Vehicle position estimation method and vehicle position estimation device

Abstract

This invention provides a vehicle position estimation method that can improve the accuracy of vehicle position estimation while suppressing an increase in processing load. [Solution] First, the driving environment around the vehicle is determined (S107). The vehicle is equipped with multiple types of sensors (e.g., GPS receiver, LiDAR, camera, radar). Next, a table of information showing the correspondence between the driving environment and the sensors used to estimate the vehicle's position is referenced, and the sensors associated with the determined driving environment are identified (S111). Subsequently, the vehicle's position is estimated based on the detection results of the identified sensors (S112).

Description

Technical Field

[0001] The present disclosure relates to a vehicle position estimation method and a vehicle position estimation device.

Background Art

[0002] Conventionally, a device has been proposed that mounts a GPS receiver, a camera, and a lidar on a vehicle and estimates the position of the vehicle itself based on the detection results of the GPS receiver, the camera, and the lidar (see, for example, Patent Document 1). In the device described in Patent Document 1, the position of the vehicle itself (hereinafter, also referred to as "individual position") is detected separately from each of the detection results of the GPS receiver, the camera, and the lidar, the detection accuracy of each detected individual position is calculated, and the adoption ratio of the individual position with high detection accuracy is increased to improve the estimation accuracy of the position of the vehicle itself.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the device described in Patent Document 1, for all of the GPS receiver, the camera, and the lidar, that is, for all sensors for estimating the position of the vehicle itself, the estimation of the position of the vehicle itself from the detection results and the calculation of the estimation accuracy of the position of the vehicle itself are performed, so there is a possibility that the processing load increases. An object of the present disclosure is to provide a vehicle position estimation method and a vehicle position estimation device capable of suppressing an increase in processing load while improving the estimation accuracy of the position of the vehicle itself.

Means for Solving the Problems

[0005] A vehicle position estimation method according to one aspect of the present disclosure involves determining the driving environment around a vehicle equipped with multiple types of sensors for estimating the vehicle's own position, referring to table information showing the correspondence between the driving environment and the sensors used to estimate the vehicle's own position, identifying the sensors associated with the determined driving environment, and estimating the vehicle's own position based on the detection results of the identified sensors.

[0006] Furthermore, a vehicle position estimation device according to one aspect of the present disclosure is mounted on a vehicle and comprises: a plurality of types of sensors for estimating the vehicle's position; a discrimination unit for determining the driving environment around the vehicle; an identification unit for referring to table information showing the correspondence between the driving environment and the sensors used to estimate the vehicle's position and identifying the sensors associated with the driving environment determined by the discrimination unit; and an estimation unit for estimating the vehicle's position based on the detection results of the sensors identified by the identification unit. [Effects of the Invention]

[0007] According to this disclosure, it is possible to provide a vehicle position estimation method and a vehicle position estimation device that can improve the accuracy of vehicle position estimation while suppressing an increase in processing load. [Brief explanation of the drawing]

[0008] [Figure 1] This figure shows the overall configuration of the automated driving system according to the present invention. [Figure 2] This diagram shows the contents of the table information. [Figure 3] This diagram shows the various functions that an information processing device can perform. [Figure 4] This is a flowchart showing the overall flow of the vehicle position estimation method according to the embodiment. [Figure 5] This figure shows the differences between the comparative example and the embodiment. [Figure 6] This figure shows the differences between the comparative example and the embodiment. [Modes for carrying out the invention]

[0009] The embodiments of this invention will be described in detail below with reference to the drawings. Note that the drawings are schematic and may differ from actual ones. Furthermore, the embodiments of the present invention shown below are illustrative examples of devices and methods for realizing the technical concept of the present invention, and the technical concept of the present invention is not limited to the structure, arrangement, etc., of the components described below. The technical concept of the present invention can be modified in various ways within the technical scope defined by the claims described in the patent claims.

[0010] (Embodiment) In this embodiment, as shown in Figure 1, an example is given in which the vehicle position estimation method and vehicle position estimation device of this disclosure are applied to the vehicle position estimation device 2 included in the automatic driving system 1 of vehicle C. Figure 1 is a diagram showing the overall configuration of the automatic driving system 1 according to this embodiment. As shown in Figure 1, the autonomous driving system 1 includes an external sensor 3, a vehicle sensor 4, an HMI 5 (Human Machine Interface), an actuator 6, and an information processing device 7.

[0011] The external sensor 3 is a sensor that detects various information (external environment information) about the external environment surrounding vehicle C. The external sensor 3 can include, for example, multiple types of sensors for estimating the vehicle's position. For example, it may include a GPS (Global Positioning System) receiver 8, LiDAR 9, camera 10, and radar 11. In the following description, the GPS receiver 8, LiDAR 9, camera 10, and radar 11 will be collectively referred to as "sensors 8-11". The GPS receiver 8 is a sensor that acquires its position by receiving signals from GPS satellites. The LiDAR 9 is a sensor that acquires the coordinates of each point of an object around vehicle C by irradiating laser light around vehicle C and detecting the reflected light. The camera 10 is a sensor that acquires image data of the area around vehicle C by imaging the area around vehicle C. For example, the camera 10 can be an RGB camera, an IR camera, or a depth image camera. The radar 11 is a sensor that detects the position of objects around vehicle C by emitting radar waves around vehicle C and detecting the reflected waves. A millimeter-wave radar emitting millimeter waves can be used as the radar 11. The detection results (external environment information) are output to the information processing device 7.

[0012] Vehicle sensor 4 is a sensor that detects various vehicle status information (vehicle information) obtained from vehicle C. Examples of vehicle sensors 4 include a vehicle speed sensor that detects the vehicle speed of vehicle C, a wheel speed sensor that detects the rotational speed of the vehicle C's tires, a three-axis acceleration sensor that detects the acceleration and deceleration of vehicle C in three axes, a steering angle sensor that detects the steering angle of the steering wheels, a gyro sensor that detects the angular velocity of vehicle C in three axes, a yaw rate sensor that detects the yaw rate of vehicle C, an accelerator sensor that detects the accelerator opening of vehicle C, and a brake sensor that detects the amount of brake operation of vehicle C. The detection results (vehicle information) are output to the information processing device 7.

[0013] HMI5 is an interface device that exchanges information between the information processing device 7 and the driver of vehicle C. For example, HMI5 can be a display that shows image information visible to the driver of vehicle C, a speaker that outputs audio information that the driver can hear, and controls that the driver can operate (e.g., buttons, levers, switches, dials, touch panels).

[0014] Actuator 6 comprises a steering actuator, an accelerator opening actuator, and a brake control actuator. The steering actuator controls the steering direction and amount of the steering wheel of vehicle C. The accelerator opening actuator controls the accelerator opening of vehicle C. The brake control actuator controls the brake system of vehicle C. Actuator 6 (including the steering actuator) controls the steering wheel, accelerator opening, and brake system of vehicle C in accordance with control signals output from the information processing device 7, thereby generating predetermined vehicle behavior in vehicle C.

[0015] The information processing device 7 comprises a processor 12 and peripheral components such as a storage device 16 that stores a program 13, map information 14, and table information 15. The processor 12 can be, for example, a CPU (Central Processing Unit) or an MPU. The storage device 16 can be, for example, a semiconductor storage device, a magnetic storage device, or an optical storage device. The storage device 16 may also include registers, cache memory, and memory such as ROM and RAM used as main memory. Each function of the information processing device 7 described below is realized, for example, by the processor 12 executing the program 13 stored in the storage device 16.

[0016] Furthermore, map information 14 is information that indicates, for example, the driving environment at each location on the map. Examples of driving environments include urban areas, indoor parking lots (including indoor and underground parking lots), tunnels, highways, underpasses, suburbs, and others (areas other than the urban to suburban areas mentioned above). Table information 15, as shown in Figure 2, is information that shows the correspondence between the driving environment and the sensors 8-11 (GPS receiver 8, LiDAR 9, camera 10, radar 11) used to estimate the vehicle C's position. In other words, table information 15 associates the sensors 8-11 used in each driving environment. Figure 2 is a diagram showing the contents of table information 15. In Figure 2, for each driving environment, the priority column for sensors 8-11 not used in that driving environment is represented as "-", and the priority column for sensors 8-11 used is represented as "1" or "2". The priority order is priority "1" > priority "2". Specifically, Table Information 15 associates urban areas with LiDAR9 and Camera 10. In urban areas, the priority of LiDAR9 is set higher than that of Camera 10. Also, Table Information 15 associates indoor parking lots with LiDAR9 and Camera 10. In indoor parking lots, the priority of LiDAR9 is set higher than that of Camera 10. Also, Table Information 15 associates tunnels with LiDAR9 and Camera 10. In tunnels, the priority of Camera 10 is set higher than that of LiDAR9. Also, Table Information 15 associates highways with GPS receiver 8, LiDAR9, and Camera 10. In highways, the priority of GPS receiver 8 and Camera 10 is set higher than that of LiDAR9. Also, Table Information 15 associates underpasses with LiDAR9 and Camera 10. Furthermore, under elevated structures, the priority of camera 10 is set higher than that of LiDAR 9. Also, table information 15 associates suburban areas with all of the multiple types of sensors 8-11. Additionally, in suburban areas, the priorities of the multiple types of sensors 8-11 are set to be the same. Note that in table information 15 of Figure 2, "Other" is not registered as a driving environment.

[0017] In the table information 15 of FIG. 2, the priorities of sensors 8 to 11 were set from the viewpoints of estimation accuracy and processing load. For example, the priorities of sensors 8 to 11 with low estimation accuracy were set to "-", and the priorities of sensors 8 to 11 with high estimation accuracy were set to "1" or "2". Among sensors 8 to 11 with high estimation accuracy, the priority of the sensor with the highest estimation accuracy was set to "1", and the priorities of the remaining sensors were set to "2". Also, when the estimation accuracy of all sensors 8 to 11 was low and the processing load of all sensors 8 to 11 was low, the estimation accuracy of all sensors 8 to 11 was set to "1". Note that the estimation accuracy and processing load shown in FIG. 2 were determined by experiments using a real vehicle.

[0018] Next, each function realized by the information processing device 7 will be described in detail. The information processing device 7 realizes the functions of a route generation unit 17, a determination unit 18, a specification unit 19, an estimation unit 20, and a travel control unit 21 as shown in FIG. 3 by the processor 12 executing the program 13. FIG. 3 is a diagram showing each function realized by the information processing device 7. Also, the route generation unit 17, the determination unit 18, the specification unit 19, and the estimation unit 20 together with the HMI 5 and sensors 8 to 11 constitute a vehicle position estimation device 2 that estimates the position of the host vehicle of the vehicle C.

[0019] Before the start of autonomous driving, the route generation unit 17 generates a target route from the current location of the vehicle C to the destination. The target route is, for example, a route for the vehicle C to travel autonomously. Also, as the current location, for example, the position of the host vehicle of the current vehicle C estimated by the estimation unit 20 can be adopted. For example, when the generation of the target route is performed immediately after turning on the ignition switch, the position of the host vehicle at the time when the ignition switch was turned off last time can be used. Also, as the destination, for example, a position manually input by operating the HMI 5 by the driver can be adopted.

[0020] The discrimination unit 18 determines the driving environment around vehicle C. For example, before the start of autonomous driving, the discrimination unit 18 refers to the map information 14 stored in the storage device 16 and identifies the driving environment for each location on the target route generated by the route generation unit 17. Then, during the execution of autonomous driving, based on the identification results (i.e., information indicating the driving environment for each location on the target route) and the current position of vehicle C, it determines where vehicle C is currently traveling on the target route and determines what kind of driving environment the current location of vehicle C is in. As the current location, for example, the current position of vehicle C estimated by the estimation unit 20 can be used.

[0021] The identification unit 19 refers to the table information 15 shown in Figure 2, which is stored in the storage device 16, and identifies sensors 8 to 11 that are associated with the driving environment determined by the discrimination unit 18 (i.e., the driving environment of the location where vehicle C is currently driving). In the following description, the sensors identified by the identification unit 19 will also be referred to as "sensors A1, A2...". Specifically, in the table information 15 shown in Figure 2, sensors with a priority of "1" or "2" are identified, and sensors with a priority of "-" are excluded. For example, if the driving environment is determined to be an urban area, the LiDAR 9 with priority "1" and the camera 10 with priority "2" are identified as sensors A1, A2.... At that time, the identification unit 19 also identifies the priority of sensors A1, A2... from the table information 15.

[0022] The estimation unit 20 estimates the vehicle's position based on the detection results of sensors A1, A2, etc., identified by the identification unit 19. For example, it individually detects the vehicle's position (hereinafter also referred to as "individual position") from each of the detection results of the identified sensors A1, A2, etc., and estimates the vehicle's position by weighting and adding all the detected individual positions. That is, it performs the calculation: Vehicle C's position = weight α * first individual position + weight β * second individual position + ... In this case, the estimation unit 20 makes the weight α of the detection result of sensor A1 (hereinafter also referred to as "first sensor A1"), which has a higher priority than the identified sensors A1, A2, etc., greater than the weight β of the detection results of sensors A2, etc. (hereinafter also referred to as "second sensor A2, etc."), which have a lower priority than the first sensor A (α > β). As a method for calculating individual positions from the LiDAR9 detection results, for example, NDT scan matching, which detects individual positions by matching the point cloud detected by LiDAR9 with the point cloud map, can be employed. Furthermore, as a method for calculating individual positions from the detection results of camera 10, an estimation method can be employed in which individual positions are detected by matching the image captured by camera 10 with an image of a 3D map captured by a virtual camera. Similarly, as a method for calculating individual positions from the detection results of radar 11, an estimation method can be employed in which individual positions are detected by matching the reflection points of the reflected waves detected by radar 11 with landmarks. The detection results of GPS receiver 8 are used directly as individual positions.

[0023] When sensors A1, A2, etc. are identified, the driving control unit 21 starts autonomous driving and causes vehicle C to autonomously drive according to the target route generated by the route generation unit 17. In autonomous driving, the driving control unit 21 generates control signals for the actuator 6 so that vehicle C autonomously drives according to the target route, based on the target route generated by the route generation unit 17, the vehicle's position estimated by the estimation unit 20, external environment information obtained by the external sensor 3, and vehicle information obtained by the vehicle sensor 4.

[0024] (operation) Next, we will explain the process for estimating the position of vehicle C. First, at the start of autonomous driving, default priorities are set for sensors 8-11 (GPS receiver 8, LiDAR 9, camera 10, and radar 11) (S101 in Figure 4). For example, the priority of GPS receiver 8 is set to "1", LiDAR 9 to "2", camera 10 to "3", and radar 11 to "4". Figure 4 is a flowchart showing the overall flow of the vehicle position estimation method in this embodiment. Next, default sensors are set from among sensors 8-11 to be used for estimating the vehicle C's own position (S102 in Figure 4). For example, GPS receiver 8, LiDAR 9, camera 10, and radar 11 are all set as default sensors.

[0025] Next, the route generation unit 17 of the information processing device 7 generates a target route from the current location of vehicle C to the destination (S103 in Figure 4). Next, the discrimination unit 18 acquires map information 14 and table information 15 from the storage device 16 (S104 in Figure 4). Next, the discrimination unit 18 refers to the map information 14 and identifies the driving environment for each location on the target route generated by the route generation unit 17 (S105 in Figure 4). Next, the driving control unit 21 starts autonomous driving and makes vehicle C drive autonomously according to the target route generated by the route generation unit 17 (S106 in Figure 4). Next, the discrimination unit 18 determines where vehicle C is driving on the target route based on the identification result in S105 and the current position of vehicle C, and determines what kind of driving environment the location where vehicle C is currently driving is (S107 in Figure 4). For example, it determines whether it is an urban area, indoor parking lot, tunnel, highway, underpass, suburb, or other.

[0026] Next, the discrimination unit 18 determines whether the same driving environment as the one determined in S107 is registered in the table information 15 (S108 in Figure 4). That is, it determines whether it is anything other than "Other". If the driving environment is determined to be "Other" (S108 "No" in Figure 4), the estimation unit 20 estimates the vehicle C's position based on the detection results of the default sensors (S109 in Figure 4). For example, the vehicle C's position (individual position) is detected individually from each of the default sensor's detection results, and all the detected individual positions are weighted and added together to estimate the vehicle C's position. Also, the weights α and β of the detection results are increased for default sensors with higher default priority. Next, if the discrimination unit 18 determines that the automatic driving has not ended (S110 "No" in Figure 4), it returns to S107 and repeats the flow from S107 to S110, starting with the determination of the current driving environment, thereby performing the above vehicle position estimation many times.

[0027] As the flow from S107 to S110 is repeated, if vehicle C enters an urban area or similar and the driving environment is determined to be something other than "other" (S108 "Yes" in Figure 4), the identification unit 19 refers to the table information 15 and identifies the sensors A1, A2, etc., associated with the driving environment determined in S107, and their priorities (S111 in Figure 4). Subsequently, the estimation unit 20 estimates the vehicle's position based on the detection results of the identified sensors A1, A2, etc. (S112 in Figure 4). For example, the vehicle's position (individual position) is detected individually from each of the sensors A1, A2, etc., and the vehicle's position is estimated by weighting and adding all the detection results of the detected individual positions. Furthermore, the weights α and β of the detection results are increased for sensors A1, A2, etc., which have higher priorities. Next, if the discrimination unit 18 determines that the autonomous driving has not ended (S110 "No" in Figure 4), it returns to S107 and repeats the flow from S107, S108, S111, and S112, starting with the determination of the current driving environment, thereby repeatedly estimating the vehicle's position as described above.

[0028] (modified version) (1) In this embodiment, the estimation unit 20 individually detects the vehicle's own position (individual position) from each of the detection results of sensors A1, A2, etc. identified by the identification unit 19, and estimates the vehicle's own position by weighting and adding all the detected individual positions. However, other configurations can also be adopted. For example, the vehicle's own position may be estimated based only on the detection result of the first sensor (the sensor with the highest priority among the detection results of sensors A1, A2, etc.) from among the identified sensors A1, A2, etc. In this case, when the vehicle's own position estimation is repeated, the vehicle's own position (individual position) is individually detected from each of sensors A1, A2, etc., and the detection accuracy of each detected individual position is calculated. If the detection accuracy of the vehicle's own position obtained from the first sensor is lower than the detection accuracy of the vehicle's own position obtained from the detection result of a sensor with lower priority than the first sensor (second sensor), the vehicle's own position is estimated based only on the detection result of the second sensor among sensors A1, A2, etc. For example, if vehicle C is driving in an indoor parking lot, as shown in Figure 2, the priority of LiDAR9 becomes "1" and the priority of camera 10 becomes "2". Therefore, in an indoor parking lot, the vehicle C's position is estimated based only on the detection results of LiDAR9 among the multiple types of sensors 8 to 11. Then, if the accuracy of vehicle position detection obtained from the LiDAR9 is lower than the accuracy of vehicle position detection obtained from the camera 10, the vehicle C's position is estimated based only on the detection results of camera 10 among the multiple types of sensors 8 to 11. Alternatively, for example, the system may be configured to estimate the vehicle's position based solely on the detection result of the first sensor from among the identified multiple types of sensors 8 to 11, without switching the sensors used to estimate the vehicle's position (i.e., switching from the first sensor to the second sensor).

[0029] (2) In this embodiment, an example was shown in which the table information 15 identifies the correspondence between the driving environment and each sensor (i.e., one sensor at a time), but other configurations can also be adopted. For example, if vehicle C has multiple sensors of the same type (for example, two or more cameras), the table information 15 may be configured to identify the correspondence between the driving environment and the type of sensor, thereby showing the correspondence between the driving environment and the sensors used to estimate the vehicle's position.

[0030] (Effects of this embodiment) (1) As a comparative example, consider a configuration in which the vehicle's position (individual position) is detected individually from the detection results of sensors 8 to 10 (excluding radar 11), and the vehicle's position is estimated by adding and averaging the detected individual positions. In this case, as shown in Figure 5, since the individual position is estimated from the detection results of all sensors 8 to 10 (excluding radar 11), the processing load may increase. In urban areas, the straight-line propagation of signals from GPS satellites is obstructed by buildings, etc., causing signal blockage and multipath errors due to signal reflection. Therefore, as shown in Figure 5, the accuracy of estimating the individual position obtained from the detection results of GPS receiver 8 may decrease in urban areas. For this reason, in the comparative example, when the detection results of GPS receiver 8 are used in urban areas, the accuracy of estimating the vehicle's position may decrease.

[0031] In contrast, in this embodiment, the driving environment around vehicle C, which is equipped with multiple types of sensors 8 to 11 for estimating the vehicle's position, is determined. Next, a table information 15 showing the correspondence between the driving environment and the sensors 8 to 11 used to estimate the vehicle C's position is referred to, and sensors A1, A2, etc. associated with the determined driving environment are identified. Subsequently, the vehicle C's position is estimated based on the detection results of the identified sensors A1, A2, etc. As a result, the process of estimating the vehicle's position from the detection results is performed only on the detection results of sensors A1, A2, etc. that are suitable for the current driving environment. Therefore, as shown in Figure 5, the increase in processing load can be suppressed compared to the above comparative example (i.e., a configuration in which the vehicle C's position is always estimated based on the detection results of sensors 8 to 10). In addition, since the detection results of sensors other than sensors A1, A2, etc., i.e., sensors that are not suitable for the current driving environment, are not used to estimate the vehicle's position, the accuracy of estimating the vehicle C's position can be improved. Thus, in this embodiment, the accuracy of estimating the vehicle's position can be improved while suppressing the increase in processing load. Figure 5 shows the differences between the comparative example and this embodiment. According to Figure 5, this embodiment shows improvements in both estimation accuracy and processing load in urban areas compared to the comparative example. Note that there was no difference in memory usage between this embodiment and the comparative example.

[0032] (2) In this embodiment, when estimating the position of vehicle C, the position of vehicle C (individual position) is detected individually from the detection results of each of the identified sensors A1, A2, etc., and all the detected individual positions are weighted and added together to estimate the position of vehicle C. The weight of the detection result of the sensor with higher priority (first sensor) among the identified sensors A1, A2, etc. is made greater than the weight of the detection result of the sensor with lower priority than the first sensor (second sensor). Alternatively, the position of vehicle C is estimated based only on the detection result of the first sensor among the identified sensors A1, A2, etc. Then, when the detection accuracy of the position obtained from the first sensor is lower than the detection accuracy of the position obtained from the second sensor, the position of vehicle C is estimated based on the detection result of the second sensor. Alternatively, the position of vehicle C is estimated based only on the detection result of the first sensor among the identified sensors A1, A2, etc. This allows for the preferential use of individual positions with high estimation accuracy, thereby improving the estimation accuracy of the position of vehicle C.

[0033] (3) In this embodiment, the driving environment includes urban areas. Subsequently, the multiple types of sensors 8 to 11 include a GPS receiver 8, a LiDAR 9, and a camera 10. Subsequently, the table information 15 associates urban areas with the LiDAR 9 and the camera 10. Here, in urban areas, the straight-line propagation of signals from GPS satellites is obstructed by buildings, etc., causing signal blockage and multipath errors due to signal reflection. Therefore, by not using the detection results of the GPS receiver 8 in urban areas, it is possible to prevent a decrease in the accuracy of estimating the vehicle's position due to signal blockage and multipath errors. Furthermore, compared to a method that also uses the detection results of the GPS receiver 8, for example, the processing load can be reduced.

[0034] (4) In urban areas, the priority of LiDAR9 is set higher than the priority of camera 10. In urban areas, there are many feature points used for the matching process of LiDAR9. Therefore, by setting a higher priority for LiDAR9 in urban areas, the accuracy of estimating the vehicle's position can be improved.

[0035] (5) In this embodiment, the driving environment includes an indoor parking lot and a tunnel. The multiple types of sensors 8 to 11 include a GPS receiver 8, a LiDAR 9, and a camera 10. The table information 15 associates the indoor parking lot and the tunnel with the LiDAR 9 and the camera 10. Here, in the indoor parking lot and the tunnel, the GPS receiver 8 cannot receive signals from GPS satellites and cannot detect the vehicle C's own position. Therefore, by not using the detection results of the GPS receiver 8 in the indoor parking lot and the tunnel, a decrease in the accuracy of estimating the vehicle's own position can be prevented. Also, for example, the processing load can be reduced compared to a method that also uses the detection results of the GPS receiver 8.

[0036] (6) In indoor parking lots and tunnels, the priority of LiDAR9 is set higher than the priority of camera 10. Here, for example, when vehicle C enters an indoor parking lot or tunnel from outdoors and the lighting conditions around vehicle C change, the accuracy of detecting the vehicle's position obtained from the camera 10's detection results may decrease. In contrast, the accuracy of detecting the vehicle's position obtained from the LiDAR9's detection results is not affected by changes in lighting conditions. Therefore, by giving LiDAR9 a higher priority in indoor parking lots and tunnels, the accuracy of estimating the vehicle's position can be improved.

[0037] (7) In this embodiment, the driving environment includes highways. The multiple types of sensors 8 to 11 include a GPS receiver 8, a camera 10, a LiDAR 9, and a radar 11. The table information 15 associates highways with the GPS receiver 8, camera 10, and LiDAR 9. This reduces the processing load compared to a method that also uses the detection results of the radar 11, for example.

[0038] (8) On highways, the priority of the GPS receiver 8 and camera 10 is set higher than the priority of the LiDAR 9. On highways, there are many locations where the point clouds detected by the LiDAR 9 are similar, and the detected point clouds may be matched to point cloud maps of different locations, which may reduce the accuracy of the vehicle's position detection obtained from the LiDAR 9 detection results. Therefore, by lowering the priority of the LiDAR 9 on highways, the accuracy of the vehicle's position estimation can be improved.

[0039] (9) In this embodiment, the driving environment includes under an overpass. The multiple types of sensors 8 to 11 include a GPS receiver 8, a LiDAR 9, and a camera 10. The table information 15 associates under the overpass with the LiDAR 9 and the camera 10. Under the overpass, the GPS receiver 8 can acquire fewer GPS satellites, or it can receive signals from GPS satellites, resulting in a decrease in the accuracy of detecting the vehicle C's own position, or it cannot detect the vehicle's position at all. Therefore, by not using the detection results of the GPS receiver 8 under the overpass, a decrease in the accuracy of estimating the vehicle's own position can be prevented. Furthermore, for example, the processing load can be reduced compared to a method that also uses the detection results of the GPS receiver 8.

[0040] (10) Under the elevated structure, the priority of camera 10 is set higher than the priority of LiDAR 9. Under the elevated structure, there are many locations where the point clouds detected by LiDAR 9 are similar, and the detected point clouds may be matched to point cloud maps of different locations, which may reduce the accuracy of the vehicle's position detection obtained from the LiDAR 9 detection results. Therefore, by lowering the priority of LiDAR 9 under the elevated structure, the accuracy of the vehicle's position estimation can be improved.

[0041] (11) In this embodiment, the driving environment includes suburban areas. The table information 15 associates suburban areas with all of the multiple types of sensors 8 to 11. As shown in Figure 6, in suburban areas, the estimation accuracy of the vehicle's position (individual position) estimated individually from each of the sensors 8 to 11 is low. Therefore, in this embodiment, the estimation accuracy of the vehicle's position is improved by using the detection results of all sensors 8 to 11 in suburban areas. In suburban areas, the processing load and memory usage when estimating individual positions from the detection results of all sensors 8 to 11 are low. Therefore, even when using the detection results of all sensors 8 to 11, the processing load and memory usage remain low. Figure 6 shows the differences between the comparative example and this embodiment. According to Figure 6, it can be seen that this embodiment shows improved estimation accuracy in suburban areas compared to the comparative example. There was no difference in processing load and memory usage between this embodiment and the comparative example. Therefore, according to this embodiment, as shown in Figures 5 and 6, the accuracy of self-position estimation can be improved regardless of whether it is an urban or suburban area, and an increase in processing load and memory usage can be prevented.

[0042] (12) In this embodiment, before the start of autonomous driving of vehicle C, a target route for vehicle C to drive autonomously is generated, and the driving environment is identified for each location on the generated target route. Furthermore, while vehicle C is driving autonomously, based on the identification results and the current position of vehicle C, it is determined where vehicle C is driving on the target route, and the driving environment of the location where vehicle C is currently driving is determined. As a result, the driving environment of each location on the target route is identified before the start of autonomous driving, so the processing load during autonomous driving can be reduced. [Explanation of Symbols]

[0043] 1...Automated driving system, 2...Vehicle position estimation system, 3...External environment sensor, 4...Vehicle sensor, 5...HMI, 6...Actuator, 7...Information processing device, 8...GPS receiver, 9...LiDAR, 10...Camera, 11...Radar, 12...Processor, 13...Program, 14...Map information, 15...Table information, 16...Storage device, 17...Route generation unit, 18...Discrimination unit, 19...Specification unit, 20...Estimation unit, 21...Driving control unit

Claims

1. The vehicle is equipped with multiple types of sensors for estimating its own position, and the surrounding driving environment is determined. Referencing table information showing the correspondence between the driving environment and the sensors used to estimate the vehicle's position, identify the sensor associated with the determined driving environment. Based on the detection results of the identified sensors, the vehicle's own position is estimated. Vehicle position estimation method.

2. In estimating the vehicle's own position, The vehicle's own position is individually detected from each of the detection results of the identified sensors, and the vehicle's own position is estimated by weighting and summing all of the detected individual positions, while the weight of the detection result of the first sensor, which is the sensor with the higher priority among the identified sensors, is made greater than the weight of the detection result of the second sensor, which has a lower priority than the first sensor. Alternatively, the vehicle's position may be estimated based solely on the detection result of the first sensor from among the identified sensors, and if the detection accuracy of the vehicle's position obtained from the first sensor is lower than the detection accuracy of the vehicle's position obtained from the second sensor, the vehicle's position may be estimated based solely on the detection result of the second sensor. Alternatively, the vehicle's position is estimated based solely on the detection result of the first sensor from among the identified sensors. The vehicle position estimation method according to claim 1.

3. The aforementioned driving environment includes urban areas, The various types of sensors include a GPS receiver, LiDAR, and a camera. The aforementioned table information associates the urban area with the LiDAR and the camera. The vehicle position estimation method according to claim 1.

4. In the aforementioned urban area, the priority of the LiDAR is set higher than the priority of the camera. The vehicle position estimation method according to claim 3.

5. The aforementioned driving environment includes indoor parking lots and tunnels. The various types of sensors include a GPS receiver, LiDAR, and a camera. The table information associates the indoor parking lot with the LiDAR and the camera, and associates the tunnel with the LiDAR and the camera. The vehicle position estimation method according to claim 1.

6. In the aforementioned indoor parking lot, the priority of the LiDAR is set higher than the priority of the camera. In the aforementioned tunnel, the priority of the camera is set higher than the priority of the LiDAR. The vehicle position estimation method according to claim 5.

7. The aforementioned driving environment includes highways, The various types of sensors include a GPS receiver, LiDAR, camera, and radar. The table information associates the highway with the GPS receiver, the LiDAR, and the camera. The vehicle position estimation method according to claim 1.

8. On the aforementioned highway, the priority of the GPS receiver and the camera is set higher than the priority of the LiDAR. The vehicle position estimation method according to claim 7.

9. The aforementioned driving environment includes under elevated structures, The various types of sensors include a GPS receiver, LiDAR, and a camera. The aforementioned table information associates the underpass with the LiDAR and the camera. The vehicle position estimation method according to claim 1.

10. Under the elevated structure, the priority of the camera is set higher than the priority of the LiDAR. The vehicle position estimation method according to claim 9.

11. The aforementioned driving environment includes suburban areas, The aforementioned table information associates the suburban area with all of the multiple types of sensors. The vehicle position estimation method according to claim 1.

12. Before the vehicle starts autonomous driving, a target route for the vehicle to drive autonomously is generated, and the driving environment is identified for each location on the generated target route. During the execution of the vehicle's autonomous driving, based on the identification results and the vehicle's current position, it is determined where the vehicle is traveling on the target route, and the driving environment of the location where the vehicle is currently traveling is determined. The vehicle position estimation method according to claim 1.

13. Equipped on the vehicle, it includes multiple types of sensors for estimating the vehicle's position, A determination unit for determining the driving environment around the vehicle, A determination unit refers to table information showing the correspondence between the driving environment and the sensors used to estimate the vehicle's position, and identifies the sensors associated with the driving environment determined by the determination unit, The vehicle comprises an estimation unit that estimates the vehicle's position based on the detection results of the sensor identified by the identification unit. Vehicle position estimation device.

Images