Automated driving system

Abstract

To provide a technique capable of efficiently using the capacity of a storage device when log data related to automated driving control is stored in an in-vehicle storage device in an automated driving system of a vehicle.SOLUTION: An automated driving system comprises a recognition sensor for detecting information around a vehicle, and one or more processors that recognize a situation around the vehicle from detection information detected by the recognition sensor and perform automated driving control of the vehicle on the basis of a recognition result. The one or more processors are further configured to execute storage processing of storing log data of the detection information in the one or more storage devices while performing the automated driving control. In the storage processing, the one or more processors identify a low-reliability area around the vehicle where the reliability of the recognition result is low, and preferentially store the log data of target detection information that is the detection information related to the low-reliability area when the low-reliability area is identified.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present disclosure relates to an autonomous driving system installed in a vehicle. [Background technology]

[0002] There is known a technology for controlling automatic driving of a vehicle using a machine learning model. Patent Document 1 discloses a method for collecting training data that can be used for learning a machine learning model. Other documents that show the technical level of this technical field include the following Patent Documents 2 and 3. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2019 / 116423 [Patent Document 2] JP 2021-174413 A [Patent Document 3] Patent Publication No. 2022-023956 Summary of the Invention [Problem to be solved by the invention]

[0004] As a method for retroactively verifying the autonomous driving control of a vehicle, it has been considered to store log data related to the autonomous driving control in an on-board storage device. However, since the capacity of the on-board storage device is limited, if the capacity of the storage device becomes tight, it may be impossible to store necessary log data. In particular, among the log data related to the autonomous driving control, log data of detection information from a recognition sensor that detects information around the vehicle is particularly necessary for verifying the autonomous driving control, but the amount of data tends to be relatively large.

[0005] In view of the above-mentioned problems, one object of the present disclosure is to provide a technique that enables efficient use of the capacity of a storage device. [Means for solving the problem]

[0006] One aspect of the present disclosure relates to an autonomous driving system mounted on a vehicle. The autonomous driving system includes a recognition sensor that detects information about the surroundings of the vehicle, one or more processors that recognize the surroundings of the vehicle from the detection information detected by the recognition sensor and perform autonomous driving control of the vehicle based on the recognition result, and one or more storage devices. The one or more processors are configured to further execute a storage process for storing log data of the detection information in the one or more storage devices while executing the autonomous driving control. In the storage process, the one or more processors include identifying a low reliability area in which the reliability of the recognition result is low around the vehicle, and when the low reliability area is identified, preferentially storing log data of target detection information related to the low reliability area among the detection information. Effect of the Invention

[0007] According to the present disclosure, the log data of the target detection information related to the low reliability area of ​​the detection information where the reliability of the recognition result is low is preferentially stored. This allows the capacity of the storage device to be used efficiently in view of the high necessity for the verification of the autonomous driving control. [Brief description of the drawings]

[0008] [Figure 1] FIG. 2 is a diagram illustrating an example of a configuration related to automatic driving control of a vehicle according to an embodiment. [Diagram 2] FIG. 2 is a diagram illustrating an example of a hardware configuration of the autonomous driving system according to the embodiment. [Diagram 3] 4 is a flowchart showing a process executed by a processor according to the embodiment. [Figure 4] 11A and 11B are diagrams illustrating low-reliability regions and object detection information. [Diagram 5] 11 is a diagram for explaining preferential storage of log data of object detection information. FIG. [Figure 6] 11 is a diagram for explaining information related to a specific recognition result; FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] 1.Autonomous Driving System This embodiment relates to an automatic driving system that is mounted on a vehicle and controls automatic driving of the vehicle. FIG. 1 is a diagram showing an example of a configuration related to automatic driving control of a vehicle 1 by the automatic driving system according to this embodiment. Automatic driving refers to automatically performing at least one of steering, acceleration, and deceleration of the vehicle 1 without the driver's driving operation. Automatic driving control is a concept that includes not only fully automatic driving control, but also risk avoidance control, lane keep assist control, and the like. The operator may be a driver who rides in the vehicle 1, or a remote operator who remotely operates the vehicle 1.

[0010] The vehicle 1 includes a sensor group 10, a recognition unit 20, a planner 30, a control amount calculation unit 40, and a traveling device 50.

[0011] The sensor group 10 includes a recognition sensor 11 that detects information about the surroundings of the vehicle 1. Examples of the recognition sensor 11 include a camera, a LIDAR (Laser Imaging Detection and Ranging), a radar, and the like. The sensor group 10 may further include a status sensor 12 that detects the status of the vehicle 1, a position sensor 13 that detects the position of the vehicle 1, and the like. Examples of the status sensor 12 include a speed sensor, an acceleration sensor, a yaw rate sensor, a steering angle sensor, and the like. Examples of the position sensor 13 include a GNSS (Global Navigation Satellite System) sensor.

[0012] The sensor detection information SEN is information detected by the sensor group 10. For example, the sensor detection information SEN includes image data captured by a camera. Furthermore, the sensor detection information SEN may include information (relative position, relative speed, class, etc.) regarding a specific object (pedestrian, preceding vehicle, white line, bicycle, road sign, etc.) captured in the image. For example, the sensor detection information SEN may include point cloud data detected by LIDAR. For example, the sensor detection information SEN may include information regarding the relative position and relative speed of an object detected by radar. The sensor detection information SEN may include vehicle state information indicating the state of the vehicle 1. The sensor detection information SEN may include position information indicating the position of the vehicle 1.

[0013] The recognition unit 20 receives the sensor detection information SEN. The recognition unit 20 recognizes the situation around the vehicle 1 based on the sensor detection information SEN detected by the recognition sensor 11. For example, the recognition unit 20 recognizes objects around the vehicle 1. Also, for example, the recognition unit 20 generates a spatial map that integrates the recognized objects into map information. Examples of objects include pedestrians, other vehicles (preceding vehicles, parked vehicles, etc.), white lines, road structures (e.g., guardrails, curbs), fallen objects, traffic lights, intersections, signs, etc. Furthermore, the recognition unit 20 may predict the behavior of objects around the vehicle 1. Furthermore, the recognition unit 20 may generate a risk map around the vehicle 1. The recognition result information RES indicates the recognition result by the recognition unit 20.

[0014] The planner 30 receives the recognition result information RES from the recognition unit 20. The planner 30 generates a driving plan for the vehicle 1 based on the received recognition result information RES. The planner 30 may further receive vehicle state information and position information. The driving plan may be for reaching a preset destination or for avoiding risks. In the driving plan, for example, driving decisions such as maintaining the current driving lane, changing lanes, overtaking, turning right or left, steering, accelerating, decelerating, stopping, etc. are given. Furthermore, the planner 30 generates a target trajectory TRJ required for the vehicle 1 to travel according to the driving plan. The target trajectory TRJ includes a target position and a target speed.

[0015] The control amount calculation unit 40 receives the target trajectory TRJ from the planning unit 30. The control amount calculation unit 40 calculates a control amount CON required for the vehicle 1 to follow the target trajectory TRJ. It can also be said that the control amount CON is a control amount required to reduce the deviation between the vehicle 1 and the target trajectory TRJ. The control amount CON includes at least one of a steering control amount, a drive control amount, and a braking control amount. Examples of the steering control amount include a target steering angle, a target torque, a target motor angle, a target motor drive current, etc. Examples of the drive control amount include a target drive force, a target engine torque, etc. Examples of the braking control amount include a target braking force, a target braking torque, etc.

[0016] The traveling device 50 includes a steering device 51, a driving device 52, and a braking device 53. The steering device 51 steers the wheels. The driving device 52 is a power source that generates a driving force. Examples of the driving device 52 include an engine, an electric motor, and an in-wheel motor. The braking device 53 generates a braking force. The traveling device 50 receives a control amount CON from the control amount calculation unit 40. The traveling device 50 operates the steering device 51, the driving device 52, and the braking device 53 according to the steering control amount, the driving control amount, and the braking control amount, respectively. As a result, the vehicle 1 travels so as to follow the target trajectory TRJ.

[0017] The recognition unit 20 includes at least one of a rule-based model and a machine learning model. The rule-based model performs recognition processing based on a group of predetermined rules. Examples of the machine learning model include a Neural Network (NN), a Support Vector Machine (SVM), a regression model, a decision tree model, and the like. The NN may be a Convolutional Neural Network (CNN), a Recurrent Neural Network (RNN), or a combination thereof. The type of each layer, the number of layers, and the number of nodes in the NN are arbitrary. The machine learning model is generated in advance through machine learning. The recognition unit 20 performs recognition processing by inputting the sensor detection information SEN to the model. The recognition result information RES is output from the model, or is generated based on the output from the model.

[0018] Similarly, the planner 30 includes at least one of a rule-based model and a machine learning model. The planner 30 performs planning processing by inputting the recognition result information RES to the model. The target trajectory TRJ is output from the model or is generated based on the output from the model.

[0019] Similarly, the control amount calculation unit 40 includes at least one of a rule-based model and a machine learning model. The control amount calculation unit 40 performs a control amount calculation process by inputting a target trajectory TRJ to the model. The control amount CON is output from the model or is generated based on the output from the model.

[0020] Two or more of the recognition unit 20, the planner 30, and the control amount calculation unit 40 may be integrally configured. All of the recognition unit 20, the planner 30, and the control amount calculation unit 40 may be integrally configured (end-to-end configuration). For example, the recognition unit 20 and the planner 30 may be integrally configured by a NN that outputs a target trajectory TRJ from the sensor detection information SEN. Even in the case of an integral configuration, intermediate products such as recognition result information RES and target trajectory TRJ may be output. For example, when the recognition unit 20 and the planner 30 are integrally configured by a NN, the recognition result information RES may be the output of an intermediate layer of the NN.

[0021] The recognition unit 20, the planner 30, and the control amount calculation unit 40 constitute an “automatic driving control unit” that controls the automatic driving of the vehicle 1.

[0022] 2 is a diagram showing an example of a hardware configuration of the autonomous driving system 100 according to this embodiment. The autonomous driving system 100 has at least the functions of the autonomous driving control unit described above. The autonomous driving system 100 may further include a sensor group 10 and a driving device 50.

[0023] The autonomous driving system 100 includes one or more processors 110 (hereinafter simply referred to as processors 110) and one or more memory devices 120 (hereinafter simply referred to as memory devices 120).

[0024] The processor 110 executes various processes. The processor 110 can be configured, for example, by a central processing unit (CPU), a graphics processing unit (GPU), etc. The recognition unit 20, the planning unit 30, and the control amount calculation unit 40 may be realized by a single processor 110, or may be realized by separate processors 110. The storage device 120 stores various information required for the execution of processes by the processor 110. The storage device 120 can be configured, for example, by a recording medium such as a read only memory (ROM), a random access memory (RAM), etc.

[0025] The storage device 120 stores a computer program 121, model data 122, and log data LOG.

[0026] The computer program 121 is executed by the processor 110. Various processes by the autonomous driving system 100 may be realized by cooperation between the processor 110, which executes the computer program 121, and the storage device 120. The computer program 121 may be recorded in a computer-readable recording medium.

[0027] The model data 122 is data of models included in the recognition unit 20, the planning unit 30, and the control amount calculation unit 40. When executing processing related to autonomous driving control, the processor 110 configures the recognition unit 20, the planning unit 30, and the control amount calculation unit 40 by selecting and using a model from the model data 122.

[0028] The processor 110 collects log data LOG related to the autonomous driving control while executing the autonomous driving control. The processor 110 stores the collected log data LOG in the storage device 120. It is assumed that the stored log data LOG is used for verifying the autonomous driving control. In this embodiment, the collected and stored log data LOG includes at least log data of the sensor detection information SEN detected by the recognition sensor 11. The log data LOG may also include log data of the sensor detection information SEN detected by the status sensor 12 and the position sensor 13.

[0029] In addition, the log data LOG may include log data of the control amount CON output from the automatic driving control unit, log data of the recognition result information RES output from the recognition unit 20, log data of the target trajectory TRJ output from the planning unit 30, etc.

[0030] 2. Save log data of detected information As described above, in this embodiment, while executing the automatic driving control, the processor 110 collects log data of at least the sensor detection information SEN detected by the recognition sensor 11 and stores it in the storage device 120. The automatic driving control unit recognizes the situation around the vehicle 1 from the sensor detection information SEN detected by the recognition sensor 11, and performs automatic driving control of the vehicle 1 based on the recognition result. For this reason, the log data of the sensor detection information SEN detected by the recognition sensor 11 is data that is particularly required for verifying the automatic driving control. Hereinafter, the sensor detection information SEN detected by the recognition sensor 11 will be simply referred to as "detection information."

[0031] The log data of the detection information includes image data, point cloud data, etc., and the amount of data tends to be relatively large. On the other hand, the capacity of the storage device 120 that stores the log data LOG is limited. If the amount of data of the log data LOG becomes large and the capacity of the storage device 120 becomes tight, a situation may occur in which necessary log data cannot be stored. For this reason, when storing log data, it is required to efficiently use the capacity of the storage device 120.

[0032] Incidentally, the recognition result information RES may include a recognition result with high reliability or may include a recognition result with low reliability. For example, the recognition result information RES regarding the recognized objects around the vehicle 1 may include a recognized object whose position, state, behavior prediction, etc. are recognized with high reliability, or may include a recognized object whose position, state, behavior prediction, etc. are recognized with low reliability. Autonomous driving control performed based on a recognition result with low reliability is likely to deviate from the intended operation of the vehicle 1. Therefore, when verifying the autonomous driving control after the fact, it is assumed that a recognition result with low reliability is more likely to require verification than a recognition result with high reliability. In other words, it is considered that log data of detection information related to a recognition result with low reliability is more necessary in verifying the autonomous driving control.

[0033] From the above viewpoint, the autonomous driving system 100 according to this embodiment makes it possible to efficiently utilize the capacity of the storage device 120.

[0034] 3 is a flowchart showing an example of a process (storage process) for storing log data of detection information, which is executed by processor 110. The process according to the flowchart shown in FIG. 3 is executed repeatedly at a predetermined processing cycle, for example, while the automatic driving control is being executed.

[0035] First, in step S110, the processor 110 acquires various pieces of information. The processor 110 acquires at least detection information and recognition result information RES. In this embodiment, the recognition result information RES is generated so as to include information on the reliability of each recognition result. For example, when the recognition unit 20 is configured with a machine learning model, the reliability is the confidence of the machine learning model for each recognition result. Also, for example, when the recognition unit 20 is configured with a rule-based model based on a statistical model, the reliability is the probability of each recognition result in the statistical model.

[0036] Next, in step S120, the processor 110 refers to the recognition result information RES to identify an area in which the reliability of the recognition result is low (hereinafter, referred to as a "low reliability area") around the vehicle 1. For example, the processor 110 identifies an area including a recognition result whose reliability is equal to or lower than a predetermined threshold as a low reliability area.

[0037] FIG. 4(A) is a diagram showing an example of the low reliability region 2. In FIG. 4(A), a plurality of recognized objects OBj (j=1, 2, ..., 9) around the vehicle 1 and the reliability level of each recognized object OBj are shown as the recognition result information RES. In the example shown in FIG. 4(A), the reliability of the recognized objects OB7, OB8, and OB9 is low. Therefore, the processor 110 specifies the low reliability region 2 so as to include the recognized objects OB7, OB8, and OB9. In this case, the shape and range of the low reliability region 2 may be suitably determined according to the environment to which the present embodiment is applied. The low reliability region 2 may include two or more regions. For example, in the example shown in FIG. 4(A), the processor 110 may specify three regions for each of the recognized objects OB7, OB8, and OB9 as the low reliability region 2.

[0038] If there is no recognition result with low reliability, the processor 110 may not specify the low reliability area 2. In addition, if a risk map is provided as the recognition result information RES, the processor 110 may specify, for example, an area including a point with a low reliability of the risk value as the low reliability area.

[0039] 3 again. If the low-confidence region 2 is not identified in step S120 (step S130; No), the processor 110 stores the log data of the detection information in a normal procedure (step S160). The normal procedure may be any known suitable procedure. After step S160, the current process ends.

[0040] If a low-reliability region 2 is identified in step S120 (step S130; Yes), the process proceeds to step S140.

[0041] In step S140, the processor 110 determines information related to the low reliability region 2 from among the detection information (hereinafter referred to as "target detection information"). The target detection information is, for example, image data captured by a camera whose imaging range includes the low reliability region 2, point cloud data detected by a LIDA whose detection range includes the low reliability region 2, point cloud data detected for the low reliability region 2, information on an object detected by a radar whose detection range includes the low reliability region 2, information on an object detected in the low reliability region 2, and the like.

[0042] FIG. 4B is a diagram for explaining an example of target detection information. In FIG. 4B, four detection ranges D1, D2, D3, and D4 of the recognition sensor 11 are shown. That is, in the example shown in FIG. 4B, detection information of the recognition sensor 11 corresponding to each of the four detection ranges D1, D2, D3, and D4 is acquired. As shown in FIG. 4B, the detection ranges D1 and D4 include the low reliability area 2. On the other hand, the detection ranges D2 and D3 do not include the low reliability area 2. Therefore, in this case, the processor 110 determines that the detection information of the recognition sensor 11 having the detection range D1 and the detection information of the recognition sensor 11 having the detection range D4 are the target detection information among the acquired detection information.

[0043] Referring again to FIG. 3, after step S140, the process proceeds to step S150. In step S150, the processor 110 stores the log data of the object detection information with priority. That is, when the low-reliability region 2 is identified, the object detection information related to the low-reliability region 2 is treated with priority in storing the log data of the detection information. After step S150, the process ends.

[0044] 5A, 5B, and 5C are conceptual diagrams showing an example of preferential storage of log data of target detection information, respectively. In particular, each of Figs. 5A, 5B, and 5C shows a case where the processor 110 stores compressed detection information as log data.

[0045] In the example shown in FIG. 5(A), the processor 110 stores the log data with a compression rate of the object detection information lower than that of the other detection information. The higher the compression rate, the more the amount of data can be reduced, but the more information is lost. Therefore, by setting the compression rate of the object detection information lower than that of the other detection information, the log data of the object detection information can be stored richer with emphasis on preserving information, while the log data of the other detection information can be stored with emphasis on reducing the amount of data. In particular, the processor 110 may be configured to apply lossless compression to the object detection information and lossy compression with a high compression rate to the other detection information. In this way, in the example shown in FIG. 5(A), the log data of the object detection information is preferentially stored.

[0046] In the example shown in FIG. 5(B), the processor 110 stores the log data without compressing the object detection information. This can also be considered as a compression rate of 0% for the object detection information. In this sense, the example shown in FIG. 5(B) can also be considered as a part of the example shown in FIG. 5(A). Therefore, in the example shown in FIG. 5(B), as in the example shown in FIG. 5(A), preferential storage of the object detection information log data is realized.

[0047] In the example shown in Fig. 5(C), the processor 110 stores log data only for the object detection information. In other words, in the example shown in Fig. 5(C), log data for other detection information is not stored. In this way, in the example shown in Fig. 5(C), the log data for the object detection information is preferentially stored.

[0048] The processor 110 may be configured to switch the method of saving the log data of the target detection information depending on the free space of the storage device 120. For example, the processor 110 may be configured to save the log data of the detection information as shown in Fig. 5(B) when there is sufficient free space, to save the log data of the detection information as shown in Fig. 5(A) when the free space falls below a first threshold, and to save the log data of the detection information as shown in Fig. 5(C) when the free space falls further below a second threshold.

[0049] In this manner, in the autonomous driving system 100 according to this embodiment, the processor 110 executes processing related to the storage of log data of the detection information. The processor 110 executes processing in this manner, thereby realizing a log data storage method according to this embodiment. The computer program 121 that causes the processor 110 to execute processing in this manner realizes a log data storage program according to this embodiment.

[0050] 3.Effects As described above, according to this embodiment, the log data of the target detection information is preferentially stored. The target detection information is information related to the low reliability area 2 where the reliability of the recognition result is low among the detection information, and is data that is more necessary for verifying the autonomous driving control. Therefore, according to this embodiment, the capacity of the storage device 120 can be efficiently used from the viewpoint of the high necessity for verifying the autonomous driving control.

[0051] 4. Variations The autonomous driving system 100 according to this embodiment may adopt the following modified aspects.

[0052] 4-1. Saving log data of target detection information Each recognition result included in the recognition result information RES may or may not be involved in the calculation of the control amount of the automatic driving control. For example, when the planner 30 generates a driving plan or a target trajectory TRJ based on the recognition result information RES, a part of the recognition result information RES may not be involved in the generation of the driving plan or the target trajectory TRJ. When the automatic driving control is verified after the fact, it is assumed that the recognition result involved in the calculation of the control amount of the automatic driving control (hereinafter, referred to as a "specific recognition result") is more likely to require verification than the recognition result not involved in the calculation of the control amount of the automatic driving control. In other words, it is considered that the log data of the detection information related to the specific recognition result is more necessary in the verification of the automatic driving control.

[0053] Therefore, when processor 110 stores the log data of the object detection information, processor 110 may be configured to store the log data of the information related to the specific recognition result from among the object detection information with a higher priority. For example, processor 110 stores the log data with a compression rate of the information related to the specific recognition result lower than the compression rate of other object detection information.

[0054] 6(A) and (B) are diagrams for explaining information related to a specific recognition result. FIG. 6(A) shows a target trajectory TRJ generated by the recognition unit 20. In FIG. 6(A), the target trajectory TRJ is generated so as to avoid a recognized object OB1, which is a parked vehicle parked on the roadside. That is, the recognized object OB1 is a specific recognition result. In FIG. 6(B), the four detection ranges D1, D2, D3, and D4 of the recognition sensor 11 and the low reliability area 2 are shown, similar to the case shown in FIG. 4(B). Furthermore, FIG. 6(B) shows a recognized object OB1, which is a specific recognition result. In FIG. 6(B), the processor 110 determines, among the acquired detection information, the detection information of the recognition sensor 11 having the detection range D1 and the detection information of the recognition sensor 11 having the detection range D4 as target detection information, similar to the case shown in FIG. 4(B). Furthermore, the detection range D1 includes the recognized object OB1, which is a specific recognition result. In other words, the detection information of the recognition sensor 11 having the detection range D1 is information related to a specific recognition result. Therefore, in this case, when storing the log data of the target detection information, the processor 110 stores the detection information of the recognition sensor 11 having the detection range D1 with higher priority.

[0055] By adopting such a modified embodiment, the capacity of the storage device 120 can be used more efficiently in view of the high necessity for verifying autonomous driving control.

[0056] 4-2. Reproduction data In the above, a case has been described in which the processor 110 stores the compressed detection information as log data. Furthermore, in order to reduce the amount of log data, the processor 110 may be configured to generate reproduction data that artificially reproduces the detection information. Then, the processor 110 may be configured to store the generated reproduction data as log data.

[0057] For example, when the detection information is image data, the processor 110 generates, as reproduction data, information capable of generating an image that reproduces and displays an object captured in the original image data in a pseudo manner. In this case, the reproduction data can be, for example, information on the position, state (blurredness, darkness, etc.), shape, etc., of the object captured in the original image data. The processor 110 may be configured to generate the reproduction data using parameters in the approximation space of the image data. Also, for example, when the detection information is point cloud data, the processor 110 generates, as reproduction data, information capable of generating a point cloud that reproduces the original point cloud data in a pseudo manner. In this case, the reproduction data can be, for example, information on the intensity of reflected light detected by the LIDAR when the original point cloud data is detected.

[0058] Since the reproduction data is information that reproduces the detection information in a pseudo manner, the amount of data is smaller than that of the original detection information. Furthermore, the detection information reproduced in a pseudo manner using the reproduction data is expected to output recognition result information RES similar to the original detection information. Therefore, by adopting such a modified embodiment, the amount of data in the log data can be reduced while ensuring the verifiability of the automatic driving control.

[0059] The amount of data for reproduction depends on how faithfully the original information is reproduced. In other words, the more faithful the reproduction, the less information is lost, but the amount of data is greater. The processor 110 may be configured to generate reproduction data for the object detection information that reproduces the original information more faithfully than other detection information. This configuration also makes it possible to realize preferential storage of the log data of the object detection information. Furthermore, processor 110 may be configured to store the compressed reproduction data as log data. In this case, processor 110 may also be configured to realize preferential storage of the log data of the object detection information, as described in FIGS. 5(A), (B), and (C). [Explanation of symbols]

[0060] 1 Vehicle, 2 Low-reliability area, 11 Recognition sensor, 100 Autonomous driving system, 110 processor, 120 storage device, LOG log data, RES recognition result information

Claims

1. An automated driving system installed in a vehicle A recognition sensor that detects information about the surroundings of the vehicle; One or more processors that recognize a situation around the vehicle from detection information detected by the recognition sensor and perform automatic driving control of the vehicle based on a recognition result; one or more storage devices; Including, The one or more processors are configured to further execute a storage process of storing log data of the detection information in the one or more storage devices while the autonomous driving control is being executed; In the storage process, the one or more processors identifying a low confidence region around the vehicle where the confidence of the recognition result is low; When the low-reliability region is identified, storing the log data of the object detection information related to the low-reliability region among the detection information with priority; Includes Autonomous driving system.

2. The automated driving system according to claim 1, In the storage process, the one or more processors Acquiring a specific recognition result involved in calculation of a control amount of the automatic driving control from among the recognition results; storing the log data of information related to the specific recognition result among the object detection information with a higher priority; Includes Characterized by Autonomous driving system.

3. The automated driving system according to claim 1 or 2, In the storing process, the one or more processors store the compressed detection information as the log data, Preferentially storing the log data of the object detection information includes lowering the compression rate of the object detection information compared to the compression rate of other detection information. Characterized by Autonomous driving system.

4. The automated driving system according to claim 1 or 2, In the storage process, the one or more processors generating reproduction data for pseudo-reproducing the detection information; storing the reproduction data as the log data; Includes Characterized by Autonomous driving system.

Images