describe the arrangement of a private vehicle 1 in which a vehicle tracking device 10 is mounted according to the embodiment of the present invention.

[0012] Fig. Figure 1 is a block diagram showing a simplified arrangement of the vehicle 1. The vehicle 1 is equipped with a vehicle tracking device 10, a LIDAR (Laser Imaging Detection and Ranging) 20, and a radar 30. Using the detection data from the LIDAR 20 and radar 30 mounted in the vehicle 1, the vehicle tracking device 10 tracks an oncoming vehicle, which serves as the target vehicle traveling in the vicinity of the vehicle 1. During the tracking of the oncoming vehicle, its position, direction of travel, and speed are determined at each detection point. The information determined during the tracking of the oncoming vehicle is used, for example, for controlling the vehicle 1 to avoid collisions (e.g., by...).a control system to warn of the immediate proximity of a vehicle ahead, an emergency braking control system, an ACC control system or a lane change control system, etc., is used.

[0013] The LIDAR 20 is an example of a sensor according to the present invention for detecting the relative position of a target object relative to the vehicle 1 as detection point data. The LIDAR 20 shines light onto the target object, receives the light reflected from the target object, and thus detects the relative position of the target object as detection point data. Specifically, the LIDAR 20 detects outline data, which consists of a plurality of detection point data that indicate the outline of the oncoming vehicle. The outline data is used by the vehicle tracking device 10 for processing and tracking the oncoming vehicle. The LIDAR 20 is mounted, for example, at the front end of the vehicle 1. The sensor according to the present invention is not limited to the LIDAR 20, but can, for example, be a radar sensor (e.g., a radar 30), an infrared sensor, an ultrasonic sensor, or the like.In short: the method for acquiring data from data points by a sensor according to the present invention is not particularly limited.

[0014] The radar 30 detects the relative speed of the object to be detected with respect to the vehicle 1. The radar 30 detects the relative speed of the object by emitting radio waves onto the object and receiving the radio waves reflected from it. The radar 30 is, for example, mounted at the front of the vehicle 1. Sensors using methods other than the radar 30 can also be used to detect the relative speed of the object.

[0015] The vehicle tracking device 10 includes, among other things, a CPU (Central Processing Unit), which is a computing and processing device, a ROM (Read Only Memory), which is a memory element for storing programs used by the CPU, calculation parameters, etc., and a RAM (Random Access Memory), which is a memory element for temporarily storing parameters, etc., that change according to the circumstances when executed by the CPU.

[0016] The vehicle tracking device 10 includes, for example, a recording part 11 and a tracking part 12.

[0017] The acquisition unit 11 receives information that is used in processing by the tracking unit 12 and outputs it to the tracking unit 12. For example, the acquisition unit 11 receives acquisition results from the LIDAR 20 and radar 30. The LIDAR 20 and radar 30 perform acquisition at each acquisition time point, with predefined time intervals between them. The information acquired by the LIDAR 20 and radar 30 is associated with the acquisition times.

[0018] Tracking part 12 performs tracking processing to track the oncoming vehicle, which serves as the target vehicle traveling in the vicinity of the user vehicle 1. Tracking part 12 comprises an estimation part 121, an adaptation part 122, a determination part 123, and a recognition part 124.

[0019] Estimation part 121 and adjustment part 122 perform the processing described below to optimize the determination of the oncoming vehicle's direction of travel by determination part 123. Specifically, estimation part 121 extrapolates a rectangular frame based on previous data, which approximates the outer shape of the oncoming vehicle. Adjustment part 122 then adjusts the outline data of the oncoming vehicle, acquired by the LIDAR 20, using this rectangular frame.

[0020] Determination part 123 determines various types of information about the oncoming vehicle. Specifically, determination part 123 determines the position, direction of travel, and speed of the oncoming vehicle. The determination of the position, direction of travel, and speed of the oncoming vehicle is performed at each acquisition time of the LIDAR 20 and radar 30. The determination of the position and direction of travel of the oncoming vehicle is based on the outline data of the oncoming vehicle acquired by the LIDAR 20. In particular, the rectangular frame extrapolated by estimation part 121 is used to determine the direction of travel of the oncoming vehicle. The determination of the speed of the oncoming vehicle is performed based on the acquisition results from radar 30.

[0021] The position of the oncoming vehicle, as determined by part 123, can be its relative position to the vehicle 1 or its absolute position. Furthermore, the speed of the oncoming vehicle, as determined by part 123, can be its relative speed to the vehicle 1 or its absolute speed.

[0022] The recognition unit 124 identifies which vehicle is the one detected by the LIDAR 20 and radar 30 (i.e., the vehicle corresponding to the detection results from the LIDAR 20 and radar 30). Specifically, based on information about the oncoming vehicle determined by the determination unit 123, the recognition unit 124 estimates the position of the oncoming vehicle at the next detection time and identifies the vehicle detected near the estimated position at the next detection time as the oncoming vehicle.

[0023] Now, with the help of the Fig. 2 an overview of the processing for tracking an oncoming vehicle 2 in the vehicle tracking device 10 is given. Fig. Figure 2 is a schematic sketch that provides an overview of the processing for tracking the oncoming vehicle 2 in the vehicle tracking device 10. Fig. Figure 2 shows the vehicle 1 and the oncoming vehicle 2, which serves as the target vehicle traveling in the vicinity of vehicle 1. The x'-y' coordinate system is the coordinate system of the LIDAR 20. The x'-axis corresponds to the direction of travel of vehicle 1, and the y'-axis lies horizontally orthogonal to the x'-axis.

[0024] As in Fig. As shown in Figure 2, during processing for tracking the oncoming vehicle 2, outline data 4, which consists of a plurality of capture point data 3 showing the outline of the oncoming vehicle 2, are captured by the LIDAR 20. The plurality of points on a frontal surface of the oncoming vehicle 2 in Fig. 2 corresponds to a plurality of data points 3. Among the outline data 4, a plurality of data points 3 are fundamentally data from the frontal area of ​​the oncoming vehicle 2 (i.e., data indicating the relative positions of points on the frontal area of ​​the oncoming vehicle 2 with respect to the vehicle 1). Using such outline data 4, the position and direction of travel of the oncoming vehicle 2 are determined.

[0025] The determining element 123 determines a direction of travel Di at a given acquisition time Ti based on the distribution of the majority of acquisition point data 3 among the outline data 4. Then, the position of a center point of the oncoming vehicle 2, which was predicted under the assumption that a majority of the acquisition point data 3 among the outline data 4 are data from the frontal area of ​​the oncoming vehicle 2, is determined as the position Pi of the oncoming vehicle 2. That is, the position Pi of the oncoming vehicle 2 is determined based on the distribution of the majority of acquisition point data 3 among the outline data 4. Fig. Figure 2 shows a direction of travel Di-1 and a position Pi-1 of the oncoming vehicle 2, which were determined in the same way at a detection time Ti-1 preceding the detection time Ti. Furthermore, in Fig. 2 a direction of travel Di+1 and a position Pi+1 of the oncoming vehicle 2 are shown, which are determined in the same way at a detection time Ti+1 following the detection time Ti.

[0026] The detection unit 124 estimates the position of the oncoming vehicle 2 at detection time Ti based on the direction of travel Di-1 and the position Pi-1 of the oncoming vehicle 2, which were determined at the previous detection time Ti-1. Specifically, the detection unit 124 estimates the position of the oncoming vehicle 2 at detection time Ti as the position that the oncoming vehicle 2 would reach if it continued traveling from position Pi-1 in the direction of travel Di-1 between detection times Ti-1 and Ti-1 at the speed present at detection time Ti-1. The detection unit 124 then decides, by comparing the estimated position of the oncoming vehicle 2 at detection time Ti with position Pi, whether a vehicle detected at position Pi at detection time Ti is the oncoming vehicle 2 or not.

[0027] For example, if position Pi is quite close to the estimated position of the oncoming vehicle 2 at the detection time Ti, the detection unit 124 recognizes the vehicle detected at position Pi at the detection time Ti as the oncoming vehicle 2. Conversely, if position Pi is far from the estimated position of the oncoming vehicle 2 at the detection time Ti, the detection unit 124 recognizes the vehicle detected at position Pi at the detection time Ti as a different vehicle from the oncoming vehicle 2.In the same way, the recognition unit 124 predicts an estimated position of the oncoming vehicle 2 for the detection time Ti+1 based on the direction of travel Di and position Pi of the oncoming vehicle 2 determined at the detection time Ti+1, compares the estimated position of the oncoming vehicle 2 for the detection time Ti+1 and the position Pi+1 and thereby decides whether a vehicle detected at position Pi+1 at the detection time Ti+1 is the oncoming vehicle 2 or not.

[0028] By tracking the oncoming vehicle 2 in this way, various pieces of information about it can be determined at each point in time. With conventional technology, the accuracy in determining the direction of travel of the oncoming vehicle 2 can decrease, making it difficult to track it effectively. For example, the outline data 4 might also contain data from a side surface of the oncoming vehicle 2. With conventional technology, this could lead to the determined direction of travel deviating from the actual direction, making it difficult to track the oncoming vehicle 2 effectively.Specifically, when determining the direction of travel of the oncoming vehicle 2, individual data points 3 acquired by the LIDAR 20 are assigned to those outer surfaces of the oncoming vehicle 2 that face the LIDAR 20 (a maximum of two surfaces), whereby the majority of data points 3 assigned to the front surface of the oncoming vehicle 2 correspond to the outline data 4. However, it is possible that the assignment of the data points 3 to the surfaces is not carried out correctly, so that the outline data 4 may contain data from a side surface adjacent to the front surface as a result of the assignment.

[0029] A situation in which data from a side surface of the oncoming vehicle 2 can easily be included in the outline data 4 is, for example, a situation in which the oncoming vehicle 2 is turning. The following will illustrate this using the... Fig. 3 and Fig. 4 explains the case where, according to a comparative example, a vehicle tracking device determines the direction of travel of the oncoming vehicle 2 in a state in which the oncoming vehicle 2 is turning. Fig. Figure 3 shows how the oncoming vehicle 2 turns. Fig. Figure 4 is a schematic sketch showing how a vehicle tracking device according to the comparative example determines the direction of travel of the oncoming vehicle 2 in a state in which the oncoming vehicle 2 is turning.

[0030] As in Fig. As shown in Figure 3, it can happen that an oncoming vehicle 2 approaching the vehicle 1 turns to change its direction of travel. Fig. Figure 4 shows the outline data 4, which is captured by the LIDAR 20 in a state where the oncoming vehicle 2 is turning. In the Fig. In the outline data shown in section 4, points in an area A1, which represents a peripheral region of the frontal area of ​​the oncoming vehicle 2 (specifically, the peripheral region on the side furthest from the own vehicle 1), are not recorded as detection point data 3. On the other hand, in the Fig. The outline data shown represents 4 points in an area A2, which is part of a side surface of the oncoming vehicle 2, as detection point data 3.

[0031] The vehicle tracking device according to the comparative example determines the normal direction of an approximation line segment 5, which represents an approximation of a plurality of detection point data 3 under the outline data 4, as the direction of travel of the oncoming vehicle 2. The approximation line segment 5 corresponds, for example, to that part of an approximation line representing an approximation of the plurality of detection point data 3, which is divided where the points at both ends of the plurality of detection point data 3 are projected onto the approximation line. However, as in Fig. Figure 4 shows that the approximate line segment 5 determined by the vehicle tracking device according to the comparison example is highly inclined relative to the line segment 6, which represents an approximation of the actual frontal area of ​​the oncoming vehicle 2. This is due to the fact that the outline data 4 also includes data on a side surface of the oncoming vehicle 2 (i.e., the data in area A2) and that part of the outline data 4 (i.e., the data in area A1) is missing. Therefore, the determined direction of travel of the oncoming vehicle 2 deviates significantly from its actual direction of travel.

[0032] In the vehicle tracking device 10 according to the present embodiment, the estimation part 121 extrapolates a rectangular frame that approximates the outer shape of the oncoming vehicle 2. The determination part 123 then determines the direction of travel of the oncoming vehicle 2 based on the outline data 4 of the oncoming vehicle 2 acquired by the LIDAR 20 and the rectangular frame. This prevents the determined direction of travel from deviating from the actual direction of travel, thus enabling the oncoming vehicle 2 to be tracked appropriately. The processing operations performed by the vehicle tracking device 10 to determine the direction of travel of the oncoming vehicle 2 are explained in detail below.

[0033] As described above, the vehicle tracking device 10 communicates with all devices mounted in the vehicle 1. Communication between the vehicle tracking device 10 and the various devices is implemented, for example, via CAN (Controller Area Network) communication.

[0034] The functions of the vehicle tracking device 10 can be distributed across multiple devices, or multiple functions can be implemented by a single device. If the functions of the vehicle tracking device 10 are distributed across multiple devices, these devices may be able to communicate with each other. <Betrieb der Fahrzeugverfolgungseinrichtung>

[0035] The operation of the vehicle tracking device 10 according to the embodiment of the present invention is described by reference to the Fig. Explained in sections 5 to 14.

[0036] Since the direction of travel of the oncoming vehicle 2 can be determined with high accuracy by the vehicle tracking device 10, as described above, tracking of the oncoming vehicle 2 is implemented in a suitable manner. As examples of processing operations performed by the vehicle tracking device 10 to determine the direction of travel of the oncoming vehicle 2, a first and a second operation are explained sequentially. The vehicle tracking device 10 can perform either only the first operation or only the second operation, or it can perform both (e.g., the first and second operations can also be performed sequentially). [First processing]

[0037] First, the sequence of the initial processing to determine the direction of travel of the oncoming vehicle 2 is described using the Fig. 5 to 10 are explained. During the first processing, the outline data 4 are adjusted based on the position of a rectangular frame 7, and the direction of travel of the oncoming vehicle 2 is determined based on the adjusted outline data 4.

[0038] Fig. Figure 5 is a flowchart showing an exemplary sequence of the initial processing performed by the vehicle tracking device 10 to determine the direction of travel of the oncoming vehicle 2. Steps S110 and S150 in Fig. 5 each correspond to starting and ending the in Fig. 5 control sequence shown.

[0039] Will the in Fig. 5. The control sequence shown is started, so in step S120 the estimation part 121 extrapolates a rectangular frame that represents an approximation of the outer shape of the oncoming vehicle 2.

[0040] Fig. Figure 6 is a schematic sketch showing an exemplary rectangular frame 7 extrapolated from the estimation part 121. Fig. Figure 6 shows the vehicle 1 and the rectangular frame 7 extrapolated from the estimation part 121. The size ratio of the vehicle 1 and the rectangular frame 7 in Fig. Figure 6 was achieved by appropriately modifying the actual size ratio and does not correspond to the actual size ratio. The xy-coordinate system has an x-axis, which is a coordinate axis in a transverse direction of the vehicle within the rectangular frame 7, and a y-axis, which is a coordinate axis in a longitudinal direction of the vehicle within the rectangular frame 7. The y-axis is orthogonal to the x-axis. The positive direction of the x-axis is the direction away from the vehicle 1, and the negative direction of the x-axis is the direction towards the vehicle 1. The positive direction of the y-axis is the direction away from the vehicle 1, and the negative direction of the y-axis is the direction towards the vehicle 1.In the following, the direction of the x-axis is also referred to as the vehicle transverse direction x of the rectangular frame 7 and the direction of the y-axis is also referred to as the vehicle longitudinal direction y of the rectangular frame 7.

[0041] The estimation part 121 extrapolates the rectangular frame 7 based on a previous direction of travel determined by the determination part 123 using earlier outline data 4. Specifically, the estimation part 121 assumes that the position of a center point of the oncoming vehicle 2 is located at the estimated position of the oncoming vehicle 2 estimated by the recognition part 124 for the detection time Ti, and extrapolates as the rectangular frame 7 for the detection time Ti a rectangular frame which, assuming that the oncoming vehicle 2 points in the direction of travel Di-1 of the oncoming vehicle 2 determined at the previous detection time Ti-1, represents an approximation of the outer shape of the oncoming vehicle 2 in the plane xy.

[0042] The rectangular frame 7 comprises two edges N1, N2 along the vehicle's transverse direction x and two edges N3, N4 along the vehicle's longitudinal direction y. Edge N1 lies on the side of edge N2 facing the vehicle 1. Edge N3 lies on the side of edge N4 facing the vehicle 1. The length of the rectangular frame 7 in the vehicle's transverse direction x is set, for example, to an average value of the widths of vehicles of different types. The length of the rectangular frame 7 in the vehicle's longitudinal direction y is set, for example, to an average value of the lengths of vehicles of different types.

[0043] Fig. Figure 6 shows the outline data 4 acquired by the LIDAR 20. As described above, the outline data 4 consists of a plurality of acquisition point data 3 that indicate the outline of the oncoming vehicle 2. Specifically, the outline data 4 comprises a plurality of data networks 8, which in turn comprise a plurality of acquisition point data 3 whose vertical positions correspond to each other (more precisely, coincide). If the vertical positions of individual acquisition point data 3 coincide, it is possible that the vertical positions of the individual acquisition point data 3 will cluster within a given area. The vertical position refers to the position in the perpendicular direction. The vertical positions of the acquisition point data 3 differ between several data networks 8. In the Fig. In the example shown in Figure 6, the outline data 4 comprise three data sets 8: a data set 8a containing a plurality of data points 3a, a data set 8b containing a plurality of data points 3b, and a data set 8c containing a plurality of data points 3c. Fig. 6 and those described below Fig. 7 to 10 are individual data sets 8 distinguished from each other by different hatching densities.

[0044] The vertical positions of the data points 3 of the individual data sets 8 increase sequentially from data set 8a to data set 8b to data set 8c. In short: data set 8a shows the outline of the oncoming vehicle 2 at a lower position compared to data sets 8b and 8c. Data set 8c shows the outline of the oncoming vehicle 2 at a higher position compared to data sets 8a and 8b. Data set 8b shows the outline of the oncoming vehicle 2 at a position between the vertical position of data set 8a and the vertical position of data set 8c. In short: the individual data sets 8 show the outlines of parts of the oncoming vehicle 2 that are located at different heights (e.g., bumper, hood, etc.). Fig. Figure 6 shows an example in which the number of data groups 8 contained in the outline data 4 is three; however, the number of data groups 8 contained in the outline data 4 can also be other than three.

[0045] Next, in step S130, the adaptation part 122 adjusts the outline data 4.

[0046] The adaptation part 122 adjusts the outline data 4, for example, in a state where the coordinate system of the outline data 4 has been changed from the coordinate system of sensor 20, i.e., the x'-y' coordinate system, to the xy coordinate system. The following is an example in which the outline data 4 is adapted in the xy coordinate system. However, the adaptation of the outline data 4 can also be performed in the x'-y' coordinate system, which is the coordinate system of sensor 20.

[0047] The adjustment section 122 performs three adjustment processes sequentially to adjust the outline data 4: a first adjustment process, a second adjustment process, and a third adjustment process. The following sections explain the first, second, and third adjustment processes in sequence.

[0048] Fig. Figure 7 is a schematic sketch showing how the outline data 4 is adapted by the first adaptation processing performed by the adaptation part 122.

[0049] During the first adaptation processing, the adaptation part 122 deletes those acquisition point data 3 of the data networks 8 that lie in the longitudinal direction y of the rectangular frame 7 on the side furthest from the own vehicle 1 (i.e., on the side positive in the longitudinal direction y of the vehicle) from the outline data 4. Fig. Figure 7 shows how the first adaptation processing is performed on the data set 8a of the data sets 8a, 8b, 8c.

[0050] Here, for example, the adaptation part 122 deletes from the outline data 4 those data points 3a of the data set 8a that are located further away from the vehicle 1 in the longitudinal direction y of the rectangular frame 7 than a first basic position y1 (i.e., data points 3a on the side positive in the longitudinal direction y of the vehicle) than the first basic position y1. The first basic position y1 is located further away from the vehicle 1 in the longitudinal direction y of the rectangular frame 7 than an average position y2 of the data points 3a of the data set 8a. The difference between the average position y2 and the first basic position y1 is adjusted to a value that allows data contained in the data set 8a to be appropriately deleted from the side surface of the oncoming vehicle 2.

[0051] During the initial customization processing, the in Fig. 7. Data point 3a located in area A3 is deleted from data set 8a. The data point 3a deleted in this way, located in area A3, is highly likely to be data from the side of the oncoming vehicle 2. It follows that the data from the side of the oncoming vehicle 2 contained in data set 8a can also be deleted. The first adjustment processing is performed in the same way for data sets 8b and 8c as for data set 8a. The data from the side of the oncoming vehicle 2 contained in the respective data sets 8 are a significant factor in the decrease in the accuracy of determining the direction of travel of the oncoming vehicle 2. It follows that deleting such data from the outline data 4 can appropriately improve the accuracy of determining the direction of travel of the oncoming vehicle 2.

[0052] Fig. Figure 8 is a schematic sketch showing how the outline data 4 is adapted by the second adaptation processing performed by the adaptation part 122.

[0053] During the second adaptation processing, the adaptation part 122 deletes those data groups 8 of the data groups 8a, 8b, 8c, which lie in the longitudinal direction y of the rectangular frame 7 on the side furthest from the own vehicle 1 (i.e. on the side positive in the longitudinal direction y of the vehicle), from the outline data 4.

[0054] Here, the adaptation part 122 deletes, for example, those data groups 8 of the majority of data groups 8a, 8b, 8c for which the average position of the acquisition point data 3 in the vehicle longitudinal direction y of the rectangular frame 7 is further away from the vehicle 1 than a second basic position y3 (i.e., data groups 8 on the side positive in the vehicle longitudinal direction y) from the outline data 4. The second basic position y3 is further away from the vehicle 1 than the average position y4 of the acquisition point data 3 of data group 8a, for which the average position of the acquisition point data 3 is closest to the vehicle 1 in the vehicle longitudinal direction y of the rectangular frame 7.The difference between the average position y4 and the second base position y3 is adjusted, as appropriate, to a value at which those data sets 8 can be suitably deleted that include data point data 3 with excessively high vertical positions. During the second adjustment processing, the average position of the data point data 3 of each data set 8 in the vehicle longitudinal direction y of the rectangular frame 7 is determined using the data point data 3 of the data sets 8 remaining after the first adjustment processing (i.e., data point data 3 excluding the data point data 3 deleted during the first adjustment processing).

[0055] In the Fig. In example 8, an average position y5 of the data point data 3 of data network 8b is closer to the vehicle 1 in the longitudinal direction y of the rectangular frame 7 than the second basic position y3. Conversely, an average position y6 of the data point data 3 of data network 8c is farther from the vehicle 1 in the longitudinal direction y of the rectangular frame 7 than the second basic position y3. It follows that, using the second adjustment processing of the data shown in Fig. Data group 8c, located in area A4, is deleted from the outline data 4.

[0056] Among the outline data 4, those detection point data 3 with excessively high vertical positions are highly likely to be data from parts of the oncoming vehicle 2 that do not form its outer shape (e.g., hood, windshield, etc.). It follows that, using the second adjustment processing, those data sets 8 that include detection point data 3 with excessively high vertical positions (e.g., data set 8c) can be deleted from the outline data 4. The data sets 8 containing detection point data 3 with excessively high vertical positions are a significant factor in the decrease in the accuracy of determining the direction of travel of the oncoming vehicle 2. Therefore, deleting such data from the outline data 4 can appropriately improve the accuracy of determining the direction of travel of the oncoming vehicle 2.

[0057] Fig. Figure 9 is a schematic sketch showing how the outline data 4 is adapted by the third adaptation processing performed by the adaptation part 122.

[0058] In the third adaptation processing, the adaptation part 122 adjusts the outline data 4 such that the majority of data networks 8 approach each other in the longitudinal direction y of the rectangular frame 7.

[0059] The adaptation part 122 adjusts the outline data 4, for example, such that the average positions of the individual data point data 3 of the majority of data networks 8a, 8b, 8c coincide in the vehicle longitudinal direction y of the rectangular frame 7. In the third adaptation processing, the average position of the data point data 3 of the individual data networks 8 in the vehicle longitudinal direction y of the rectangular frame 7 is determined using the data point data 3 of the data networks 8 existing after the first adaptation processing (i.e., data point data 3 excluding the data point data 3 deleted during the first adaptation processing).

[0060] For example, if data group 8c was deleted from outline data 4 during the second adjustment processing, then adjustment part 122 is shifted during the third adjustment processing, as shown in Fig. Figure 9 shows that the data set 8b is shifted in the vehicle's longitudinal direction y in a negative direction such that it approaches the data set 8a. Specifically, the individual data points 3b of the data set 8b are shifted in the vehicle's longitudinal direction y in a negative direction such that the average position of the data points 3b of the data set 8b in the vehicle's longitudinal direction y of the rectangular frame 7 reaches the average position y4.

[0061] Then, the position of the center point of the oncoming vehicle 2, which was predicted under the assumption that a majority of the detection point data 3 among the outline data 4 are data from the frontal area of ​​the oncoming vehicle 2, is determined as the position of the oncoming vehicle 2. It follows that, with regard to the proper determination of the position of the oncoming vehicle 2, it is preferable to shift the individual data sets 8 in the longitudinal direction y of the vehicle in such a way that they approach the data set 8a that is closest to the own vehicle 1.

[0062] As described below, the determination of the direction of travel of the oncoming vehicle 2 is specifically carried out based on the approximation line segment 5, which represents a linear approximation of the majority of acquisition point data 3 among the outline data 4 available after the third fitting process. It follows that the acquisition point data 3 of the majority of data networks 8 can be used to determine the direction of travel of the oncoming vehicle 2 by fitting the outline data 4 such that the majority of data networks 8 approximate each other in the vehicle longitudinal direction y of the rectangular frame 7. Consequently, even if the number of acquisition point data 3 with corresponding vertical positions (i.e.,Since the number of data points 3 in the respective data networks 8) is relatively small, the accuracy in determining the direction of travel of the oncoming vehicle 2 can be improved in a suitable manner.

[0063] Next, in step S140, the determining part 123 determines the direction of travel of the oncoming vehicle 2, and the one in Fig. The control sequence shown in point 5 is terminated.

[0064] Fig. Figure 10 is a schematic sketch showing how the direction of travel Di of the oncoming vehicle 2 is determined by the determining part 123 at a detection time Ti.

[0065] The determining part 123 determines the direction of travel Di based on the outline data 4 adapted by the adaptation part 122. Specifically, the determining part 123 determines the normal direction of the approximation line segment 5, which represents a linear approximation of the majority of acquisition point data 3 under the adapted outline data 4, as the direction of travel Di. The approximation line segment 5 corresponds, for example, to that part of the approximation line representing an approximation of the majority of acquisition point data 3 under the adapted outline data 4, which is divided where the points at both ends of the majority of acquisition point data 3 are projected onto the approximation line. The determining part 123 determines the approximation line segment 5, for example, using the method of least squares. Fig. Figure 10 shows an approximation line segment 5 in the case that, during the second adaptation processing, the data set 8c was deleted from the outline data 4 and, during the third adaptation processing, the data set 8b was shifted in the longitudinal direction y of the vehicle so that it approaches the data set 8a.

[0066] In this way, during the first adaptation processing, the outline data 4 is adjusted based on the position of the rectangular frame 7, and the direction of travel of the oncoming vehicle 2 is determined based on the adapted outline data 4. This allows the outline data 4 to be optimized as the data used for the direction of travel of the oncoming vehicle 2, and the direction of travel of the oncoming vehicle 2 can be determined with high accuracy. Consequently, a deviation of the determined direction of travel from the actual direction of travel is prevented, so that the oncoming vehicle 2 can be tracked appropriately. [Second processing]

[0067] Next, the process of the second processing step to determine the direction of travel of the oncoming vehicle 2 will be described using the Fig. Sections 11 to 14 explain. In the second processing step, the direction of travel of the oncoming vehicle 2 is determined based on the arrangement of the outline data 4 with respect to a base edge, which is that of the edges of the rectangular frame 7 running in the transverse direction x of the vehicle and facing the own vehicle 1 (i.e., edge N1). Hereinafter, edge N1 is also referred to as base edge N1.

[0068] Fig. Figure 11 is a flowchart showing an exemplary sequence of the second processing step performed by the vehicle tracking device 10 to determine the direction of travel of the oncoming vehicle 2. Steps S210 and S240 in Fig. 11 each correspond to starting and ending the in Fig. 11 control sequence shown.

[0069] Will the in Fig. 11. The control sequence shown is started, so in step S220 the estimation part 121 extrapolates a rectangular frame 7, which represents an approximation of the outer shape of the oncoming vehicle 2.

[0070] The estimation part 121 extrapolates, as in step S120 of the first processing described above, a rectangular frame 7 that represents an approximation of the outer shape of the oncoming vehicle 2. Specifically, the estimation part 121 assumes that the position of the center point of the oncoming vehicle 2 is located at the estimated position of the oncoming vehicle 2 determined by the recognition part 124 for the detection time Ti, and extrapolates as a rectangular frame 7 for the detection time Ti a rectangular frame which, assuming that the oncoming vehicle 2 is pointing in the direction of travel Di-1 of the oncoming vehicle 2 determined at the previous detection time Ti-1, represents an approximation of the outer shape of the oncoming vehicle 2 in the plane xy.As described above, depending on the relative positions of the vehicle 1 and the oncoming vehicle 2, the intended direction of travel of the oncoming vehicle 2 may deviate more easily from its actual direction of travel. In such a case, the position of the rectangular frame 7 is more likely to differ from the position of the oncoming vehicle 2.

[0071] Fig. Figure 12 is a schematic sketch showing how the position of the rectangular frame 7 differs from the position of the oncoming vehicle 2.

[0072] In Fig. Figure 12 shows the change in the detection point data 3 under the outline data 4, which occurs when a direction of travel D0 of the oncoming vehicle 2 is maintained as the direction towards the own vehicle 1 and the oncoming vehicle 2 passes a point relatively close to the own vehicle 1. As in the example above, Fig. 3 and Fig. In the case described in section 4, if the oncoming vehicle 2 turns, it will also be difficult in this case to record points on one of the outer sides (specifically on the side further away from the own vehicle 1) of the front surface of the oncoming vehicle 2 as recording point data 3.

[0073] In the Fig. In the example shown, at a detection time T10, detection point data 3 are recorded across all areas of the frontal surface of the oncoming vehicle 2. However, at a detection time T20 following detection time T10, points in an area A5 on one of the edges of the frontal surface of the oncoming vehicle 2 are not recorded as detection point data 3. Furthermore, at a detection time T30 following detection time T20, points in an area A6 on one of the edges of the frontal surface of the oncoming vehicle 2, which is even larger than area A5, are not recorded as detection point data 3.

[0074] Since the direction of travel of the oncoming vehicle 2 is determined based on the approximation line segment 5 that approximates the outline data 4, if it becomes difficult to capture points on one of the edge sides (specifically on the side furthest from the vehicle 1) of the frontal surface of the oncoming vehicle 2 as detection point data 3, the determined direction of travel of the oncoming vehicle 2 deviates more easily from the actual direction of travel D0. In the Fig. In the example shown in Figure 12, the direction of travel D10 determined at detection time T10, the direction of travel D20 determined at detection time T20, and the direction of travel D30 determined at detection time T30 change in that order such that they rotate clockwise when viewed from above. If the determined direction of travel of the oncoming vehicle 2 changes in this way so that it deviates from the actual direction of travel D0, then the position of the rectangular frame 7 deviates relative to the position of the oncoming vehicle 2 due to this change in the determined direction of travel of the oncoming vehicle 2. In the example shown in Fig. In the example shown in Figure 12, the rectangular frame 7 changes such that it rotates clockwise in top view from the time of acquisition T10 to the time of acquisition T20 to the time of acquisition T30 in this sequence.

[0075] In the second processing step, the direction of travel of the oncoming vehicle 2 can be determined with high accuracy by means of a processing step 123, as described below, which is carried out by the determining part 123 and which is aimed at the feature that the position of the rectangular frame 7 deviates from the position of the oncoming vehicle 2 due to the change in the determined direction of travel of the oncoming vehicle 2, even in the case that the determined direction of travel of the oncoming vehicle 2 deviates more easily from the actual direction of travel.

[0076] Next, in step S230, the determining part 123 corrects the normal direction of the approximation line segment 5, which represents a linear approximation of a plurality of detection point data 3 under the outline data 4, by rotation and determines the direction resulting from the rotational correction of the normal direction of the approximation line segment 5 as the direction of travel of the oncoming vehicle 2. Now the process ends in Fig. The control sequence shown in Figure 11 is as follows. The rotational correction of the normal direction of the approach line segment 5 is performed based on the arrangement of the outline data 4 relative to the base edge N1 of the rectangular frame 7.

[0077] Fig. Figure 13 is a schematic sketch showing an exemplary arrangement of outline data 4 in relation to the base edge N1 of the rectangular frame 7. Fig. 14 is a schematic sketch showing how in the Fig. In the example shown in Figure 13, the normal direction of the approximation line segment 5 is corrected by rotation. As described above, the approximation line segment 5 corresponds, for example, to that part of the approximation line representing an approximation of the majority of acquisition point data 3 under the outline data 4, which is divided where the points at both ends of the majority of acquisition point data 3 are projected onto the approximation line.

[0078] In the following, one end of the approach line segment 5 in the vehicle transverse direction x of the rectangular frame 7 (specifically, in the positive direction of the vehicle transverse direction x, i.e., on the side farther from the own vehicle 1) is referred to as the first line end E51, and the other end of the approach line segment 5 in the vehicle transverse direction x of the rectangular frame 7 (specifically, in the negative direction of the vehicle transverse direction x, i.e., on the side closer to the own vehicle 1) is referred to as the second line end E52. Furthermore, one end of the base edge N1 in the vehicle transverse direction x of the rectangular frame 7 is referred to as the first rectangular frame end E71, and the other end of the base edge N1 in the vehicle transverse direction x of the rectangular frame 7 is referred to as the second rectangular frame end E72.

[0079] The determining element 123 corrects the normal direction of the approximation line segment 5 by rotating it, based on the result of a comparison of the positional ratio of the first line end E51 of the approximation line segment 5 to the first rectangular frame end E71 of the base edge N1 and the positional ratio of the second line end E52 of the approximation line segment 5 to the second rectangular frame end E72 of the base edge N1. Specifically, the determining element 123 corrects the normal direction of the approximation line segment 5 by rotating it, based on the result of a comparison of the distance between the first line end E51 and the first rectangular frame end E71 and the distance between the second line end E52 and the second rectangular frame end E72.

[0080] For example, the determining part 123 uses a first angle θ1 as an indicator for the distance between the first line end E51 and the first rectangular frame end E71. This angle is formed between the direction running from the center point C1 of the rectangular frame 7 to the first line end E51 and the direction running from the center point C1 of the rectangular frame 7 to the first rectangular frame end E71. Furthermore, the determining part 123 uses a second angle θ2 as an indicator for the distance between the second line end E52 and the second rectangular frame end E72. This angle is formed between the direction running from the center point C1 of the rectangular frame 7 to the second line end E52 and the direction running from the center point C1 of the rectangular frame 7 to the second rectangular frame end E72. The determining part 123 corrects, for example...based on the result of a comparison of the first angle θ1 and the second angle θ2, the normal direction of the approximation line segment 5 is determined by rotation.

[0081] If the distance between the second line end E52 and the second rectangular frame end E72 is greater than the distance between the first line end E51 and the first rectangular frame end E71, the determining part 123 corrects the normal direction of the approximation line segment 5 by rotating it around the midpoint C1 of the rectangular frame 7 in the direction from the second rectangular frame end E72 to the first rectangular frame end E71 (in the example according to Fig. 13 (in top view counterclockwise). On the other hand, if the distance between the first line end E51 and the first rectangular frame end E71 is greater than the distance between the second line end E52 and the second rectangular frame end E72, the determining part 123 corrects the normal direction of the approximation line segment 5 by rotating it around the midpoint C1 of the rectangular frame 7 in the direction from the first rectangular frame end E71 to the second rectangular frame end E72 (in the example according to Fig. 13 (in top view, rotating clockwise)

[0082] For example, the determining part 123 corrects the normal direction of the approximation line segment 5 by rotating it around the midpoint C1 of the rectangular frame 7 from the rectangular frame end belonging to the larger of the angles first angle θ1 and second angle θ2 to the rectangular frame end belonging to the smaller of the angles first angle θ1 and second angle θ2. In the example according to Fig. 13. The second angle θ2 is larger than the first angle θ1. It follows that the determining part 123 corrects the normal direction of the approximation line segment 5 by adjusting it as in Fig. 14 shown, rotating counterclockwise in top view. Fig. Figure 14 shows a rotation angle θcorr of the rotation correction, a normal direction Dn of the approximation line segment 5 before the rotation correction, a normal direction Dn' of the approximation line segment 5 after the rotation correction, and an approximation line segment 5' corresponding to this normal direction Dn' (i.e., the line segment obtained by rotating the approximation line segment 5 by exactly the rotation angle θcorr about the center point C1). On the other hand, if the first angle θ1 is larger than the second angle θ2, the determining element 123 corrects the normal direction of the approximation line segment 5 by rotating it clockwise in the top view.

[0083] If, as above based on the Fig. As explained in section 12, if the distance between the second line end E52 and the second rectangular frame end E72 is greater than the distance between the first line end E51 and the first rectangular frame end E71, this corresponds to the case where the specified direction of travel is more easily in the direction of a rotation about the center point C1 of the rectangular frame 7 from the first rectangular frame end E71 to the second rectangular frame end E72 (in the example according to Fig. 13 in top view clockwise) deviates from the actual direction of travel. It follows that by rotating the normal direction of the approximation line segment 5 about the center point C1 of the rectangular frame 7 in the direction from the second rectangular frame end E72 to the first rectangular frame end E71 (in the example according to Fig. 13 (in top view counterclockwise) the accuracy of determining the direction of travel of the oncoming vehicle 2 can be improved.

[0084] If, on the other hand, the distance between the first line end E51 and the first rectangular frame end E71 is greater than the distance between the second line end E52 and the second rectangular frame end E72, this corresponds to the case where the determined direction of travel is more easily in the direction of a rotation about the center point C1 of the rectangular frame 7 from the second rectangular frame end E72 to the first rectangular frame end E71 (in the example according to Fig. 13 in top view counterclockwise) deviates from the actual direction of travel. It follows that by rotating the normal direction of the approximation line segment 5 about the center point C1 of the rectangular frame 7 in the direction from the first rectangular frame end E71 to the second rectangular frame end E72 (in the example according to Fig. 13 (top view clockwise) the accuracy of determining the direction of travel of the oncoming vehicle 2 can be improved.

[0085] Here, the determining part 123, based on the larger of the angles first angle θ1 and second angle θ2, defines the rotation angle θcorr of the rotation correction of the normal direction of the approximation line segment 5. For example, in the example according to Fig. 13 the determining part 123 the rotation angle θcorr based on the second angle θ2 is fixed.

[0086] Specifically, the determining part 123 defines an angle, obtained by multiplying a gain G by the larger of the angles first angle θ1 and second angle θ2, as the rotation angle θcorr of the rotation correction of the normal direction of the approximation line segment 5. For example, in the example according to Fig. 13 the determining part 123 determines the angle of rotation θcorr using the following formula (1). θcorr=G×θ2

[0087] The term gain G encompasses various types of gain (PID) used in control systems where the second angle θ2 is understood as the deviation between a setpoint and a controlled variable. Examples of gain G include proportional gain (P gain), integral gain (I gain), and differential gain (D gain).

[0088] As described above, in the second processing step, the direction of travel of the oncoming vehicle 2 is determined based on the arrangement of the outline data 4 relative to the base edge N1 of the rectangular frame 7. This allows the direction of travel of the oncoming vehicle 2 to be determined with high accuracy, even if it deviates slightly from the actual direction of travel. Consequently, a deviation of the determined direction of travel from the actual direction of travel is prevented, thus enabling the oncoming vehicle 2 to be tracked appropriately. <Wirkungen der Fahrzeugverfolgungseinrichtung>

[0089] The effects of the vehicle tracking device 10 according to the embodiment of the present invention are explained.

[0090] In the vehicle tracking device 10 according to the present embodiment, during the first processing step, the estimation part 121 extrapolates the rectangular frame 7, which approximates the outer shape of the target vehicle (e.g., the oncoming vehicle 2), based on previous data. Furthermore, during the first processing step, the adaptation part 122 adjusts outline data 4 based on the position of the rectangular frame 7. This outline data consists of a plurality of detection point data 3, which indicate the outline of the target vehicle as detected by a sensor (e.g., by the LIDAR 20) mounted in the vehicle 1 that captures the relative position of the object to be detected relative to the vehicle 1 as detection point data 3. Finally, during the first processing step, the determination part 123 determines the direction of travel of the target vehicle based on the adapted outline data 4.This allows the outline data 4 to be optimized as data used for the target vehicle's direction of travel, and the target vehicle's direction of travel can be determined with high accuracy. Consequently, a deviation of the determined direction of travel from the actual direction of travel is prevented, so that the target vehicle can be tracked appropriately.

[0091] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that during the first processing, the outline data 4 comprise a data set 8 which includes a plurality of detection point data 3 with corresponding vertical positions, and that the adaptation part 122 deletes from the outline data 4 those detection point data 3 of the data set 8 which lie on a side far removed from the target vehicle 1 in the longitudinal direction y of the rectangular frame 7. This allows data contained in a data set 8 from the side surfaces of the target vehicle to be deleted from the outline data 4. It follows that the accuracy in determining the direction of travel of the target vehicle can be improved accordingly.

[0092] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that during the first processing, the adaptation part 122 selects those detection point data 3 of the data network 8 which are further away from the vehicle 1 in the longitudinal direction y of the rectangular frame 7 than the first basic position (e.g., the first basic position y1 in Fig. 7), deletes from the outline data 4, and that the first basic position in the vehicle longitudinal direction y of the rectangular frame 7 is further away from the own vehicle 1 than the average position (e.g. the average position y2 in Fig. 7) the data point data 3 of the data network 8. This allows data contained in a data network 8 from the side surfaces of the target vehicle to be deleted from the outline data 4 in a more suitable manner. It follows that the accuracy in determining the direction of travel of the target vehicle can be improved in a more suitable manner.

[0093] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that during the first processing, the outline data 4 comprise a plurality of data networks 8, which include a plurality of detection point data 3 with corresponding vertical positions, and that the adaptation part 122 adapts the outline data 4 such that the plurality of data networks 8 approximate each other in the longitudinal direction y of the rectangular frame 7. This allows detection point data 3 from a plurality of data networks 8 to be used to determine the direction of travel of the oncoming vehicle 2. Consequently, even if the number of detection point data 3 with corresponding vertical positions (i.e.,Since the number of data points 3 in the respective data networks 8) is relatively small, the accuracy in determining the direction of travel of the target vehicle can be improved in a suitable manner.

[0094] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that during the first processing, the adaptation part 122 adjusts the outline data 4 such that the average positions of the individual detection point data 3 of the plurality of data networks 8 coincide in the longitudinal direction y of the vehicle within the rectangular frame 7. This allows detection point data 3 of a plurality of data networks 8 to be used even more effectively for determining the direction of travel of the oncoming vehicle 2. Consequently, even if the number of detection point data 3 with corresponding vertical positions among the detected outline data 4 (i.e., the number of detection point data 3 in the respective data networks 8) is relatively small, the accuracy in determining the direction of travel of the target vehicle can be improved in a more suitable manner.

[0095] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that during the first processing, the adaptation part 122 deletes from the majority of data groups 8 in the outline data 4 that is furthest from the vehicle 1 in the longitudinal direction y of the rectangular frame 7. This allows the data group 8 comprising detection point data 3 with excessively high vertical positions to be deleted from the outline data 4. Consequently, the accuracy in determining the direction of travel of the target vehicle can be improved accordingly.

[0096] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that during the first processing, the adaptation part 122 selects from the majority of data networks 8 that data set 8 for which the average position of the detection point data 3 in the longitudinal direction y of the rectangular frame 7 is further away from the vehicle 1 than the second basic position (e.g., the second basic position y3 in Fig. 8), deletes from the outline data 4, and that the second basic position in the vehicle longitudinal direction y of the rectangular frame 7 is further away from the own vehicle 1 than the average position (e.g. the average position y4 in Fig.8) of the data point data 3 of that data set 8 for which the average position of the data point data 3 is closest to the vehicle 1. This allows that data set 8 which includes data point data 3 with excessively high vertical positions to be deleted from the outline data 4 in a more suitable manner. It follows that the accuracy in determining the direction of travel of the target vehicle can be improved in a more suitable manner.

[0097] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that during the first processing, the adaptation part 122 adapts the outline data 4 in a state in which the coordinate system of the outline data 4 has been changed from the x'-y' coordinate system of the sensor (e.g., the LIDAR 20) to the xy coordinate system with one coordinate axis in the transverse direction x of the rectangular frame 7 and one coordinate axis in the longitudinal direction y of the rectangular frame 7. This reduces the computational load in the processing operations based on the position of the rectangular frame 7 for adapting the outline data 4 (e.g., the first adaptation processing, the second adaptation processing, and the third adaptation processing).

[0098] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that during the first processing, the determining element 123 determines the normal direction of the approximation line segment 5, which represents a linear approximation of a plurality of detection point data 3 of the adapted outline data 4, as the direction of travel. This allows the direction of travel of the target vehicle to be determined in a suitable manner using the outline data 4 optimized for the direction of travel of the target vehicle.

[0099] In the vehicle tracking device 10 according to the present embodiment, during the second processing step, the estimation part 121 extrapolates the rectangular frame 7, which approximates the outer shape of the target vehicle (e.g., the oncoming vehicle 2), based on previous data. Furthermore, during the second processing step, the determination part 123 determines the direction of travel of the target vehicle based on the arrangement of the outline data 4, which consists of a plurality of detection point data 3 that are measured by a sensor (e.g., a radar).The outline of the target vehicle detected by the LIDAR 20 is displayed, the sensor being mounted in the vehicle 1 and recording the relative position of the object to be detected relative to the vehicle 1 as detection point data 3, relative to the base edge N1, which is the edge of the rectangular frame 7 running in the transverse direction x of the vehicle and facing the vehicle 1. This allows the target vehicle's direction of travel to be determined with high accuracy, even if the determined direction of travel of the target vehicle deviates slightly from the actual direction of travel. It follows that a deviation of the determined direction of travel from the actual direction of travel is prevented, so that the target vehicle can be tracked appropriately.

[0100] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that during the second processing, the determining element 123 determines as the direction of travel the direction resulting from the rotational correction of the normal direction of the approach line segment 5, which represents a linear approximation of the majority of detection point data 3 of the outline data 4, based on the arrangement of the outline data 4 relative to the base edge N1. This allows the determination of the target vehicle's direction of travel based on the arrangement of the outline data 4 relative to the base edge N1 to be implemented in a suitable manner.

[0101] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that during the second processing, the determining part 123 corrects the normal direction of the approach line segment 5 by rotation based on the result of the comparison of the position ratio of the first line end E51, which is one end of the approach line segment 5 in the vehicle transverse direction x of the rectangular frame 7, to the first rectangular frame end E71, which is one end of the base edge N1 in the vehicle transverse direction x of the rectangular frame 7, and the position ratio of the second line end E52, which is the other end of the approach line segment 5 in the vehicle transverse direction x of the rectangular frame 7, to the second rectangular frame end E72, which is the other end of the base edge N1 in the vehicle transverse direction x of the rectangular frame 7.This allows the normal direction of the approach line segment 5 to be corrected by rotation based on the result of comparing the distance between the first line end E51 and the first rectangular frame end E71, and the distance between the second line end E52 and the second rectangular frame end E72. It follows that the normal direction of the approach line segment 5 can be corrected by rotation based on the inclination with which the specified direction of travel of the target vehicle deviates from the actual direction of travel.

[0102] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that during the second processing, the determining part 123 corrects the normal direction of the approach line segment 5 by rotation based on the result of the comparison of the first angle θ1 between the direction extending from the center point C1 of the rectangular frame 7 to the first line end E51 and the direction extending from the center point C1 of the rectangular frame 7 to the first rectangular frame end E71 and the second angle θ2 between the direction extending from the center point C1 of the rectangular frame 7 to the second line end E52 and the direction extending from the center point C1 of the rectangular frame 7 to the second rectangular frame end E72.This allows the rotational correction of the normal direction of the approach line segment 5 to be appropriately implemented based on the result of comparing the distance between the first line end E51 and the first rectangular frame end E71 and the distance between the second line end E52 and the second rectangular frame end E72. It follows that the rotational correction of the normal direction of the approach line segment 5 can be appropriately implemented based on the inclination with which the specified direction of travel of the target vehicle deviates from the actual direction of travel.

[0103] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that during the second processing step, the determining element 123 corrects the normal direction of the approach line segment 5 by rotating it around the center point C1 of the rectangular frame 7 from the rectangular frame end corresponding to the larger of the angles θ1 and θ2 to the rectangular frame end corresponding to the smaller of the angles θ1 and θ2. It follows that the normal direction of the approach line segment 5 can be appropriately corrected by rotating it in a direction corresponding to the direction in which the determined direction of travel of the target vehicle deviates from the actual direction of travel.

[0104] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that, during the second processing step, the determining element 123 defines the rotation angle θcorr of the rotation correction of the normal direction of the approach line segment 5 based on the larger of the first angle θ1 and the second angle θ2. It follows that the normal direction of the approach line segment 5 can be corrected by rotating it by a correction value related to the magnitude of the deviation of the determined direction of travel of the target vehicle from the actual direction of travel.

[0105] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that, during the second processing step, the determining element 123 defines the angle obtained by multiplying the gain G by the larger of the angles θ1 and θ2 as the rotation angle θcorr of the rotation correction of the normal direction of the approach line segment 5. This allows the correction value of the rotation correction of the normal direction of the approach line segment 5 to be further optimized based on the magnitude of the deviation of the determined direction of travel of the target vehicle from the actual direction of travel.

[0106] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that the estimating part 121 extrapolates the rectangular frame 7 based on a previous direction of travel determined by the determining part 123 using previous outline data 4. This improves the accuracy of the estimation of the rectangular frame 7, which approximates the external shape of the target vehicle. Consequently, the adjustment of the outline data 4 based on the position of the rectangular frame 7 can be optimized.

[0107] Furthermore, for the vehicle tracking device 10 according to the present embodiment, it is preferred that the target vehicle is a vehicle 2 approaching the vehicle 1. Generally, the front surface of a vehicle is more curved than its rear surface. It follows that if the target vehicle is a preceding vehicle, the orientation of the rear surface, which is relatively close to a flat surface, is determined as the direction of travel, whereas if the target vehicle is an oncoming vehicle 2, the orientation of the relatively strongly curved front surface is determined as the direction of travel. Consequently, with conventional technology, it was particularly difficult in the case of an oncoming vehicle 2 as the target vehicle to extract data from the side surface from the majority of detection point data 3 within the outline data 4.It follows that, in the case of an oncoming vehicle 2 as the target vehicle, the effect of being able to determine the direction of travel of the target vehicle with high accuracy can be used particularly effectively in the first or second processing described above.

[0108] A preferred embodiment of the present invention has been explained above with reference to the drawings. However, the present invention is naturally not limited to the embodiment described above. Rather, all possible modifications or improvements in the categories disclosed in the claims fall within the technical scope of the present invention.

[0109] For example, the processing steps described in this document using flowcharts do not necessarily have to be carried out in the order shown in the flowcharts. Some processing steps can also be performed in parallel. Furthermore, additional processing steps can be applied. Some processing steps can also be omitted.

[0110] Furthermore, for example, a number of the control processing operations of the vehicle tracking device 10 described above can be implemented using software, hardware, or a combination of software and hardware. Programs that comprise a software application are, for example, stored in advance on a storage medium that is located inside or outside an information processing device. [List of reference symbols] 1 own vehicle 2 Oncoming vehicle (target vehicle) 3, 3a, 3b, 3c Data collection point data 4 Outline data 5 Approximation line segment 7 Rectangular frame 8, 8a, 8b, 8c Data network 10 Vehicle tracking device 11 Recording section 12. Pursuit section 20 LiDAR (sensor) 30 radar 121 Estimation section 122 Fitting part 123 Determination section 124 Recognition part C1 Center E51 First Line End E52 Second end of line E71 First rectangular frame end E72 Second rectangular frame end N1 Base edge y1 First basic position y2 average position y3 Second basic position y4 average position y5 average position y6 Average position θ1 First angle θ2 Second Angle θcorr rotation angle

Claims

[1] Vehicle tracking device (10) for tracking a target vehicle (2) traveling in an environment of a host vehicle (1), comprising: an estimation part (121) which, based on previous data, extrapolates a rectangular frame (7) which represents an approximation of the external shape of the target vehicle (2), an adjustment part (122) which, based on the position of the rectangular frame (7), adjusts outline data (4), which is a plurality of detection point data (3) indicating an outline of the target vehicle (2) detected by a sensor (20) mounted in the own vehicle (1) and detecting the relative position of the object to be detected with respect to the own vehicle (1) as detection point data (3), and a determination part (123) that determines the traveling direction of the target vehicle (2) based on the adjusted outline data (4). [2] The vehicle tracking device according to claim 1, wherein the outline data (4) comprises a composite data (8) comprising a plurality of detection point data (3) having mutually corresponding vertical positions, and the adjusting part (122) deletes from the outline data (4) those detection point data (3) of the composite data (8) which are located on a side far away from the host vehicle (1) in a vehicle longitudinal direction of the rectangular frame (7). [3] The vehicle tracking device according to claim 2, wherein the adjustment part (122) deletes from the outline data (4) those detection point data (3) of the data set (8) which are farther away from the host vehicle (1) in the vehicle longitudinal direction of the rectangular frame (7) than a first basic position, and the first basic position is farther away from the host vehicle (1) in the vehicle longitudinal direction of the rectangular frame (7) than an average position of the detection point data (3) of the data set (8). [4] A vehicle tracking device according to any one of claims 1 to 3, wherein the outline data (4) comprises a plurality of data links (8) including a plurality of detection point data (3) having mutually corresponding vertical positions, and the adjusting part (122) adjusts the outline data (4) such that the plurality of data links (8) approach each other in the vehicle longitudinal direction of the rectangular frame (7). [5] The vehicle tracking device according to claim 4, wherein the adjusting part (122) adjusts the outline data (4) such that the average positions of the individual detection point data (3) of the plurality of data links (8) in the vehicle longitudinal direction of the rectangular frame (7) coincide. [6] Vehicle tracking device according to claim 4 or 5, wherein the adjusting part (122) deletes from the outline data (4) that data group (8) of the plurality of data groups (8) which is located on the side remote from the own vehicle (1) in the vehicle longitudinal direction of the rectangular frame (7). [7] The vehicle tracking device according to claim 6, wherein the adjusting part (122) deletes from the outline data (4) that data set (8) of the plurality of data sets (8) for which the average position of the detection point data (3) in the vehicle longitudinal direction of the rectangular frame (7) is farther away from the host vehicle (1) than the second basic position, and the second basic position in the vehicle longitudinal direction of the rectangular frame (7) is farther away from the host vehicle (1) than an average position of the detection point data (3) of that data set (8) for which the average position of the detection point data (3) is closest to the host vehicle (1). [8] The vehicle tracking device according to any one of claims 1 to 7, wherein the adjusting part (122) adjusts the outline data (4) in a state in which a coordinate system of the outline data (4) has been changed from a coordinate system of the sensor (20) to a coordinate system having a coordinate axis in the vehicle transverse direction of the rectangular frame (7) and a coordinate axis in a vehicle longitudinal direction of the rectangular frame (7). [9] The vehicle tracking device according to any one of claims 1 to 8, wherein the estimation part (121) extrapolates the rectangular frame (7) based on the previous traveling direction determined by the determination part (123) using the previous outline data (4). [10] The vehicle tracking device according to any one of claims 1 to 9, wherein the determining part (123) determines a normal direction of an approximate line segment (5) which is a line-shaped approximation of a plurality of detection point data (3) of the adjusted outline data (4) as the traveling direction.

Images