Automatic driving system

Abstract

To provide an automatic driving system capable of ensuring storage of log data used for post-verification of a calculation result of automatic driving control.SOLUTION: An automatic driving system 100 stores output data about automatic driving control in a storage part 7 of a vehicle 1, the output data calculated based on a result detected by a sensor 2 of the vehicle 1. The automatic driving system 100 comprises: a log acquisition part for associating the output data and a driving scene where the output data is calculated, and acquiring them as log data; an influence degree acquisition part for acquiring a preset influence degree at which the driving scene gives influence to the periphery of the vehicle 1 based on the driving scene; and a log storage part for storing the log data in the storage part 7 of the vehicle 1 at a priority degree according to the influence degree of the driving scene associated with the log data.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to automated driving systems. [Background technology]

[0002] There is a known technology that assigns a label to external environmental information collected by a camera mounted on the vehicle depending on whether or not the vehicle has taken evasive action (for example, International Publication No. 2019 / 116423). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2019 / 116423 Summary of the Invention [Problem to be solved by the invention]

[0004] In an autonomous driving system, calculations for autonomous driving control are performed based on the detection results of on-board sensors. In order to perform post-mortem verification of the calculation results for autonomous driving control, it is conceivable to collect and store output data related to autonomous driving control while the vehicle is in operation as log data. However, if the amount of output data becomes large, the vehicle's memory unit may not be able to store all the log data, which may cause problems in collecting the log data.

[0005] The present disclosure aims to provide an autonomous driving system that can ensure the storage of log data used for post-mortem verification of calculation results of autonomous driving control. [Means for solving the problem]

[0006] One aspect of the present disclosure is an autonomous driving system that stores output data related to autonomous driving control calculated based on the detection results of the vehicle's onboard sensors in a vehicle memory unit, and includes a log acquisition unit that associates the driving scene for which the output data was calculated with the output data and acquires it as log data, an impact acquisition unit that acquires a predetermined impact degree that the driving scene has on the surroundings of the vehicle based on the driving scene, and a log memory unit that stores the log data in the vehicle memory unit with a priority according to the impact degree of the driving scene associated with the log data.

[0007] In the autonomous driving system according to an embodiment of the present disclosure, log data in which output data and driving scenes are associated with each other is stored in the storage unit of the vehicle with a priority according to the impact of the driving scene. For example, useful log data can be effectively collected by preferentially storing log data corresponding to driving scenes that have a large impact on the surroundings of the vehicle in the storage unit of the vehicle. Therefore, according to the autonomous driving system according to an embodiment of the present disclosure, it is possible to ensure the storage of log data used for post-mortem verification of the calculation results of the autonomous driving control.

[0008] In one embodiment, the autonomous driving system includes a capacity shortage prediction unit that predicts a future capacity shortage of the vehicle's storage unit based on the time of autonomous driving control to the vehicle's destination and road conditions, and the log storage unit may store log data in the vehicle's storage unit with a priority according to the degree of impact when a capacity shortage of the vehicle's storage unit is predicted. For example, when the vehicle's storage unit runs out of capacity, there is a possibility that the recording of log data corresponding to a driving scene that has a large impact on the surroundings of the vehicle will fail. Therefore, when a capacity shortage of the vehicle's storage unit is predicted, the log data corresponding to the driving scene that has a large impact on the surroundings of the vehicle is stored in the vehicle's storage unit with a priority according to the degree of impact, thereby ensuring the storage of log data corresponding to the driving scene that has a large impact on the surroundings of the vehicle.

[0009] In one embodiment, the autonomous driving system includes an external storage unit that stores log data transmitted from the vehicle via a network, and a communication unit that is provided in the vehicle and communicates with the external storage unit, and the log storage unit may transmit the log data to the external storage unit and delete the transmitted log data from the storage unit of the vehicle. In this case, by storing the log data in the external storage unit and deleting the log data from the storage unit of the vehicle, it is possible to ensure the storage of the log data while preventing a shortage of capacity in the storage unit of the vehicle. Effect of the Invention

[0010] According to the present disclosure, it is possible to ensure the storage of log data used for post-mortem verification of calculation results of autonomous driving control. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 is a schematic block diagram illustrating a configuration of an autonomous driving system according to an embodiment. [Diagram 2] FIG. 2 is a schematic block diagram illustrating a configuration of a vehicle in the autonomous driving system of FIG. 1. [Diagram 3] 2 is a flowchart illustrating a process of the ECU of FIG. 1; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, exemplary embodiments will be described with reference to the drawings. In the following description, the same or corresponding elements are denoted by the same reference numerals, and duplicated description may be omitted.

[0013] As shown in FIG. 1 and FIG. 2, an autonomous driving system 100 according to the embodiment is a system that autonomously drives a vehicle 1. The vehicle 1 is not particularly limited and may be a passenger car or a cargo vehicle. The vehicle 1 is an autonomous driving vehicle capable of autonomous driving. The vehicle 1 may also be capable of manual driving by a driver.

[0014] The autonomous driving system 100 stores output data related to autonomous driving control, which is calculated based on detection results of an on-board sensor of the vehicle 1, in a storage unit of the vehicle 1. The autonomous driving system 100 includes a sensor 2, a map database 3, a GNSS receiver 4, an actuator 5, a communication unit 6, a storage unit 7, and an ECU [Electronic Control Unit] 10.

[0015] The sensor 2 is an on-board sensor that detects information regarding at least one of the driving state and the surrounding conditions of the vehicle 1. The sensor 2 has an external sensor and an internal sensor. The external sensor is a sensor that acquires the surrounding environment of the vehicle 1. The external sensor includes at least one of a camera, a millimeter wave radar, a LiDAR (Light Detection and Ranging), etc. The internal sensor is a detection device that detects the driving state of the vehicle 1. The internal sensor includes at least one of a vehicle speed sensor, an acceleration sensor, a yaw rate sensor, etc. The sensor 2 transmits the detection result to the ECU 10.

[0016] The map database 3 is a storage device that stores map information. The map database 3 is formed in a storage medium such as a HDD (Hard Disk Drive) mounted on the vehicle 1. The map information includes information such as the positions and shapes of roads, intersections, and branching points (for example, the positions and shapes of intersections, the widths of lanes, the positions of division lines and crosswalks, etc.). The map information may also include information on structures (such as the positions and types of traffic lights installed at intersections) and various types of traffic rule information (such as one-way streets, no right or left turns, and no turns). Note that a part of the map information included in the map database 3 may be stored in a storage device other than the HDD. The map database 3 may be formed in a computer in a facility such as a management center that can communicate with the vehicle 1.

[0017] The GNSS receiver 4 receives signals from positioning satellites to measure the current vehicle position (for example, the latitude and longitude of the vehicle 1) of the vehicle 1. The GNSS receiver 4 transmits the measured position information of the vehicle 1 to the ECU 10.

[0018] The actuator 5 is a controller for controlling the traveling of the vehicle 1. The actuator 5 may include, for example, an actuator for controlling the output of an engine or a motor, or a brake actuator, in order to control the speed of the vehicle 1. The actuator 5 may include, for example, a steering actuator for controlling a steering angle or a steering torque in order to control the steering of the vehicle 1.

[0019] The communication unit 6 is a communication device that controls wireless communication with the outside of the vehicle 1. The communication unit 6 communicates various information with a server (external storage unit) 50 for an autonomous driving system via a communication network N, for example. The communication unit 6 communicates various information with other vehicles (external storage units) 20 around the vehicle 1 via the communication network N, for example. The communication unit 6 is not particularly limited, and various known communication devices can be used.

[0020] The storage unit 7 stores log data to be used for post-verification of the calculation results of the autonomous driving control. The storage unit 7 is mounted on the vehicle 1. The storage unit 7 is, for example, a recording device such as an HDD mounted on the vehicle 1. The storage unit 7 has a log storage area for storing the log data. The log storage area has a finite capacity. The log data will be described in detail later.

[0021] The ECU 10 is an electronic control unit having a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), etc. The ECU 10 realizes various functions by, for example, loading a program recorded in the ROM into the RAM and executing the program loaded into the RAM with the CPU. The ECU 10 may be composed of multiple electronic control units. The ECU 10 has, as its functional configuration, an automatic driving processing unit 11, a log acquisition unit 12, an impact degree acquisition unit 13, a capacity shortage prediction unit 14, and a log storage unit 15.

[0022] The automatic driving processing unit 11 executes automatic driving control based on the detection result of the sensor 2. The automatic driving processing unit 11 calculates output data related to automatic driving control based on the detection result of the sensor 2, map information, and the vehicle position. The output data is calculated in the ECU 10 and output to the actuator 5, and is data for performing automatic driving. The output data includes, for example, data for controlling at least one of driving, braking, and steering of the vehicle 1.

[0023] As an example, the automatic driving processing unit 11 calculates output data related to automatic driving control using a machine learning model 11a. For example, the automatic driving processing unit 11 inputs the detection result of the sensor 2 to the machine learning model 11a and outputs a control signal obtained thereby to the actuator 5, thereby controlling the driving, braking, and steering of the vehicle 1 to perform automatic driving.

[0024] The machine learning model 11a is a recurrent deep learning model. The machine learning model 11a is a recurrent neural network [RNN: Recurrent neural network]. A convolutional neural network [CNN: Convolutional Neural Network] including a plurality of layers including a plurality of convolution layers and a pooling layer may be used as at least a part of the neural network. In the machine learning model 11a, deep learning by deep learning is performed. The machine learning model 11a is a trained model trained using data of the vehicle 1 under a predetermined training condition. For example, the machine learning model 11a may be trained using teacher data regarding the output to the actuator 5 when various detection results of the sensor 2 are input.

[0025] The log acquisition unit 12 specifies the driving scene when the automatic driving processing unit 11 calculates the output data based on the detection result of the sensor 2 of the vehicle 1, the map information, and the vehicle position. The driving scene means the driving situation of the vehicle 1. The driving scene includes, for example, the road shape (straight road shape, curved shape, etc.) of the road on which the vehicle 1 is driving, the situation of other vehicles around the vehicle 1, and the progress status of the vehicle 1. Examples of the driving scene include lane changes, turning right or left at an intersection, the approach of an emergency vehicle, passing through a construction site, driving in a traffic jam, and driving in bad weather. Bad weather includes, for example, rain, snow, fog, etc.

[0026] The log acquisition unit 12 associates the driving scene for which the output data was calculated with the output data and acquires it as log data. The log data includes, for example, output data for each driving scene as a set of time-series data. In the log data, the output data is basically compressed. In the log data, compression of the output data is not essential. The log acquisition unit 12 calculates the amount of data for each driving scene as the storage capacity required to store one event.

[0027] The impact degree acquisition unit 13 acquires the impact degree of the driving scene on the surroundings of the vehicle 1 based on, for example, the driving scene specified by the log acquisition unit 12. The impact degree means the possibility of impact, risk, or inconvenience on traffic safety for other vehicles. The impact degree is set in advance according to the driving scene in which the vehicle 1 may be placed. The impact degree may be set to a value that is larger as the impact degree is more important in the post-mortem verification of the calculation result of the automatic driving control. For example, the impact degree of turning right or left at an intersection may be larger than the impact degree of lane change. The impact degree of turning right or left at an intersection where a pedestrian or an oncoming vehicle is present may be larger than the impact degree of turning right or left at an intersection where a pedestrian or an oncoming vehicle is not present. The impact degree may be stored in the ECU 10 or may be acquired from outside the vehicle 1 via the communication unit 6.

[0028] The capacity shortage prediction unit 14 predicts a future capacity shortage in the storage unit 7 of the vehicle 1 based on the time of autonomous driving control to the destination of the vehicle 1 and road conditions. Capacity shortage in the storage unit 7 means a state in which the remaining capacity of the storage unit 7 becomes small, causing log data that cannot be stored in the storage unit 7, thereby hindering collection of log data. The capacity shortage prediction unit 14 predicts a future capacity shortage in the storage unit 7 of the vehicle 1, for example, by comparing the difference between the current remaining capacity of the storage unit 7 and an estimated value of the data amount of the log data with a predetermined capacity shortage threshold.

[0029] The approximate value of the data amount of the log data is an estimated value of the data amount of the log data used to predict a future capacity shortage of the storage unit 7 of the vehicle 1. The approximate value of the data amount of the log data is set in advance for each driving scene based on the data amount (storage capacity required to store one event) for each driving scene acquired in the past by a test or a simulation, for example. The approximate value of the data amount of the log data may be calculated or corrected in real time for each driving scene based on the data amount (storage capacity required to store one event) for each driving scene acquired in the past by the log acquisition unit 12. Note that, for log data of regular driving scenes acquired as a routine at a certain period, the approximate value of the data amount of the log data may be calculated according to the time of the automatic driving control to the destination and the average data amount of the log data.

[0030] The capacity shortage threshold is a threshold of the remaining capacity of the storage unit 7 for determining whether or not a future capacity shortage of the storage unit 7 of the vehicle 1 is predicted. The capacity shortage threshold may be a preset parameter or may be adjusted in real time. The capacity shortage threshold may be a different value based on the road conditions for each driving scene. The capacity shortage threshold may be adjusted according to the road type.

[0031] Depending on the road type, the estimated data amount of the log data tends to be large at an intersection and small if driving straight. The estimated data amount of the log data tends to be large if there is a lot of traffic, for example. The estimated data amount of the log data tends to be large when the log data includes multiple driving scenes, for example. The estimated data amount of the log data tends to be large on general roads other than expressways, compared to expressways. This is because expressways have a simple structure, a fixed driving direction, and a low possibility of a vehicle making a sudden change in behavior. Therefore, for example, in a road type in which the data amount of the log data tends to be larger on average than other road types, the capacity shortage threshold may be set to a larger value than other road types in accordance with this tendency.

[0032] The capacity shortage threshold may be adjusted based on the number of near-miss events. For example, the capacity shortage threshold may be set to a larger value in road conditions where near-miss events are likely to occur than in road conditions where near-miss events are unlikely to occur, because the amount of log data of non-stationary events is likely to increase in road conditions where near-miss events are likely to occur. Note that a near-miss event means an event that is not anticipated in advance for each driving scene. Specific examples of road conditions where near-miss events are likely to occur include branching points, merging points, narrow roads, etc. This is because narrow roads have a risk of pedestrians and the like running out into the road, and the distance between other traffic participants tends to be close.

[0033] The log storage unit 15 stores the log data in the storage unit 7 of the vehicle 1. For example, when a future capacity shortage of the storage unit 7 of the vehicle 1 is predicted, the log storage unit 15 stores the log data in the storage unit 7 of the vehicle 1 with a priority according to the impact of the driving scene associated with the log data. Note that, when a future capacity shortage of the storage unit 7 of the vehicle 1 is not predicted, the log storage unit 15 may store the log data in the storage unit 7 of the vehicle 1 regardless of the impact of the driving scene on the surroundings of the vehicle 1.

[0034] When a future capacity shortage of the storage unit 7 of the vehicle 1 is predicted, and an external storage unit is available via the communication unit 6, the log storage unit 15 may transmit log data to the external storage unit. In this case, for log data that was stored in the storage unit 7 of the vehicle 1, the log storage unit 15 may delete the transmitted log data from the storage unit 7 of the vehicle 1. For log data that was acquired by the log acquisition unit 12 and transmitted without being stored in the storage unit 7 of the vehicle 1, the log storage unit 15 may not store the log data in the storage unit 7 of the vehicle 1.

[0035] The log storage unit 15 may determine whether or not the external storage unit can be used based on the communication quality of the communication network N. The log storage unit 15 may use, for example, radio wave strength or a transmission frame rate obtained in association with the transmission of an image or the like, as the communication quality of the communication network N. The log storage unit 15 may also determine whether or not the external storage unit can be used based on the remaining capacity of the external storage unit.

[0036] An example of the external storage unit may be a server 50 for an autonomous driving system. The server 50 for an autonomous driving system is a server capable of communicating with the vehicle 1. The server 50 for an autonomous driving system includes a communication unit 51, a storage unit 52, and an analysis event DB 53. The communication unit 51 communicates various information with the vehicle 1 via, for example, a communication network N. The communication unit 51 is not particularly limited, and various known communication devices may be used.

[0037] The storage unit 52 stores log data to be used for post-verification of the calculation results of the autonomous driving control. The storage unit 52 is, for example, a recording device such as an HDD provided in the autonomous driving system server 50. The storage unit 52 has a log storage area for storing the log data. That is, the autonomous driving system server 50 functions as an external storage unit that stores the log data transmitted from the vehicle 1 via the communication network N. The log storage area has a finite capacity.

[0038] The analysis event DB 53 is a storage device that stores analysis information. The analysis event DB 53 is formed in a recording device such as an HDD provided in the autonomous driving system server 50. The analysis information includes information on the presence or absence of analysis that performs post-mortem verification of the computation results of the autonomous driving control using log data, and information on the analysis results.

[0039] Another example of the external storage unit may be another vehicle 20. The other vehicle 20 is a vehicle in the vicinity of the vehicle 1 that exists within a communication range of the vehicle 1. The other vehicle 20 includes an ECU 21, a communication unit 22, and a storage unit 23. The ECU 21 is an electronic control unit that executes processes related to communication with the vehicle 1 and storage of log data in the other vehicle 20. The communication unit 22 communicates various information with the vehicle 1 via, for example, the communication network N. The communication unit 22 is not particularly limited, and various known communication devices may be used.

[0040] The storage unit 23 stores log data to be used for post-verification of the calculation results of the autonomous driving control. The storage unit 23 is, for example, a recording device such as an HDD provided in the autonomous driving system server 50. The storage unit 23 has a log storage area for storing the log data. That is, the other vehicle 20 functions as an external storage unit that stores the log data transmitted from the vehicle 1 via the communication network N. The log storage area has a finite capacity.

[0041] When the log storage unit 15 can determine, based on the analysis information in the analyzed event DB 53, that the log data acquired by the log acquisition unit 12 is an analyzed event using log data already stored in the autonomous driving system server 50 or the other vehicle 20, the log storage unit 15 does not need to store the log data acquired by the log acquisition unit 12 in the storage unit 7 of the vehicle 1. In this case, the log storage unit 15 may store, in the storage unit 7, information specific to the vehicle 1 among the log data acquired by the log acquisition unit 12.

[0042] If the log memory unit 15 can determine, based on the analysis information in the analyzed event DB 53, that the log data acquired by the log acquisition unit 12 is an unanalyzed event due to log data already stored in the server 50 for the autonomous driving system or another vehicle 20, it may store all of the log data acquired by the log acquisition unit 12 in the memory unit 7 or an external memory unit.

[0043] The information specific to the vehicle 1 may include, for example, information on the specifications of the vehicle 1, a calculation result by the machine learning model 11a, and a control instruction value for the actuator 5. The information specific to the vehicle 1 may include differential data between log data already stored in the autonomous driving system server 50 or the other vehicle 20 and log data acquired by the log acquisition unit 12. The information specific to the vehicle 1 may include, among the detection results of the sensor 2, data based on the detection results of an on-board sensor of a type that the other vehicle 20 does not have.

[0044] The log storage unit 15 may use a storage device of an information processing terminal carried by an occupant of the vehicle 1 as an external storage unit to store the log data acquired by the log acquisition unit 12 for backup.

[0045] Next, the processing of the ECU 10 will be described with reference to Fig. 3. Fig. 3 is a flowchart showing an example of a log storage process of the ECU. The flowchart shown in Fig. 3 is repeatedly executed at predetermined intervals, for example, every time the autonomous driving processing unit 11 calculates a certain amount of output data during autonomous driving of the vehicle 1.

[0046] 3, the ECU 10 acquires log data by the log acquisition unit 12 in step S11. The autonomous driving processing unit 11 calculates output data related to autonomous driving control based on the detection result of the sensor 2 of the vehicle 1, for example, using a machine learning model 11a. The log acquisition unit 12 identifies the driving scene when the output data was calculated based on the detection result of the sensor 2 of the vehicle 1, map information, and the vehicle position. The log acquisition unit 12 associates the driving scene for which the output data was calculated with the output data and acquires the output data as log data.

[0047] In S12, the ECU 10 acquires the impact degree by the impact degree acquisition unit 13. The impact degree acquisition unit 13 acquires, for example, a preset impact degree that the driving scene specified by the log acquisition unit 12 has on the surroundings of the vehicle 1.

[0048] In S13, the ECU 10 predicts a future capacity shortage in the memory unit 7 of the vehicle 1 by the capacity shortage prediction unit 14. The capacity shortage prediction unit 14 predicts a future capacity shortage in the memory unit 7 of the vehicle 1 based on the time of the automatic driving control to the destination of the vehicle 1 and the road conditions.

[0049] In S14, the ECU 10 determines whether or not a future capacity shortage of the storage unit 7 of the vehicle 1 is predicted by the log storage unit 15. The log storage unit 15 determines whether or not a future capacity shortage of the storage unit 7 of the vehicle 1 is predicted based on, for example, a comparison result between the future capacity of the storage unit 7 of the vehicle 1 estimated by the capacity shortage prediction unit 14 and a predetermined capacity threshold value.

[0050] If it is determined that future capacity shortage of the storage unit 7 of the vehicle 1 is not predicted (S14: NO), the ECU 10 proceeds to processing of S15. In S15, the ECU 10 causes the log storage unit 15 to store the log data in the storage unit 7 of the vehicle 1. The log storage unit 15 stores the log data in the storage unit 7 of the vehicle 1, for example, regardless of the degree of influence that the driving scene has on the surroundings of the vehicle 1. Thereafter, the ECU 10 ends the processing of FIG. 3.

[0051] On the other hand, when it is determined that the storage unit 7 of the vehicle 1 is predicted to be insufficient in the future (S14: YES), the ECU 10 proceeds to processing in S16. In S16, the ECU 10 determines whether or not communication with an external storage unit is possible using the log storage unit 15. The log storage unit 15 determines whether or not communication with the autonomous driving system server 50 or other vehicles 20 around the vehicle 1 is possible based on, for example, communication quality information of the communication network N.

[0052] When it is determined that communication with the external storage unit is possible (S16: YES), the ECU 10 proceeds to the process of S17. In S17, the ECU 10 transmits the log data to the external storage unit by the log storage unit 15. The log storage unit 15 transmits, for example, the log data acquired by the log acquisition unit 12 and not stored in the storage unit 7 of the vehicle 1, or the log data stored in the storage unit 7 of the vehicle 1, to the autonomous driving system server 50 or the other vehicle 20. Next, in S18, the ECU 10 deletes the transmitted log data from the storage unit 7 of the vehicle 1 by the log storage unit 15. For example, the log storage unit 15 deletes the log data transmitted in S17 from the storage unit 7 of the vehicle 1, for the log data stored in the storage unit 7 of the vehicle 1. For the log data acquired by the log acquisition unit 12 and not stored in the storage unit 7 of the vehicle 1, the log storage unit 15 can perform the process of S19 of not storing the log data in the storage unit 7 of the vehicle 1, instead of deleting the log data from the storage unit 7 of the vehicle 1. Then, the process of FIG. 3 is terminated.

[0053] On the other hand, when it is determined that communication with the external storage unit is not possible (S16: NO), the ECU 10 proceeds to processing of S19. In S19, the ECU 10 causes the log storage unit 15 to store the log data in the storage unit 7 of the vehicle 1 with a priority according to the impact level. The log storage unit 15 stores the log data in the storage unit 7 of the vehicle 1 with a priority according to the impact level of the driving scene associated with the log data. Thereafter, the processing of FIG. 3 is terminated.

[0054] Here, a trained model generated by machine learning, such as the machine learning model 11a, may not necessarily be capable of realizing appropriate autonomous driving control in all driving scenes. As an example, when the surrounding environment of the vehicle 1 changes, the inference result by the trained model may change. For example, whether or not the calculation result of the autonomous driving control using the trained model can be said to be appropriate may change depending on the traffic environment such as the weather, time of day, and traffic volume. Therefore, in order to appropriately operate the autonomous driving control using the trained model, it is considered that a mechanism that enables the autonomous driving control to be verified after the fact is useful. For this reason, it is considered to collect log data used for the after-fact verification of the calculation result of the autonomous driving control. However, the log data of the calculation result of the autonomous driving control may become enormous.

[0055] In this regard, as described above, in the autonomous driving system 100, log data in which output data and driving scenes are associated with each other is stored in the storage unit 7 of the vehicle 1 with a priority according to the impact of the driving scene. For example, useful log data can be effectively collected by preferentially storing log data corresponding to driving scenes that have a large impact on the surroundings of the vehicle 1 in the storage unit 7 of the vehicle 1. Therefore, according to the autonomous driving system 100, it is possible to ensure the storage of log data used for post-mortem verification of the calculation results of the autonomous driving control.

[0056] The autonomous driving system 100 includes a capacity shortage prediction unit 14 that predicts a future capacity shortage of the storage unit 7 of the vehicle 1 based on the time of autonomous driving control to the destination of the vehicle 1 and road conditions. When a capacity shortage of the storage unit 7 of the vehicle 1 is predicted, the log storage unit 15 stores log data in the storage unit 7 of the vehicle 1 with a priority according to the degree of impact. For example, when the storage unit 7 of the vehicle 1 runs out of capacity, there is a possibility that the recording of log data corresponding to a driving scene that has a large impact on the surroundings of the vehicle 1 will fail. Therefore, when a capacity shortage of the storage unit 7 of the vehicle 1 is predicted, the log data corresponding to the driving scene that has a large impact on the surroundings of the vehicle 1 is stored in the storage unit 7 of the vehicle 1 with a priority according to the degree of impact, thereby ensuring the storage of log data corresponding to the driving scene that has a large impact on the surroundings of the vehicle 1.

[0057] The autonomous driving system 100 includes an external storage unit (an autonomous driving system server 50 or another vehicle 20) that stores log data transmitted from the vehicle 1 via the communication network N, and a communication unit 6 that is provided in the vehicle 1 and communicates with the external storage unit. The log storage unit 15 transmits the log data to the external storage unit and deletes the transmitted log data from the storage unit 7 of the vehicle 1. By storing the log data in the external storage unit and deleting the log data from the storage unit 7 of the vehicle 1 in this manner, it is possible to ensure the storage of the log data while preventing a shortage of capacity in the storage unit 7 of the vehicle 1.

[0058] [Variations] Although various exemplary embodiments have been described above, various omissions, substitutions, and modifications may be made without being limited to the above-described exemplary embodiments.

[0059] In the above embodiment, when a capacity shortage in the storage unit 7 of the vehicle 1 is predicted, the log storage unit 15 stores the log data in the storage unit 7 of the vehicle 1 with a priority according to the degree of impact, but this example is not limiting. The prediction of a capacity shortage in the storage unit 7 of the vehicle 1 may be omitted. The log storage unit 15 may store the log data in the storage unit 7 of the vehicle 1 with a priority according to the degree of impact, regardless of the remaining capacity of the storage unit 7 of the vehicle 1.

[0060] In the above embodiment, the log storage unit 15 transmits the log data to the external storage unit and deletes the transmitted log data from the storage unit 7 of the vehicle 1, but this is not limiting. It is not essential for the log storage unit 15 to use an external storage unit.

[0061] In the above embodiment, the autonomous driving system 100 according to the embodiment calculates the output data related to autonomous driving control using the machine learning model 11a, but the use of a machine learning model is not essential.

[0062] The capacity shortage prediction unit 14 may determine whether or not a capacity shortage is predicted when a route for automatic driving is set. When a capacity shortage is predicted, the capacity shortage prediction unit 14 may notify the occupant of the vehicle 1 using the HMI of the vehicle 1 or the like. The capacity shortage prediction unit 14 may notify the occupant of the vehicle 1 to stop the vehicle 1 for a certain period of time using the HMI of the vehicle 1 or the like. The capacity shortage prediction unit 14 may notify the occupant of the vehicle 1 to manually drive the vehicle 1 for a certain period of time using the HMI of the vehicle 1 or the like. The certain period of time may be, for example, a period of time until a capacity shortage in the storage unit 7 is no longer predicted by the log storage unit 15 performing the transmission and deletion process of the log data to the external storage unit (S17 and S18 above). The log storage unit 15 may be configured to be capable of transmitting the log data to the external storage unit by manual operation using the HMI of the vehicle 1 or the like.

[0063] The log acquisition unit 12 may acquire log data only when a predetermined event occurs. [Explanation of symbols]

[0064] 1...vehicle, 6...communication unit, 7...memory unit, 12...log acquisition unit, 13...impact acquisition unit, 14...capacity shortage prediction unit, 15...log memory unit, 20...other vehicles (external memory unit), 50...server for autonomous driving system (external memory unit), 100...autonomous driving system.

Claims

1. An autonomous driving system that stores output data related to autonomous driving control calculated based on detection results of an on-board sensor of a vehicle in a storage unit of the vehicle, a log acquisition unit that associates the driving scene for which the output data was calculated with the output data and acquires the output data as log data; an impact degree acquisition unit that acquires a preset impact degree that the driving scene has on the surroundings of the vehicle based on the driving scene; An autonomous driving system comprising: a log memory unit that stores the log data in the memory unit of the vehicle with a priority according to the impact degree of the driving scene associated with the log data.

2. A capacity shortage prediction unit that predicts a future capacity shortage of a storage unit of the vehicle based on a time taken for the autonomous driving control to reach a destination of the vehicle and road conditions, The autonomous driving system according to claim 1 , wherein the log storage unit stores the log data in the storage unit of the vehicle with a priority according to the degree of impact when a capacity shortage of the storage unit of the vehicle is predicted.

3. an external storage unit that stores the log data transmitted from the vehicle via a network; a communication unit provided in the vehicle and configured to communicate with the external storage unit; The autonomous driving system according to claim 1 , wherein the log storage unit transmits the log data to the external storage unit and deletes the transmitted log data from the storage unit of the vehicle.

Images