External perception device

Abstract

Provided is an external recognition device 1 that can accurately estimate the relative relationship between a cartographic feature on a map and a target detected by a sensor. The external recognition device 1 comprises: a self-position estimation unit 3 that estimates self-positions 10P1 to 10P4, which are the positions of a vehicle 10 on a map, on the basis of external information acquired by an external sensor 2 mounted on the vehicle 10; and a target recognition unit 4 that recognizes a target around the vehicle 10 on the basis of the external information. The external recognition device 1 further comprises: a map information acquisition unit 5 that acquires map information including a map point cloud 16G, which is a collection of feature points on the map, and cartographic feature information including information on the positions and types of cartographic features; a sensor point cloud acquisition unit 6 that acquires, from the target recognition unit 4, a sensor point cloud 15G around the target recognized by the target recognition unit 4; and a point cloud matching unit 7 that estimates the position of the target on the map through matching between the sensor point cloud 15G acquired by the sensor point cloud acquisition unit 6 and the map point cloud 16G.

Description

[Technical Field]

[0001] The present invention relates to an external environment recognition device mounted on a vehicle. [Background technology]

[0002] In recent years, driving assistance systems and automatic driving systems for vehicles have been developed. In these driving assistance systems and automatic driving systems, it is important to estimate the position of a target.

[0003] Patent Document 1 describes a technique in which a target is detected from two points using a camera or a sensor mounted on the vehicle, two error distributions are obtained, and standard errors E1 and E2 are compared.

[0004] The technology described in Patent Document 1 is a technology that determines whether to select a sampling point from one of two error distributions or a sampling point in the overlapping region of the two error distributions based on the magnitude relationship between the standard errors E1 and E2, and estimates the target using the selected sampling point. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 2018-185156 Summary of the Invention [Problem to be solved by the invention]

[0006] However, Patent Document 1 does not take into consideration the case where there is an error between the sensor point cloud detected using a camera or sensor and the map point cloud. If there is an error between the sensor point cloud and the map point cloud, it becomes difficult to accurately estimate the relative relationship between the map features and the targets detected by the sensor.

[0007] If it is difficult to accurately estimate the relative relationship between map features and targets detected by sensors, it is difficult to provide highly accurate driving assistance or autonomous driving of vehicles.

[0008] An object of the present invention is to provide an external environment recognition device and an external environment recognition method that can accurately estimate the relative relationship between map features and targets detected by a sensor. [Means for solving the problem]

[0009] In order to achieve the above object, the present invention is configured as follows.

[0010] an external environment recognition device including a self-position estimation unit that estimates a self-position, which is the position of the host vehicle on a map stored in a map database, based on external environment information acquired by an external environment sensor mounted on the host vehicle; a target recognition unit that recognizes targets around the host vehicle based on the external environment information; a map information acquisition unit that acquires map information including a map point cloud, which is a collection of feature points on the map, and feature information including information on the positions and types of features; a sensor point cloud acquisition unit that acquires from the target recognition unit a sensor point cloud around the target recognized by the target recognition unit; and a point cloud matching unit that estimates the position of the target on the map by matching the sensor point cloud acquired by the sensor point cloud acquisition unit with the map point cloud; a target selection unit that selects a target that satisfies a predetermined condition from the targets recognized by the target recognition unit, based on the self-position estimated by the self-position estimation unit, the recognition result of the target recognized by the target recognition unit, and the feature information; Equipped with The sensor point cloud acquisition unit acquires a sensor point cloud around the selected target from a target recognition unit. . and a point cloud matching unit that estimates the position of the target on the map by matching the sensor point cloud acquired by the sensor point cloud acquisition unit with the map point cloud. The point cloud matching unit determines whether point cloud matching is successful, and if point cloud matching is successful, outputs the position of the target on the map estimated by matching the sensor point cloud with the map point cloud. and a point cloud matching unit that estimates the position of the target on the map by matching the sensor point cloud acquired by the sensor point cloud acquisition unit with the map point cloud. The point cloud matching unit determines whether point cloud matching is successful, and if point cloud matching is unsuccessful, outputs the position of the target on the map calculated using the position of the target and the recognition result of the target recognized by the target recognition unit.

[0011] In addition, in the external environment recognition method, a map stored in a map database is recognized based on external environment information acquired by an external environment sensor mounted on the vehicle. Before Estimating a self-position, which is a position of the vehicle, and recognizing targets around the vehicle based on the external environment information; The aforementioned Map information including a map point cloud, which is a set of feature points on a map, and feature information including information on the location and type of features is acquired, and the recognized The aforementioned The sensor point cloud around the target is acquired. The aforementioned Sensor point cloud and The aforementioned By matching with map point clouds The aforementioned landmark The aforementionedEstimate location on map Then, based on the estimated self-position, the recognition results of the recognized targets, and the feature information, the system selects, from the recognized targets, those that satisfy predetermined conditions, and acquires a sensor point cloud around the selected targets. [Effects of the Invention]

[0012] According to the present invention, it is possible to provide an external environment recognition device and an external environment recognition method that can accurately estimate the relative relationship between map features and targets detected by a sensor.

[0013] In the present invention, a point cloud surrounding a sensor target (a target detected by a sensor) is extracted, point cloud matching is performed between the map point cloud and the sensor target, the position of the sensor target on the map is estimated, and the relative relationship between the map feature and the sensor target is accurately estimated. [Brief explanation of the drawings]

[0014] [Figure 1] 1 is a schematic configuration diagram of an external environment recognition device according to a first embodiment of the present invention. [Figure 2] FIG. 10 is an explanatory diagram of a scale error. [Figure 3] FIG. 10 is an explanatory diagram of a scale error. [Figure 4] FIG. 10 is an explanatory diagram of an angle error. [Figure 5] FIG. 10 is an explanatory diagram of an angle error. [Figure 6] FIG. 2 is a diagram illustrating the operation of point cloud matching in the first embodiment. [Figure 7] FIG. 10 is a diagram showing an example of an information table in which combinations of sensor targets and map features are set. [Figure 8] FIG. 10 is an explanatory diagram of the selection of a sensor target. [Figure 9] FIG. 10 is a diagram showing an information table showing the relationship between sensor targets, map features, and thresholds. [Figure 10] FIG. 1 is a schematic configuration diagram of an external environment recognition device according to a second embodiment of the present invention. [Figure 11] FIG. 10 is a diagram illustrating the operation of a speed prediction unit. [Figure 12] FIG. 10 is a schematic configuration diagram of an external environment recognition device according to a third embodiment of the present invention. [Figure 13] FIG. 4 is a diagram illustrating the operation of an intended speed prediction unit. [Figure 14] FIG. 10 is a schematic configuration diagram of an external environment recognition device according to a fourth embodiment of the present invention. [Figure 15] 10 is an explanatory diagram showing that there is a high-precision range in the external environment recognition results obtained by the target recognition unit. FIG. [Figure 16] FIG. 10 is an explanatory diagram of self-generated map according to the fourth embodiment. [Figure 17] FIG. 10 is a schematic configuration diagram of an external environment recognition device according to a fifth embodiment of the present invention. [Figure 18] FIG. 10 is a diagram illustrating the operation of a trajectory prediction unit according to the fifth embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0015] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will now be described in detail with reference to the accompanying drawings. [Example]

[0016] Example 1 FIG. 1 is a schematic configuration diagram of an external environment recognition device 1 according to a first embodiment of the present invention.

[0017] In FIG. 1, the external environment recognition device 1 includes a self-position estimation unit 3, a target recognition unit 4, a map information acquisition unit 5, a sensor point cloud acquisition unit 6, a point cloud matching unit 7, and a target selection unit 8.

[0018] The self-position estimation unit 3 estimates the self-position, which is the position of the vehicle 10 on a map, based on external information acquired by an external sensor 2 mounted on the vehicle 10 (shown in FIG. 2).

[0019] The target recognition unit 4 recognizes targets around the vehicle 10 based on external information detected by the external sensor 2. The external sensor 2 is a sensor such as a camera or radar, and detects external information about the vehicle 10.

[0020] The map information acquisition unit 5 acquires map information including a map point cloud, which is a collection of feature points on a map stored in a storage unit mounted on the vehicle 10 or transmitted from an external device, and feature information, which includes information on the positions and types of features. In Fig. 1, the map information is stored in a map database 9. The map database 9 may be stored in a storage unit mounted on the vehicle 10, or may be information transmitted from an external device.

[0021] The sensor point cloud acquisition unit 6 acquires a sensor point cloud, which is a plurality of positions around the target recognized by the target recognition unit 4, from the external world information detected by the external sensor 2.

[0022] The point cloud matching unit 7 estimates the position on the map of the target recognized by the target recognition unit 4 by matching the sensor point cloud acquired by the sensor point cloud acquisition unit 6 with the map point cloud acquired by the map information acquisition unit 5. In other words, the point cloud matching unit 7 determines whether the point cloud matching is successful, and if the point cloud matching is successful, outputs the position on the map of the target estimated by matching the sensor point cloud with the map point cloud. If the point cloud matching is unsuccessful, the point cloud matching unit 7 outputs the position on the map of the target calculated using the self-position and the recognition result of the target recognized by the target recognition unit 4.

[0023] The target selection unit 8 selects targets that satisfy predetermined conditions from among the targets recognized by the target recognition unit 4, based on the self-position of the vehicle 10 estimated by the self-position estimation unit 3, the recognition results of the targets recognized by the target recognition unit 4, and the feature information acquired by the map information acquisition unit 5.

[0024] The following describes the positional relationship between the vehicle 10 and the other vehicle 11, and the positional relationship between the vehicle 10 and the road and the other vehicle 11. The positional relationship between the vehicle 10 and the external target detected using a self-generated map based on odometry (a method of estimating the current position from the rotation angle of the vehicle's wheels) and the external sensor 2 may be inaccurate compared to the actual positional relationship (true positional relationship) between the vehicle 10 and the external target.

[0025] The other vehicle 11 is also included in the targets.

[0026] The scale error will be described with reference to FIGS.

[0027] 2(a) is a diagram showing a true position map 12T1, which is the actual positional relationship (true positional relationship) of the host vehicle 10 with respect to external objects. In FIG. 2(a), another vehicle 11 is traveling ahead of the host vehicle 10 in the opposite direction to the traveling direction of the host vehicle 10. The host vehicle 10 and the other vehicle 11 are traveling within a road boundary 12, and the other vehicle 11 is traveling at a position midway across the turn-around area position 17.

[0028] FIG. 2(b) is a diagram showing a self-generated map 12G1 generated by the vehicle 10 using odometry.

[0029] When the true position map 12T1 shown in Fig. 2(a) is compared with the self-generated map 12G1 shown in Fig. 2(b), the position of the escape area position 17 is shifted. This is because there is a scale error due to an odometry error.

[0030] FIG. 3(a) is a diagram showing a sensing result 12S1 indicating the positional relationship between the host vehicle 10 and another vehicle 11 detected using the external sensor 2 of the host vehicle 10. FIG.

[0031] FIG. 3(b) is a diagram showing a composite result map 12GS1 obtained by combining the sensing result 12S1 shown in FIG. 3(a) and the self-generated map 12G1 shown in FIG. 2(b).

[0032] When comparing the true position map 12T1 shown in (a) of Figure 2 with the synthesis result map 12G1 in (b) of Figure 3, the position of the evacuation area position 17 is shifted, and the relative relationship between the other vehicle 11 and the evacuation area 17 is inaccurate.

[0033] In other words, in the true position map 12T1 shown in (a) of Figure 2, the other vehicle 11 is in a position crossing the evacuation area position 17, whereas in the synthesized result map 12GS1 shown in (b) of Figure 3, the position 11P1 of the other vehicle is in a state before crossing the evacuation area position 17.

[0034] If the relative relationship between the other vehicle 11 and the avoidance area 17 is inaccurate due to scale error, it is difficult to perform driving assistance or automatic driving of the host vehicle 10 with high accuracy.

[0035] The angle error will be described with reference to FIGS.

[0036] 4(a) is a diagram showing a true position map 12T2 that shows the actual positional relationship (true positional relationship) between the host vehicle 10 and external targets. In FIG. 4(a), another vehicle 11 is traveling in a direction that is approximately perpendicular to the traveling direction of the host vehicle 10 on a road that is approximately perpendicular to the road on which the host vehicle 10 is traveling.

[0037] FIG. 4(b) is a diagram showing a self-generated map 12G2 generated by the vehicle 10 using odometry.

[0038] When comparing the true position map 12T2 shown in Fig. 4(a) with the self-generated map 12G2 shown in Fig. 4(b), the angle of the dividing line 14 is shifted. This is because there is an angle error due to an odometry error.

[0039] FIG. 5(a) is a diagram showing a sensing result 13R2 indicating the positional relationship between the host vehicle 10 and another vehicle 11 detected using the external sensor 2 of the host vehicle 10. FIG.

[0040] FIG. 5(b) is a diagram showing a composite result map 12GS2 obtained by combining the sensing result 13R2 shown in FIG. 5(a) and the self-generated map 12G2 shown in FIG. 4(b).

[0041] When the true position map 12T2 shown in FIG. 4(a) is compared with the synthesis result map 12GS2 shown in FIG. 5(b), the relative relationship between the other vehicle position 11P2 on the map and the lane marking 14 is inaccurate.

[0042] In other words, in the true position map 12T2 shown in (a) of Figure 4, the other vehicle 11 is positioned to travel in a direction parallel to the dividing line 14, whereas in the synthesized result map 12GS2 shown in (b) of Figure 5, the position 11P2 of the other vehicle is positioned to travel in a direction that intersects with the dividing line 14.

[0043] If the relative relationship between the other vehicle 11 and the lane marking 14 is inaccurate due to an angle error, it will be difficult to perform driving assistance or automatic driving of the vehicle 10 with high precision, just as with a scale error.

[0044] In the first embodiment of the present invention, a point cloud around a target detected by an external sensor 2 is extracted, and point cloud matching is performed between the extracted point cloud around the target and a map point cloud obtained from map information, and the position on the map of the target detected by the external sensor 2 is estimated, thereby correcting scale error and angle error, and accurately estimating the relative relationship between the host vehicle 10, other vehicles 11, and the features on the map.

[0045] FIG. 6 is a diagram illustrating the operation of point cloud matching in the first embodiment of the present invention.

[0046] 6(a), a plurality of points 15 (shown as rectangles in FIG. 6) in a sensor point cloud 15G around another vehicle 11 are extracted from the external environment information detected by the external sensor 2. In the area of ​​the sensor point cloud 15G around another vehicle 11, the sensor points 15 have relatively high accuracy.

[0047] The process of extracting the multiple points 15 in the sensor point cloud 15G is executed by the external environment recognition device 1. That is, the target recognition unit 4 recognizes targets from external environment information detected by the external environment sensor 2. Furthermore, the self-position estimation unit 3 estimates the map information acquired by the map information acquisition unit 5 and the self-position, which is the position of the host vehicle 10 detected by the external environment sensor 2. Then, the target selection unit 8 selects targets from the map information acquired by the map information acquisition unit 5, the self-position estimated by the self-position estimation unit 3, and the targets recognized by the target recognition unit 4. Then, the sensor point cloud acquisition unit 6 extracts the sensor point cloud 15G based on the targets recognized by the target recognition unit 4 and the targets selected by the target selection unit 8.

[0048] Next, the point cloud matching unit 7 performs a matching process between the sensor points 15 and the map point cloud 16 based on the map information acquired by the map information acquisition unit 5 and the sensor point cloud 15G extracted by the sensor point cloud acquisition unit 6.

[0049] Fig. 6(b) is an explanatory diagram of the matching process between the sensor points 15 and the map points 16. The circles in Fig. 6(b) indicate the map points 16. The map point group 16G made up of the map points 16 has relatively high accuracy in a local area.

[0050] By performing matching processing so that the sensor point 15 and the map point 16 overlap, the relative relationship between the other vehicle position 11P3 on the map and the escape area 17 is made more accurate.

[0051] For the angle error shown in FIG. 5, the relative relationship between the other vehicle position 11P2 on the map and the lane marking 14 is made more accurate by performing the above-described matching process.

[0052] The success / failure determination of the matching process in the point cloud matching unit 7 will be described.

[0053] If the number of points included in the map is equal to or less than a predetermined threshold, the matching is determined to be unsuccessful, and the matching process is not executed from the beginning.

[0054] Next, if the number of points included in the map exceeds a predetermined threshold, and the average value of the distance between the corresponding sensor points 15 and map points 16 is equal to or greater than the predetermined threshold, it is determined to be a failure. Also, if the number of corresponding points between the sensor points 15 and map points 16 is equal to or less than the predetermined threshold, it is determined to be a failure.

[0055] If the matching process is successful, the point cloud matching unit 7 outputs the position and attitude of the sensor target on the map estimated by the matching process. The host vehicle 10 performs driving assistance or automatic driving processing based on the position and attitude of the sensor target output from the point cloud matching unit 7.

[0056] If the matching process fails, the point cloud matching unit 7 can output the position and orientation of the sensor target on the map estimated from the self-position estimation result and the sensing result (the provisional position of the target selection unit 8).

[0057] Next, the operation of the target selection unit 8 will be described.

[0058] Combinations of sensor targets and map features whose relative relationships are important are set in advance. Fig. 7 is a diagram showing an example of an information table 19 in which combinations of sensor targets and map features are set.

[0059] FIG. 8 is an explanatory diagram of the selection of the sensor target.

[0060] The target selection unit 8 calculates a tentative position 11P4 of the sensor feature from the self-location estimation result 10P3 and the sensing result 13R1. Then, the target selection unit 8 selects a sensor target whose distance to the corresponding map feature (the escape area in the example shown in FIG. 8) is equal to or less than a predetermined threshold.

[0061] As shown in Figure 7, by setting in advance combinations of sensor targets and map features whose relative relationships are important, and by setting in advance combinations of sensor targets and map features whose relative relationships are important in terms of the distance between the features and the sensor targets, the processing time for target selection can be shortened.

[0062] The target selection unit 8 may be configured to change the target selection threshold (the distance between a ground object and a sensor target) according to the size, speed, weather, and brightness of the sensor target, thereby enabling appropriate target selection.

[0063] For example, the larger the size of the sensor target, the larger the threshold value, or the faster the speed of the sensor target (the arrival time of the sensor target to the map feature can be used instead of the distance).

[0064] Also, the worse the weather, the higher the threshold value can be set, and the darker the brightness, the higher the threshold value can be set.

[0065] 9 is a diagram showing an example of an information table 19A that shows the relationship between sensor targets, map features, and thresholds by adding and storing thresholds for the sensor targets and map features in the information table 19. The predetermined thresholds am, bm, cm, dm, etc. for each can be changed according to at least one of the size, speed, weather, and brightness of the sensor target.

[0066] An external environment recognition method according to a first embodiment of the present invention will be described.

[0067] Based on external environment information acquired by an external sensor 2 mounted on the vehicle 10, the vehicle's own position, which is the position of the vehicle 10 on a map stored in a map database 9, is estimated, targets around the vehicle 10 are recognized based on the external environment information, map information including a map point cloud, which is a collection of feature points on the map, and feature information including information on the position and type of features, is acquired, a sensor point cloud around the recognized target is acquired, and the position of the target on the map is estimated by matching the acquired sensor point cloud with the map point cloud.

[0068] As described above, according to the first embodiment of the present invention, a configuration is made in which a peripheral point cloud of a sensor target is extracted, point cloud matching is performed between the map point cloud and the sensor target, and the position of the sensor target on the map is estimated. Therefore, it is possible to provide an external environment recognition device and an external environment recognition method that can accurately estimate the relative relationship between map features and targets detected by a sensor.

[0069] Example 2 Next, a second embodiment of the present invention will be described.

[0070] Fig. 10 is a schematic configuration diagram of an external environment recognition device 1 according to a second embodiment of the present invention. The second embodiment differs from the first embodiment shown in Fig. 1 in that a speed prediction unit 20 is connected to the point cloud matching unit 7 of the first embodiment. Other configurations of the second embodiment are the same as those of the first embodiment.

[0071] The speed prediction unit 20 predicts the position and speed of another vehicle 11 ahead on the map estimated by the point cloud matching unit 7, and is used for the adaptive cruise control (ACC) function to appropriately control the speed of the host vehicle 10.

[0072] FIG. 11 is a diagram illustrating the operation of the speed prediction unit 20.

[0073] 11(a), the host vehicle 10P3 automatically travels at a constant speed while maintaining a constant distance from another vehicle 11P4 traveling ahead by using an adaptive cruise control function. When the other vehicle 11P4 traveling ahead is approaching a stop line 21, the speed prediction unit 20 can predict that the other vehicle 11P4 will reduce its speed.

[0074] If the speed of the host vehicle 10P3 is controlled based on the speed prediction by the speed prediction unit 20, the adaptive cruise control function can be executed with high accuracy.

[0075] As described above, according to the second embodiment of the present invention, it is possible to obtain the same effects as those of the first embodiment, and also to execute the adaptive cruise control function with high accuracy.

[0076] Example 3 Next, a third embodiment of the present invention will be described.

[0077] 12 is a schematic configuration diagram of an external environment recognition device 1 according to a third embodiment of the present invention. The third embodiment differs from the first embodiment shown in FIG. 1 in that an intention prediction unit 24 is connected to the point cloud matching unit 7 of the first embodiment. Other configurations of the third embodiment are the same as those of the first embodiment.

[0078] The intention prediction unit 24 is used in an automatic driving device and predicts the intention of a pedestrian based on the relative relationship between the pedestrian's position on the map estimated by the point cloud matching unit 7 and the crosswalk on the map.

[0079] This allows the intention of the pedestrian to be predicted with high accuracy, and the behavior of the vehicle 10 to be planned appropriately.

[0080] FIG. 13 is a diagram illustrating the operation of the intended speed prediction unit 20. As shown in FIG.

[0081] 13(a), another vehicle 11P5 ahead of the host vehicle 10P4 is located on the crosswalk 23, and a pedestrian 22 on the map estimated by point cloud matching is located just before the crosswalk 23. In this case, the intention prediction unit 24 predicts that the pedestrian 22 intends to cross the crosswalk 23. Based on the prediction by the intention prediction unit 24, an automatic driving device (not shown) can control the host vehicle 10P4 to stop just before the crosswalk.

[0082] 13(b), another vehicle 11P5 ahead of the host vehicle 10P4 is located on the crosswalk 23, and the pedestrian 22 on the map estimated by point cloud matching is not in front of the crosswalk 23 but is located away from the crosswalk 23 and approximately halfway between the host vehicle 10P4 and the crosswalk 23. In this case, the intention prediction unit 24 predicts that the pedestrian 22 is standing still, and based on the prediction of the intention prediction unit 24, the automatic driving device (not shown) can control the host vehicle 10P4 to travel without stopping in front of the crosswalk.

[0083] Similarly, the intention prediction unit 24 predicts the intention of a stopped vehicle, which is a sensor target, or a time-limited parking section, which is a map feature, and the automatic driving device can control the vehicle based on the result. For example, the intention prediction unit 24 predicts whether the stopped vehicle is a parked vehicle or a vehicle that is temporarily stopped, such as waiting at a traffic light.

[0084] When the other vehicle 11P5 is located on the crosswalk 23, the external sensor 2 may not be able to detect the crosswalk 23. In this case, it is necessary to obtain information from the map database 9 and determine whether or not the other vehicle 11P5 is the crosswalk 23.

[0085] The intention prediction unit 24 can accurately predict the intention of the pedestrian 22 based on the relative relationship between the position of the pedestrian 22 on the map estimated by the point cloud matching unit 7 and the crosswalk 23 on the map.

[0086] As described above, according to the third embodiment of the present invention, it is possible to obtain the same effects as those of the first embodiment, and also to predict the intentions of pedestrians and the like and appropriately control the vehicle 10 through automatic driving.

[0087] Example 4 Next, a fourth embodiment of the present invention will be described.

[0088] Fig. 14 is a schematic configuration diagram of an external environment recognition device 1 according to a fourth embodiment of the present invention. The fourth embodiment differs from the first embodiment shown in Fig. 1 in that an own map generation unit 26 is connected to the target object recognition unit 4 of the first embodiment, and the own map generation unit 26 generates an own map from information from the odometry unit 25 and target object information from the target object recognition unit 4, and stores the map in the map database 9. Other configurations of the fourth embodiment are the same as those of the first embodiment.

[0089] As shown in Figure 15, in Example 4, if a high-precision range 27 exists among the external environment recognition results by the target recognition unit 4 at each time, a self-map is generated for this range 27 by combining the map by the odometry unit 25 and the information by the target recognition unit 4.

[0090] As shown in FIG. 16(a), the external environment recognition result 29 of the host vehicle 10P one time ago is not within a high accuracy range, so the host map generating unit 26 does not generate a map.

[0091] Then, since the escape area 17A of the host vehicle 10N at a later time is a high-precision range, the host map generation unit 26 generates a host map from the odometry estimated position 28 and the target information from the target recognition unit 4 and stores it in the map database 9.

[0092] The state shown in Figure 16(b) is the state after further time has passed from the state shown in Figure 16(a). Since the escape area 17 in the vicinity of the host vehicle 10N1 in the state shown in Figure 16(b) is a high-precision range, the host map generation unit 26 generates a host map from the odometry estimated position 28 and the target object information from the target object recognition unit 4, and stores the generated map in the map database 9. The host map generation unit 26 saves the sensor point cloud 15G as a map for the periphery of the feature described in the information table 19 in which the type of target object is associated with the type of feature.

[0093] In this way, a relatively high-precision map is generated and stored in the map database 9.

[0094] In the above example, a relatively high-precision map is generated and stored in the map database 9, but it is also possible to configure the system to store point clouds only around map features whose relative positional relationships with sensor targets are important. For example, it is also possible to store point clouds only around the evacuation area.

[0095] In this way, the capacity of the maps stored in the map database 9 can be reduced.

[0096] As described above, according to the fourth embodiment of the present invention, it is possible to obtain the same effects as those of the first embodiment, and in addition, it is possible to generate a map with relatively high accuracy, and further to reduce the map capacity to be stored in the map database 9.

[0097] Example 5 Next, a fifth embodiment of the present invention will be described.

[0098] 17 is a schematic configuration diagram of an external environment recognition device 1 according to a fifth embodiment of the present invention. The fifth embodiment differs from the first embodiment shown in FIG. 1 in that a trajectory prediction unit 30 is connected to the point cloud matching unit 7 of the first embodiment. Other configurations of the fifth embodiment are the same as those of the first embodiment.

[0099] As shown in (a) and (b) of Figure 18, the trajectory prediction unit 30 predicts the trajectory of the oncoming vehicle 11P6 based on the relative relationship between the oncoming vehicle position 11P6 on the map estimated by the point cloud matching unit 7 and the escape area 17 on the map.

[0100] In the state shown in (a) of Figure 18, the trajectory prediction unit 30 predicts that there is a possibility that the oncoming vehicle will enter the turn-off area 17. In the state shown in (b) of Figure 18, the trajectory prediction unit 30 predicts that the oncoming vehicle will not enter the turn-off area 17.

[0101] The trajectory prediction unit 30 can accurately predict the trajectory of an oncoming vehicle, so that the behavior of the vehicle can be appropriately planned.

[0102] As described above, according to the fifth embodiment of the present invention, it is possible to obtain the same effects as those of the first embodiment, and in addition, it is possible to accurately predict the trajectory of an oncoming vehicle, thereby making it possible to appropriately plan the behavior of the vehicle.

[0103] In the above-described Examples 1 to 5, the target selection unit 8 is a component of the external environment recognition device 1, but it is also possible to omit the item selection unit 8, and examples in which the item selection unit 8 is omitted are also included in the embodiments of the present invention.

[0104] In an example in which the item selection unit 8 is omitted, the self-position estimated by the self-position estimation unit 3 is output to the point cloud matching unit 7, and the target information recognized by the target recognition unit 4 is output only to the sensor point cloud acquisition unit 6. Then, the point cloud matching unit 7 matches the point clouds based on the self-position estimated by the self-position estimation unit 3, the sensor point cloud acquired by the sensor point cloud acquisition unit 6, and the map information from the map information acquisition unit 5. [Explanation of symbols]

[0105] 1 Appearance recognition device, 2 External sensor, 3 Self-position estimation unit, 4 Target recognition unit, 5 Map information acquisition unit, 6 Sensor point cloud acquisition unit, 7 Point cloud matching unit, 8 Target selection unit, 9 Map database, 10 Host vehicle, 10P1, 10P2, 10P3, 10P4, 10P5 Host vehicle position on map, 11 Other vehicles, 11P1, 11P2, 11P3, 11P4, 11P5, 11P6 Other vehicle positions on map, 11P4 Provisional other vehicle positions, 12 Road boundaries, 12G1, 12G2 Self-generated map, 12GS1, 12GS2 Map sensing synthesis result state, 12S1...sensing result state, 12T1, 12T2...true state, 13R1, 13R2...relative position of vehicle and other vehicles, 14...margin, 15...extracted point, 15G...extracted point cloud, 16...map point, 17, 17A...evacuation area position, 18TH...threshold distance, 19, 19A...information table, 20...speed prediction unit, 21...stop line, 22...pedestrian position, 23...pedestrian crossing, 24...intention prediction unit, 25...odometry unit, 26...self-map generation unit, 27...good accuracy range, 28...odometry estimated position, 29...external world recognition result from one time point ago, 30...trajectory prediction unit

Claims

1. a self-position estimation unit that estimates a self-position of the vehicle on a map stored in a map database based on external environment information acquired by an external environment sensor mounted on the vehicle; a target recognition unit that recognizes targets around the host vehicle based on the external environment information; a map information acquisition unit that acquires map information including a map point cloud, which is a collection of feature points on the map, and feature information including information on the positions and types of features; a sensor point cloud acquisition unit that acquires, from the target recognition unit, a sensor point cloud around the target recognized by the target recognition unit; a point cloud matching unit that estimates a position of the target on the map by matching the sensor point cloud acquired by the sensor point cloud acquisition unit with the map point cloud; a target selection unit that selects a target that satisfies a predetermined condition from the targets recognized by the target recognition unit, based on the self-position estimated by the self-position estimation unit, the recognition result of the target recognized by the target recognition unit, and the feature information; wherein the sensor point cloud acquisition unit acquires a sensor point cloud around the selected target from a target recognition unit.

2. The external environment recognition device according to claim 1, the target selection unit refers to an information table in which the types of targets and the types of features are associated with each other, and selects the target when a feature corresponding to the target recognized by the target recognition unit exists in the vicinity of the target.

3. The external environment recognition device according to claim 2, the target selection unit estimates a tentative position of the target on the map using the self-location and a recognition result of the target recognized by the target recognition unit; an external environment recognition device, characterized in that the target is selected when the distance between the provisional position and the position of the feature corresponding to the target in the information table is equal to or less than a threshold.

4. The external environment recognition device according to claim 3, the information table holds threshold values ​​used by the target selection unit for each correspondence between a target type and a feature type; The external environment recognition device, wherein the target selection unit changes the threshold value stored in the information table based on at least one of the target recognition result, weather, and brightness.

5. A self-position estimation unit that estimates a self-position, which is the position of the vehicle on a map stored in a map database, based on external information acquired by an external sensor mounted on the vehicle; a target recognition unit that recognizes targets around the host vehicle based on the external environment information; a map information acquisition unit that acquires map information including a map point cloud, which is a collection of feature points on the map, and feature information including information on the positions and types of features; a sensor point cloud acquisition unit that acquires, from the target recognition unit, a sensor point cloud around the target recognized by the target recognition unit; a point cloud matching unit that estimates a position of the target on the map by matching the sensor point cloud acquired by the sensor point cloud acquisition unit with the map point cloud; Equipped with the point cloud matching unit determines whether point cloud matching is successful, and if point cloud matching is successful, outputs a position of the target on the map estimated by matching the sensor point cloud with the map point cloud.

6. A self-position estimation unit that estimates a self-position, which is the position of the vehicle on a map stored in a map database, based on external information acquired by an external sensor mounted on the vehicle; a target recognition unit that recognizes targets around the host vehicle based on the external environment information; a map information acquisition unit that acquires map information including a map point cloud, which is a collection of feature points on the map, and feature information including information on the positions and types of features; a sensor point cloud acquisition unit that acquires, from the target recognition unit, a sensor point cloud around the target recognized by the target recognition unit; a point cloud matching unit that estimates a position of the target on the map by matching the sensor point cloud acquired by the sensor point cloud acquisition unit with the map point cloud; Equipped with the point cloud matching unit determines whether point cloud matching has been successful, and if point cloud matching has failed, outputs a position of the target on the map calculated using the self-position and a recognition result of the target recognized by the target recognition unit.

7. The external environment recognition device according to claim 1, an external environment recognition device further comprising a prediction unit that predicts at least one of a trajectory, a speed, and an intention of the target based on the position of the target on the map estimated by the point cloud matching unit and the position of the feature on the map.

8. The external environment recognition device according to claim 1, an own map generation unit that generates the map using the relative position and attitude of the own vehicle estimated by odometry, the recognition result of the target recognized by the target recognition unit, and the sensor point cloud.

9. The external environment recognition device according to claim 8, The self-map generation unit stores the sensor point cloud as the map in the map database regarding the periphery of the feature described in an information table in which the type of the target and the type of the feature are associated. An external environment recognition device.

10. Estimating a vehicle's own position, which is the position of the vehicle on a map stored in a map database, based on external environment information acquired by an external environment sensor mounted on the vehicle; Recognizing targets around the vehicle based on the external environment information; acquiring map information including a map point cloud, which is a collection of feature points on the map, and feature information including information on the positions and types of features; Acquire a sensor point cloud around the recognized target; estimating the position of the target on the map by matching the acquired sensor point cloud with the map point cloud; Based on the estimated self-location, the recognition result of the recognized target, and the feature information, Selecting the targets that satisfy a predetermined condition from among the recognized targets; An external environment recognition method, comprising: acquiring a sensor point cloud around the selected target.

Images