Data accumulation apparatus, data accumulation method, and data accumulation program

Abstract

To eliminate the need to perform detailed condition settings for data accumulation when accumulating learning data.SOLUTION: A data accumulation apparatus that accumulates learning data of driving operations of a vehicle 2 in a database 322 includes: a comparison unit 332 that compares manual driving operation data acquired when the vehicle is manually driven under predetermined environmental conditions with comparison driving operation data including at least one of system driving operation data generated on an assumption that the vehicle is autonomously driven by a system of the vehicle under the same environmental conditions as the predetermined environmental conditions, and similar driving operation data obtained when the vehicle is driven under environmental conditions similar to the predetermined environmental conditions; and an accumulation determination unit 333 that determines whether to accumulate the manual driving operation data as learning data in the database as it is on the basis of a comparison result between the manual driving operation data and the comparison driving operation data by the comparison unit.SELECTED DRAWING: Figure 7

Description

[Technical Field]

[0001] The present disclosure relates to a data storage device, a data storage method, and a data storage program. Regarding. [Background technology]

[0002] It is known that driving operations performed by a driver are stored in a storage unit (Patent Documents 1 to 3). In particular, Patent Document 1 proposes a technology for a driving assistance device that learns the driving operations of a driver to reflect the driver's driving characteristics in automatic driving of a vehicle, in which driving operations in driving situations that are determined to be inappropriate for learning are excluded from the learning target. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2019-127207 [Patent Document 2] Japanese Patent Publication No. 2022-129400 [Patent Document 3] Japanese Patent Publication No. 2023-61084 Summary of the Invention [Problem to be solved by the invention]

[0004] In Patent Document 1, whether a driving situation is inappropriate for learning is determined based on whether various conditions are met, such as whether the vehicle's acceleration, deceleration, or yaw rate is equal to or greater than a threshold. Therefore, it is necessary to set all the conditions under which a driving situation is appropriate for learning in advance. As such, the device disclosed in Patent Document 1 requires a lot of work to set the conditions for determining whether a driving situation is appropriate for learning in detail.

[0005] In view of the above-mentioned problems, an object of the present disclosure is to eliminate the need to set detailed conditions for accumulating learning data. [Means for solving the problem]

[0006] The gist of the present disclosure is as follows.

[0007] (1) A data storage device that stores learning data for vehicle driving operations in a database, a comparison unit that compares manual driving operation data acquired when the vehicle is being manually driven under predetermined environmental conditions with comparative driving operation data including at least one of system driving operation data generated under the assumption that the vehicle is being driven autonomously by a system of the vehicle under the same environmental conditions as the predetermined environmental conditions and similar driving operation data when the vehicle is driven under environmental conditions similar to the predetermined environmental conditions; A data storage device having an accumulation decision unit that decides whether to store the manual driving operation data as is in the database as learning data based on the comparison result between the manual driving operation data and the comparison driving operation data by the comparison unit. (2) The data storage device described in (1) above, wherein the learning data of the driving operation is used to construct a machine learning model through predetermined learning, and autonomous driving by the vehicle system is performed using a model different from the machine learning model constructed through the predetermined learning. (3) a classification unit that classifies environmental conditions when the vehicle is traveling; a determination unit that determines whether or not similar driving operation data when the vehicle was driven under environmental conditions similar to the environmental conditions under which the vehicle traveled is stored in a storage unit based on the classification result by the classification unit, The data storage device according to (1) or (2) above, wherein the comparison unit uses the similar driving operation data as the comparison driving operation data when the determination unit determines that similar driving operation data under similar environmental conditions is stored in the memory unit. (4) The data storage device described in (3) above, wherein the classification unit classifies the environmental conditions under which the vehicle traveled based on at least one of geographical conditions, temporal conditions, weather conditions, and conditions related to traffic participants around the vehicle. (5) A data storage device described in any one of (1) to (4) above, wherein the storage decision unit further decides whether to modify at least a portion of the manual driving operation data that has been determined not to be stored in the database as the learning data and then store the modified manual driving operation data in the database as the learning data. (6) A data storage unit that stores data in a database based on the decision of the storage decision unit is further provided, The data storage device described in (5) above, wherein when the storage decision unit decides to modify at least a portion of the manual driving operation data and store it as the learning data, the data storage unit modifies at least a portion of the manual driving operation data to the corresponding comparison driving operation data and stores it in the database as the learning data. (7) The data storage device described in (5) or (6) above, wherein the storage decision unit determines whether to modify a portion of the manual driving operation data and store it in the database as learning data for each environmental condition under which the vehicle is traveling. (8) A standard setting unit sets an accumulation standard based on the magnitude of an average deviation between the manual driving operation data and the comparative driving operation data for each environmental condition, The data storage device described in any one of (1) to (7) above, wherein the storage decision unit decides whether or not to store the manual driving operation data as is in the database as learning data based on the storage criteria corresponding to the environmental conditions under which the vehicle was traveling. (9) A driver state determination unit is further provided that determines the state of the driver based on an output of a sensor that detects the state of the driver, The data storage device described in any one of (1) to (8) above, wherein the storage decision unit decides whether or not to store the manual driving operation data as is in the database as the learning data based on the driver's state determined by the driver state determination unit. (10) The vehicle further includes a specifying unit that specifies an area where the driver's state is likely to become unsuitable for driving based on a plurality of determination results by the driver state determining unit, The data storage device according to (9) above, wherein the storage decision unit decides whether or not to store the manual driving operation data as is in the database as the learning data based on different criteria for each area. (11) A data storage method executed by a processor for storing learning data of a vehicle driving operation in a database, comprising: comparing manual driving operation data acquired when the vehicle is being manually driven under predetermined environmental conditions with comparative driving operation data including at least one of system driving operation data generated under the assumption that the vehicle is being driven autonomously by a system of the vehicle under the same environmental conditions as the predetermined environmental conditions and similar driving operation data when the vehicle is driven under environmental conditions similar to the predetermined environmental conditions; and determining whether or not to store the manual driving operation data as is in the database as learning data based on a comparison result between the manual driving operation data and the comparative driving operation data. (12) A data storage program for storing learning data for vehicle driving operations in a database, comprising: comparing manual driving operation data acquired when the vehicle is being manually driven under predetermined environmental conditions with comparative driving operation data including at least one of system driving operation data generated under the assumption that the vehicle is being driven autonomously by a system of the vehicle under the same environmental conditions as the predetermined environmental conditions and similar driving operation data when the vehicle is driven under environmental conditions similar to the predetermined environmental conditions; determining whether to store the manual driving operation data as is in the database as the learning data based on a comparison result between the manual driving operation data and the comparative driving operation data; A data storage program that causes a computer to execute the following. [Effects of the Invention]

[0008] According to the present disclosure, when accumulating learning data, there is no need to set detailed conditions for accumulating data. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is a schematic configuration diagram of a data storage system according to the first embodiment. [Figure 2] FIG. 2 is a schematic configuration diagram of a vehicle. [Figure 3] FIG. 3 is a functional block diagram of the vehicle processor. [Figure 4] FIG. 4 is a schematic diagram of the server. [Figure 5] FIG. 5 is a functional block diagram of the server processor. [Figure 6] FIG. 6 is a diagram showing the change in the speed of a vehicle when the vehicle is traveling on a certain road. [Figure 7] FIG. 7 is a flowchart showing the flow of the data accumulation process. [Figure 8] FIG. 8 is a functional block diagram of a server processor according to the second embodiment. [Figure 9] FIG. 9 is a flowchart showing the flow of data accumulation processing according to the second embodiment. [Figure 10] FIG. 10 is a functional block diagram of a server processor according to the third embodiment. [Figure 11] FIG. 11 is a flowchart showing the flow of data accumulation processing according to the third embodiment. [Figure 12] FIG. 12 is a functional block diagram of a server processor according to a modification of the third embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, the embodiments will be described in detail with reference to the drawings. In the following description, like components are designated by like reference numerals.

[0011] First embodiment <Data storage system configuration> First, a data accumulation system 1 according to the first embodiment will be described with reference to Fig. 1. Fig. 1 is a schematic configuration diagram of the data accumulation system 1 according to the first embodiment. The data accumulation system 1 accumulates learning data for vehicle driving operations in a database.

[0012] 1, the data accumulation system 1 includes a plurality of vehicles 2 that can communicate with each other, and a server 3. Each of the plurality of vehicles 2 and the server 3 are configured to be able to communicate with each other via a communication network 4 that is configured, for example, by optical communication lines, and a wireless base station 5 that is connected to the communication network 4 via a gateway (not shown). For communication between the vehicles 2 and the wireless base station 5, various wide-area wireless communications with long communication distances can be used, such as communications that comply with any communication standard such as 4G, LTE, 5G, or WiMAX established by 3GPP or IEEE.

[0013] In the data storage system 1 according to this embodiment, the server 3 stores learning data based on data received from a plurality of vehicles 2. The server 3 also performs learning of a machine learning model related to autonomous driving of the vehicle 2 based on the stored learning data, and transmits the completed learning machine learning model to the vehicle 2. The vehicle 2 then performs autonomous driving of the vehicle 2 using the machine learning model transmitted in this manner. In this embodiment, the machine learning model is, for example, a machine learning model obtained by end-to-end learning, and therefore outputs values ​​of control parameters for controlling the vehicle actuators 26 when environmental data is input. However, the machine learning model may also be a model trained by a learning method other than end-to-end learning.

[0014] <Vehicle configuration> Next, the configuration of the vehicle 2 will be described with reference to Figures 2 and 3. The vehicle 2 can be operated by a driver and driven manually, or can be driven autonomously by the ECU of the vehicle 2. The vehicle 2 periodically transmits manual driving operation data when being manually driven by the driver and environmental data when this manual driving operation data was obtained to the server 3. The vehicle 2 also performs autonomous driving using a machine learning model transmitted from the server 3.

[0015] 2 is a schematic configuration diagram of a vehicle 2. In this embodiment, the vehicle 2 has a surrounding environment information sensor 21, a host vehicle environment information sensor 22, a driver operation sensor 23, a driver state sensor 24, an exterior communication module 25, a vehicle actuator 26, and an electronic control unit (hereinafter referred to as "ECU") 27. These components of the vehicle 2 are communicatively connected to one another via an in-vehicle network 28 that complies with a standard such as CAN (Controller Area Network), or are directly connected to one another via signal lines.

[0016] The surrounding environment information sensor 21 is a sensor that detects the environment or situation around the vehicle 2 and generates surrounding environment data that represents the environment or situation around the vehicle 2. The surrounding environment information sensor 21 includes, for example, an exterior camera 211 and a distance measurement sensor 212. The exterior camera 211 captures images of the surroundings of the vehicle 2, and in this embodiment, captures images of the area ahead of the vehicle 2. The distance measurement sensor 212 measures the distance to an object that exists around the vehicle 2, which in this embodiment is an object that exists ahead of the vehicle 2. The distance measurement sensor 212 is, for example, a radar such as a millimeter-wave radar, a LiDAR, or a sonar.

[0017] The host vehicle environment information sensor 22 is a sensor that detects the situation of the vehicle 2 and generates vehicle environment data that represents the environment of the vehicle 2 (hereinafter, the surrounding environment data and the vehicle environment data will be collectively referred to as environment data). The host vehicle environment information sensor 22 includes, for example, a positioning sensor 221 and a running state sensor 222. The positioning sensor 221 measures the host position of the vehicle 2. The positioning sensor 221 is, for example, a GNSS receiver. The running state sensor 222 detects the running state of the vehicle 2. The running state sensor 222 detects, for example, the speed, acceleration, and rate of change of the yaw angle when turning (yaw rate) of the vehicle 2.

[0018] The driver operation sensor 23 is a sensor that detects the operation status of the vehicle 2's operating devices (e.g., accelerator pedal, brake pedal, and steering wheel) by the driver and generates manual driving operation data that represents the operation status of the operating devices. The driver operation sensor 23 includes, for example, an accelerator sensor 231 that detects the amount of depression of the accelerator pedal, a brake sensor 232 that detects the amount of depression of the brake pedal, and a steering sensor 233 that detects the steering angle of the steering wheel.

[0019] The manual driving operation data is data representing the accelerator pedal depression amount, the brake pedal depression amount, and the steering angle of the steering wheel. Alternatively, the manual driving operation data may be data representing the state of the vehicle 2 as a result of operating each operating device. Therefore, the manual driving operation data may be data representing changes in the speed or acceleration of the vehicle 2, changes in turning radius, etc., resulting from the operation of each operating device. In this case, the manual driving operation data is generated in the ECU 27 based on the operation state of the operating devices of the vehicle 2 detected by the driver operation sensor 23.

[0020] The driver state sensor 24 is a sensor that detects the state of the driver of the vehicle 2 and generates driver state data that represents the state of the driver. The driver state sensor 24 includes, for example, a driver monitor camera 241. The driver monitor camera 241 captures an image of the driver of the vehicle 2.

[0021] The surrounding environment information sensor 21, the vehicle environment information sensor 22, the driver operation sensor 23, and the driver state sensor 24 output the generated data to the ECU 27 via the in-vehicle network 28 at predetermined intervals.

[0022] The exterior-vehicle communication module 25 communicates with devices outside the vehicle. The exterior-vehicle communication module 25 is a device that performs wireless communication with the wireless base station 5 in accordance with a predetermined mobile communication standard. For example, the exterior-vehicle communication module 25 receives data generated by the various sensors 21 to 24 of the vehicle 2 from these sensors and transmits the data to the server 3 via the wireless base station 5. The exterior-vehicle communication module 25 also receives a machine learning model generated or updated by the server 3 via the wireless base station 5 and transmits the model to the ECU 27.

[0023] Vehicle actuator 26 is an actuator used to control the operation of vehicle 2. Specifically, vehicle actuator 26 includes, for example, a drive actuator 261 that controls a prime mover (an internal combustion engine or an electric motor) for driving vehicle 2, a braking actuator 262 that controls a brake that brakes vehicle 2, and a steering actuator 263 that controls steering of vehicle 2. Vehicle actuator 26 controls the operation of vehicle 2 in accordance with a control signal received from ECU 27.

[0024] The ECU 27 transmits data generated by the various sensors 21 to 24 of the vehicle 2 to the server 3. The ECU 27 also controls the operation of the vehicle 2 using a machine learning model. The ECU 27 includes a communication interface 271, a vehicle storage unit 272, and a vehicle processor 273. The communication interface 271 includes a circuit for connecting the ECU 27 to an in-vehicle network 28 or the like. The vehicle storage unit 272 stores data. The vehicle storage unit 272 includes, for example, at least one of a volatile semiconductor memory, a non-volatile semiconductor memory, a hard disk drive (HDD), and a solid state drive (SSD). The vehicle storage unit 272 stores a computer program, for example, a machine learning model, to be executed by the vehicle processor 273. The vehicle processor 273 includes one or more central processing units (CPUs) and their peripheral circuits. The vehicle processor 273 executes the computer program stored in the vehicle storage unit 272.

[0025] Fig. 3 is a functional block diagram of vehicle processor 273. As shown in Fig. 3, vehicle processor 273 includes a data transmission unit 275, a driving control unit 276, and a model update unit 277. Each of these units included in vehicle processor 273 is a functional module realized by a computer program running on vehicle processor 273, for example.

[0026] The data transmission unit 275 transmits data generated by the various sensors 21 to 24 of the vehicle 2 to the server 3 via the exterior communication module 25. The data transmission unit 275 may transmit the data generated by the various sensors 21 to 24 of the vehicle 2 sequentially to the server 3, or may temporarily store the data in the vehicle storage unit 272 and then transmit the data collectively to the server 3. Furthermore, the data transmission unit 275 may transmit all of the data generated by the various sensors 21 to 24 to the server 3, or may transmit only a portion of the data to the server.

[0027] The driving control unit 276 controls the driving of the vehicle 2. The driving control unit 276 may manually control the driving of the vehicle 2 in accordance with the operation of the driver of the vehicle 2, or may autonomously control the driving of the vehicle 2 in accordance with a machine learning model. When manually controlling the driving of the vehicle 2, the driving control unit 276 controls the vehicle actuators 26 based on manual driving operation data generated by the driver operation sensor 23. In particular, the driving control unit 276 controls the vehicle actuators 26 so that the vehicle 2 is operated in accordance with the operation status of the controls of the vehicle 2 by the driver. On the other hand, when autonomously controlling the driving of the vehicle 2, the driving control unit 276 inputs environmental data generated by the surrounding environment information sensor 21, the host vehicle environment information sensor 22, etc. into a machine learning model and controls the vehicle actuators 26 in accordance with the output of the machine learning model. Therefore, the driving control unit 276 controls the driving of the vehicle 2 based on data generated by the other sensors 21 and 22, rather than based on the manual driving operation data generated by the driver operation sensor 23.

[0028] The model update unit 277 updates the machine learning model used for the autonomous driving of the vehicle 2 to the model received from the server 3. The model update unit 277 may update parameters (weights, etc.) used in the machine learning model, and may also update the configuration of the machine learning model (for example, the number of layers or nodes, etc.).

[0029] <Server configuration> Next, the configuration of the server 3 will be described with reference to Figures 4 and 5. The server 3 receives manual driving operation data, environmental data, and driver status data when the vehicle 2 is manually driven from multiple vehicles 2 with which it can communicate via the communication network 4. The server 3 also stores at least a portion of the data received in this manner as learning data to be used for learning a machine learning model. In addition, once a certain amount of learning data has been accumulated, the server 3 constructs a machine learning model by predetermined machine learning based on the accumulated learning data. The construction of the machine learning model by machine learning may be performed automatically by the server 3, or may be performed manually using the learning data stored in the server 3. Furthermore, once the machine learning model has been constructed by machine learning, the server 3 transmits the trained machine learning model to each vehicle 2 via the communication network 4.

[0030] 4 is a schematic configuration diagram of the server 3. In this embodiment, the server 3 includes an external communication module 31, a server storage unit 32, and a server processor 33. The external communication module 31 and the server storage unit 32 are connected to the server processor 33 via signal lines. The server 3 may further include input devices such as a keyboard and a mouse, and output devices such as a display and a speaker. The server 3 may also be composed of multiple computers.

[0031] The external communication module 31 communicates with devices outside the server 3. The external communication module 31 has an interface circuit for connecting the server 3 to the communication network 4. The external communication module 31 is configured to be able to communicate with the vehicle 2 via the communication network 4 and a wireless base station 5.

[0032] The server storage unit 32 stores data. The server storage unit 32 includes, for example, an HDD, an SSD, or an optical recording medium. The server storage unit 32 may also include a volatile semiconductor memory (for example, a RAM), a non-volatile semiconductor memory (for example, a ROM), or the like. The server storage unit 32 stores a computer program executed by the server processor 33. In addition, the server storage unit 32 includes a database that stores various types of data. In this embodiment, the server storage unit 32 includes a raw database 321 that stores data received from the vehicle 2, and a training database 322 that stores training data used to train the machine learning model.

[0033] The server processor 33 has one or more CPUs and their peripheral circuits. The server processor 33 may further have other arithmetic circuits such as a logic unit, a numerical calculation unit, or a graphics processing unit. The server processor 33 executes various processes based on computer programs stored in the server storage unit 32. In this embodiment, the server processor 33 functions as a data storage device that accumulates learning data for the driving operation of the vehicle 2 in a database. In this embodiment, the server processor 33 also executes a data storage method for accumulating learning data for the driving operation of the vehicle in a database.

[0034] 5 is a functional block diagram of the server processor 33. As shown in FIG. 5, the server processor 33 includes a system operation generation unit 331, a comparison unit 332, an accumulation determination unit 333, a data accumulation unit 334, a standard setting unit 335, and a classification unit 336. Each of these units included in the server processor 33 is a functional module realized by, for example, a computer program running on the server processor 33. The computer program includes a data accumulation program that accumulates learning data for vehicle driving operations in a database.

[0035] The system operation generation unit 331 generates driving operation data (system driving operation data) assuming that the vehicle 2 is driven autonomously by the driving control model under any environmental condition. Therefore, when the system operation generation unit 331 receives environmental data generated by the vehicle 2 when the vehicle 2 is driven under any environmental condition, it outputs driving operation data assuming that the vehicle 2 is driven autonomously by the driving control model under that environmental condition.

[0036] The system operation generation unit 331 receives the environmental data transmitted from the vehicle 2 and stored in the raw database 321 of the server storage unit 32. The system operation generation unit 331 then outputs to the comparison unit 332 system driving operation data that would be generated if the vehicle 2 were driven autonomously using the driving control model under environmental conditions corresponding to the input environmental data.

[0037] Here, the environmental conditions refer to conditions related to the driving environment of the vehicle 2 when the vehicle 2 is being driven. Therefore, the environmental conditions include, for example, geographical conditions related to the road on which the vehicle 2 is traveling, time conditions, weather conditions, and conditions related to traffic participants around the vehicle 2.

[0038] The geographical conditions include the position of the road on which the vehicle 2 travels, the topography (such as the curvature, gradient, and number of lanes of the road), and the shapes of objects around the road (such as buildings and trees). The geographical conditions are identified, for example, based on surrounding environment data from the surrounding environment information sensor 21. The geographical conditions are also identified, for example, based on self-position data generated by the positioning sensor 221 and map data stored in the server storage unit 32.

[0039] The time conditions include, for example, the season and the time of day, and the brightness around the vehicle 2 changes depending on these conditions. The time conditions are identified based on time information measured by the ECU 27. The weather conditions include, for example, the weather around the vehicle 2 and are classified as sunny, cloudy, rainy, snowy, etc. The weather conditions are identified based on, for example, the surrounding environment data from the surrounding environment information sensor 21. The weather conditions are identified based on the vehicle's own position data generated by the positioning sensor 221 and the weather information for each region stored in the server storage unit 32. The traffic participants around the vehicle 2 refer to moving objects that affect the traveling of the vehicle 2, such as other vehicles and pedestrians around the vehicle 2, and the conditions related to the traffic participants around the vehicle 2 include the density of such traffic participants and their relative positional relationships. The conditions related to the traffic participants around the vehicle 2 are identified based on, for example, the surrounding environment data from the surrounding environment information sensor 21.

[0040] Furthermore, in this embodiment, the driving control model used by the system operation generation unit 331 is a different model from the machine learning model used for autonomous driving of the vehicle 2. Therefore, the hypothetical autonomous driving performed by the system operation generation unit 331 is performed using a model different from the machine learning model constructed by machine learning using training data. For example, while the machine learning model is a model obtained by end-to-end learning, the driving control model may be a model trained by a learning method other than end-to-end learning, or may be a rule-based model generated without training by a learning method. In either case, the driving control model is a model that does not necessarily produce the same output as the machine learning model when the same input is received. As a result, training data is accumulated based on a model different from the machine learning model.

[0041] The comparison unit 332 compares the manual driving operation data acquired when the vehicle 2 is manually driven under arbitrary environmental conditions with the system driving operation data (comparison driving operation data) generated under the assumption that the vehicle 2 is being driven autonomously by the system (using a driving control model) under the same environmental conditions as the arbitrary environmental conditions. In particular, in this embodiment, the degree of deviation between the manual driving by the driver and the autonomous driving by the system is calculated based on the manual driving operation data and the system driving operation data.

[0042] The comparison unit 332 receives the manual driving operation data transmitted from the vehicle 2 and stored in the raw database 321 of the server storage unit 32, and the system driving operation data output from the system operation generation unit 331 after inputting the corresponding environmental data. The comparison unit 332 then outputs the result of the comparison between the manual driving operation data and the system driving operation data (e.g., the degree of deviation) to the accumulation determination unit 333. The comparison unit 332 also outputs the input manual driving operation data and system driving operation data to the accumulation determination unit 333 as is.

[0043] Specifically, the comparison unit 332 calculates the deviation as the average or integrated value of the difference between the control parameters represented by the manual driving operation data and the control parameters represented by the system driving operation data when the vehicle 2 is located at the same position in its traveling direction. For example, the deviation is calculated as the average or integrated value of the difference between the accelerator pedal depression amount in the manual driving operation data and the accelerator pedal depression amount in the system driving operation data. Alternatively, if the driving operation data includes the speed of the vehicle 2, the deviation is calculated as the average or integrated value of the difference between the speed of the vehicle 2 in the manual driving operation data and the speed of the vehicle 2 in the system driving operation data. Note that the comparison unit 332 may calculate the deviation by a method different from the above-described method as long as it can calculate the deviation between the manual driving by the driver and the autonomous driving by the system. Furthermore, the comparison unit 332 may calculate the value of an index other than the deviation as long as the index represents the difference between the manual driving operation data and the system driving operation data.

[0044] The accumulation determination unit 333 determines whether to accumulate the manual driving operation data as is as learning data in the learning database 322 of the server storage unit 32 based on the comparison result between the manual driving operation data and the system driving operation data by the comparison unit 332. In this embodiment, the accumulation determination unit 333 determines whether to accumulate the manual driving operation data as is as learning data in the learning database 322 based on accumulation criteria corresponding to the environmental conditions (environmental classification classified by the classification unit 336) under which the vehicle 2 was traveling when the manual driving operation data was obtained. Note that in this embodiment, the accumulation criteria are set by the criteria setting unit 335 for each environmental condition (in this embodiment, for each environmental classification).

[0045] For example, the accumulation determination unit 333 determines whether to accumulate the manual driving operation data as learning data as is based on the deviation calculated by the comparison unit 332 (i.e., an index representing the difference between the control parameters represented by the manual driving operation data and the control parameters represented by the system driving operation data). In particular, in this embodiment, the accumulation determination unit 333 determines to accumulate the manual driving operation data as learning data as is when the deviation is equal to or less than the accumulation reference value corresponding to the environmental conditions. On the other hand, the accumulation determination unit 333 determines not to accumulate the manual driving operation data as learning data as is when the deviation is greater than the accumulation reference value corresponding to the environmental conditions.

[0046] Fig. 6 is a diagram showing the change in speed of vehicle 2 when vehicle 2 travels on a certain road. In particular, Fig. 6 shows the change in elevation of the road on which vehicle 2 travels, as well as the accompanying change in speed of vehicle 2. In the figure, the solid line related to speed represents the change in speed of vehicle 2 when vehicle 2 is manually driven by a driver in environmental conditions accompanied by changes in elevation as shown in Fig. 6. Furthermore, the dashed line related to speed represents the change in speed of vehicle 2 when vehicle 2 is autonomously driven by a system in the same environmental conditions.

[0047] As shown in FIG. 6, even if the elevation of the road on which vehicle 2 is traveling changes, the driver generally does not change the amount of depression of the accelerator pedal that much. As a result, as shown by the solid line in FIG. 6, the speed of vehicle 2 tends to increase when vehicle 2 travels downhill, and the speed of vehicle 2 tends to decrease when vehicle 2 travels uphill. On the other hand, when vehicle 2 is driven autonomously by a system, the driving of vehicle 2 is generally controlled so that the speed of vehicle 2 does not change. Therefore, as shown by the dashed line in FIG. 6, the speed of vehicle 2 does not change that much even when vehicle 2 travels downhill or uphill.

[0048] As shown by the solid line in FIG. 6 , when the speed of vehicle 2 changes on an uphill or downhill slope, the speed of other vehicles following vehicle 2 also changes accordingly, which may ultimately result in a traffic jam. Therefore, if the speed of vehicle 2 manually driven by the driver changes significantly when the elevation of the road on which vehicle 2 is traveling changes, it is preferable not to use the manual driving operation data generated as a result of such manual driving as learning data. In this embodiment, in such cases, the degree of deviation between the manual driving by the driver and the autonomous driving by the system becomes large. As a result, the manual driving operation data generated as a result of such sliding driving is not accumulated as learning data in the learning database 322. As a result, the accumulation of inappropriate learning data is suppressed. Furthermore, in this embodiment, whether to accumulate the manual driving operation data as is in the learning database 322 is determined based on the degree of deviation automatically calculated by the comparison unit 332. Therefore, in this embodiment, learning data can be appropriately accumulated without detailed condition setting for data accumulation.

[0049] Furthermore, for at least a portion of the manual driving operation data that has been determined not to be stored as learning data in the learning database 322, the storage determination unit 333 determines whether to modify at least a portion of the manual driving operation data and store it as learning data in the learning database 322. In this embodiment, the storage determination unit 333 determines whether to modify a portion of the manual driving operation data and store it as learning data in the database for each environmental condition under which the manual driving operation data was obtained.

[0050] For example, the accumulation determination unit 333 determines whether to correct at least a portion of the manual driving operation data and store it as learning data based on whether sufficient learning data has already been stored for the environmental conditions under which the manual driving operation data was obtained. Specifically, for example, if the learning database 322 has stored a predetermined number of pieces of learning data or more for the environmental classification under which the manual driving operation data was obtained, the accumulation determination unit 333 does not correct the manual driving operation data and store it in the learning database 322. On the other hand, for example, if the learning database 322 has stored less than the predetermined number of pieces of learning data under the environmental classification under which the manual driving operation data was obtained, the accumulation determination unit 333 corrects the manual driving operation data and stores it in the learning database 322.

[0051] Here, there may be environmental conditions in which there is often a large discrepancy between manual driving by the driver and autonomous driving by the system (for example, at a location with a stop sign, manual driving by the driver often slows down the vehicle but does not come to a complete stop, whereas autonomous driving by the system brings the vehicle to a complete stop). Under such environmental conditions, if the determination of whether to accumulate data is based on the discrepancy, no learning data will be accumulated. In contrast, in this embodiment, even when the discrepancy is large, at least a portion of the manual driving operation data is corrected before accumulation of learning data, thereby preventing learning data from not being accumulated under some environmental conditions. Furthermore, in this embodiment, if sufficient learning data has already been accumulated for each environmental condition, at least a portion of the manual driving operation data is not corrected before accumulation of learning data. This prevents unnecessary accumulation of corrected data.

[0052] The accumulation determination unit 333 may determine to accumulate as learning data all of the manual driving operation data that has been determined not to be accumulated as learning data after correcting at least a portion of each piece of manual driving operation data. Alternatively, the accumulation determination unit 333 may determine whether to accumulate as learning data after correcting at least a portion of the manual driving operation data based on the speed at which the manual driving operation data is accumulated for the environment classification when the manual driving operation data was obtained. In this case, for the environment classification in which the speed at which the manual driving operation data is accumulated is slow, the accumulation determination unit 333 determines to accumulate as learning data after correcting at least a portion of the manual driving operation data.

[0053] The accumulation determination unit 333 receives as input the result of the comparison by the comparison unit 332 (deviation), the environmental classification classified by the classification unit 336, and the reference value (reference deviation) set by the reference setting unit 335. In addition, the accumulation determination unit 333 receives as input the manual driving operation data and the system driving operation data from the comparison unit 332. The accumulation determination unit 333 then outputs the determined data accumulation mode to the data accumulation unit 334. The accumulation determination unit 333 also outputs the input manual driving operation data and the system driving operation data to the accumulation determination unit 333 as is.

[0054] The data accumulation unit 334 accumulates data in the learning database 322 based on the data accumulation mode determined by the accumulation determination unit 333. When the accumulation determination unit 333 determines that the manual driving operation data should be accumulated as learning data as is, the data accumulation unit 334 accumulates the manual driving operation data as is together with the corresponding environmental data in the learning database 322 of the server storage unit 32 as learning data. On the other hand, when the accumulation determination unit 333 determines that the manual driving operation data should be corrected at least in part before being accumulated as learning data, the data accumulation unit 334 corrects at least in part the manual driving operation data and accumulates it in the learning database 322 as learning data together with the corresponding environmental data.

[0055] Specifically, in this embodiment, when the data accumulation unit 334 corrects at least a portion of the manual driving operation data and stores it as learning data, it corrects at least a portion of the manual driving operation data to the corresponding system driving operation data and stores it in the database as learning data. For example, if the deviation between the manual driving operation data and the system driving operation data in a specific section of the manual driving operation data is large, the data accumulation unit 334 replaces the manual driving operation data for that specific section with the system driving operation data and stores it as learning data in the learning database 322. This allows relatively appropriate data to be stored in the learning database 322 as learning data without performing complex processing when the deviation is large.

[0056] Alternatively, the data accumulation unit 334 may correct at least a portion of the manual driving operation data according to a predetermined correction rule. In this case, for example, if the deviation from the system driving operation data is large in a specific area of ​​the manual driving operation data, the data accumulation unit 334 corrects the manual driving operation data for that specific area according to the predetermined correction rule and stores the data in the database as learning data. According to the correction rule, for example, if the vehicle speed fluctuates on a slope, the manual driving operation data is corrected so that the vehicle speed remains constant. Also, according to the correction rule, for example, if there is a stop sign, the manual driving operation data is corrected so that the vehicle comes to a complete stop. This allows data that at least partially reflects the manual driving operation data to be stored in the learning database 322 even if the deviation is large.

[0057] The data accumulation unit 334 receives the data accumulation mode determined by the accumulation determination unit 333 from the accumulation determination unit 333. In addition, the data accumulation unit 334 receives the manual driving operation data and the system driving operation data via the comparison unit 332 and the accumulation determination unit 333. The data accumulation unit 334 then accumulates the learning data in the learning database 322.

[0058] The standard setting unit 335 sets an accumulation standard (in this embodiment, an accumulation standard value) based on the magnitude of the average deviation between the manual driving operation data and the system driving operation data for each environmental condition (environmental classification). For example, the standard setting unit 335 sets the accumulation standard for an environmental classification so that the greater the average deviation in that environmental classification, the more likely manual driving operation data is collected. Specifically, for example, the greater the average deviation in that environmental classification, the greater the accumulation standard value for that environmental classification.

[0059] For example, in a traffic jam environment, the steering angle of the steering wheel tends to be relatively large, and the deviation tends to be large. In this embodiment, the accumulation reference value for the steering angle is set to be relatively large in such an environment classification. Therefore, according to this embodiment, it is possible to accumulate the necessary manual driving operation data even in such an environment classification.

[0060] Specifically, the standard setting unit 335 receives, for each vehicle 2, the manual driving operation data, the system driving operation data, and the environmental classification classified by the classification unit 336. Then, based on this data input for the multiple vehicles 2, the standard setting unit 335 calculates the average value of the deviation degree for the multiple vehicles 2 for each environmental classification. Then, the standard setting unit 335 sets the accumulation standard value to be larger, stepwise or continuously, for an environmental classification with a larger deviation degree. The standard setting unit 335 outputs the accumulation standard (in this embodiment, the accumulation standard value) set for each environmental classification to the accumulation determination unit 333.

[0061] In this embodiment, the standard setting unit 335 sets the accumulation standard value based on the average deviation between the control parameters represented by the manual driving operation data and the control parameters represented by the system driving operation data for each environment classification. However, if an index other than the deviation is used to represent the difference between the control parameters of the manual driving operation data and the system driving operation data, the standard setting unit 335 sets an accumulation standard corresponding to this index. In this case, too, the accumulation standard for a given environment classification is set so that the larger the average deviation in that environment classification, the more likely manual driving operation data is to be collected.

[0062] The classification unit 336 classifies the environmental conditions at the time when the vehicle 2 traveled into a plurality of environmental classifications based on the environmental data at that time. In particular, the classification unit 336 classifies the environmental conditions at the time when the vehicle 2 traveled into environmental classifications that are separated based on at least one of geographical conditions, time conditions, weather conditions, and conditions related to traffic participants around the vehicle. Therefore, the classification unit 336 classifies the environmental conditions at the time when the vehicle 2 traveled into different environmental classifications for each topography of the road on which the vehicle 2 travels, and also into different environmental classifications for each shape of objects around the road on which the vehicle 2 travels. This makes it possible to appropriately classify the environmental conditions at the time when the vehicle 2 traveled into each condition that affects the deviation degree.

[0063] The classification unit 336 receives environmental data generated by the vehicle 2. Based on the input environmental data, the classification unit 336 outputs a classification result (environment classification) of the environmental conditions corresponding to the input environmental data to the accumulation determination unit 333 and the standard setting unit 335. As a result, the accumulation determination unit 333 receives the manual driving operation data via the comparison unit 332, and also receives the environmental classification when the manual driving operation data was obtained.

[0064] <Data accumulation processing> Next, the flow of data storage processing by the data storage system 1 according to the first embodiment will be described with reference to Fig. 7. Fig. 7 is a flowchart showing the flow of the data storage processing. The data storage processing shown in Fig. 7 is executed by the server processor 33.

[0065] 7, in the data accumulation process, first, the server processor 33 acquires data from the raw database 321 of the server storage unit 32 (step S11). Specifically, the comparison unit 332 acquires manual driving operation data, and the system operation generation unit 331 and the classification unit 336 acquire environmental data corresponding to the manual driving operation data (environmental data representing the environmental conditions when the manual driving operation data was obtained).

[0066] Next, the classification unit 336 classifies the environmental conditions under which the vehicle 2 was traveling into environmental categories based on the environmental data acquired in step S11 (step S12). Next, the system operation generation unit 331 generates system driving operation data as comparative driving operation data to be compared with the manual driving operation data based on the environmental data acquired in step S11 (step S13). Thereafter, the comparison unit 332 calculates the degree of deviation between the manual driving by the driver and the autonomous driving by the system based on the manual driving operation data acquired in step S11 and the comparative driving operation data generated in step S13 (step S14).

[0067] Next, the accumulation determination unit 333 specifies an accumulation reference value corresponding to the environmental classification based on the environmental classification classified by the classification unit 336 (step S15). The relationship between the environmental classification and the accumulation reference value is set by the standard setting unit 335, and the accumulation determination unit 333 specifies the accumulation reference value based on the environmental classification classified by the classification unit 336 using this relationship set by the standard setting unit 335. Next, the accumulation determination unit 333 determines whether the deviation calculated in step S14 is equal to or less than the accumulation reference value specified in step S15 (step S16).

[0068] If it is determined in step S16 that the deviation is equal to or less than the accumulation reference value, the accumulation determination unit 333 determines that the manual driving operation data is to be learned, that is, that the manual driving operation data is to be accumulated as learning data in the learning database 322 (step S17). As a result, the data accumulation unit 334 accumulates the manual driving operation data as learning data in the learning database 322. At this time, in addition to the manual driving operation data, the data accumulation unit 334 accumulates environmental data corresponding to this manual driving operation data as learning data.

[0069] On the other hand, if it is determined in step S16 that the deviation is greater than the accumulation reference value, the accumulation determination unit 333 determines whether or not the correction data accumulation condition is met (step S18). The correction data accumulation condition is, for example, a condition that is met when, as described above, less than the reference number of pieces of training data are accumulated in the training database 322 for the corresponding environment classification. Therefore, the accumulation determination unit 333 determines whether or not the correction data accumulation condition is met based on, for example, the type and amount of training data accumulated in the training database 322.

[0070] If it is determined in step S18 that the correction data accumulation condition is met, the accumulation determination unit 333 determines that the correction data obtained by correcting the manual driving operation data will be used as the learning data, i.e., that the correction data will be accumulated in the learning database 322 as learning data (step S19). As a result, the data accumulation unit 334 accumulates the correction data as learning data in the learning database 322. At this time, the data accumulation unit 334 accumulates the corresponding environmental data as learning data in addition to the correction data. On the other hand, if it is determined in step S18 that the correction data accumulation condition is not met, the accumulation determination unit 333 determines that neither the manual driving operation data nor the correction data will be used as the learning data, i.e., that neither the manual driving operation data nor the correction data will be accumulated in the learning database 322 as learning data.

[0071] Second embodiment Next, a data storage system 1 according to a second embodiment will be described with reference to Figures 8 and 9. The configuration and control of the data storage system 1 according to the second embodiment are basically the same as the configuration and control of the data storage system 1 according to the first embodiment. The following description will focus on the parts that are different from the configuration and control of the data storage system 1 according to the first embodiment.

[0072] In the first embodiment, the comparative driving operation data to be compared with the manual driving operation data is system driving operation data generated under the assumption that the vehicle 2 is being driven autonomously by the system under the same environmental conditions as those under which the manual driving operation data was acquired. In contrast, in the second embodiment, the comparative driving operation data is also similar driving operation data obtained when the vehicle 2 is driven under environmental conditions similar to those under which the manual driving operation data was acquired.

[0073] Fig. 8 is a functional block diagram of the server processor 33 according to the second embodiment. As shown in Fig. 8, the server processor 33 according to this embodiment includes a determination unit 337 in addition to the units included in the server processor 33 according to the first embodiment.

[0074] Based on the classification result by the classification unit 336, the determination unit 337 determines whether or not similar driving operation data exists in the server storage unit 32 when the vehicle 2 was driven under environmental conditions similar to those under which the vehicle 2 traveled. In this embodiment, the similar driving operation data is past manual driving operation data acquired when the vehicle 2 was manually driven by the driver of the vehicle 2 or a driver other than the vehicle 2, for each environmental classification. In particular, in this embodiment, the similar driving operation data is past manual driving operation data acquired when the vehicle was manually driven by a good driver, for each environmental classification. Therefore, the determination unit 337 determines whether or not past manual driving operation data acquired when the vehicle 2 was manually driven by a good driver exists in the environmental classification classified by the classification unit 336.

[0075] In this embodiment, the server storage unit 32 stores past manual driving operation data acquired when the vehicle was manually driven by a good driver, along with the classification of the environmental conditions (environment classification) when the manual driving operation data was acquired. Therefore, the server storage unit 32 stores various environmental classifications and past manual driving operation data by good drivers corresponding to those environmental classifications. However, the server storage unit 32 does not necessarily store past manual driving operation data by good drivers corresponding to all environmental classifications.

[0076] The determination unit 337 receives the environmental classification classified by the classification unit 336. Based on the environmental classification classified by the classification unit 336, the determination unit 337 determines whether past manual driving operation data by a good driver corresponding to the environmental classification is stored in the server storage unit 32. If the determination unit 337 determines that past manual driving operation data by a good driver corresponding to the environmental classification is stored in the server storage unit 32, it outputs the past manual driving operation data to the comparison unit 332 as similar driving operation data. On the other hand, if the determination unit 337 determines that past manual driving operation data by a good driver corresponding to the environmental classification is not stored in the server storage unit 32, it outputs a signal indicating that past manual driving operation data is not stored, i.e., that similar driving operation data is not stored, to the system operation generation unit 331.

[0077] When the determination unit 337 determines that similar driving operation data is stored in the server storage unit 32, i.e., when similar driving operation data is input from the determination unit 337, the comparison unit 332 uses the similar driving operation data as comparison driving operation data. Therefore, in this case, the comparison unit 332 compares the manual driving operation data acquired when the vehicle 2 is manually driven under any environmental condition (any environmental classification) with the similar driving operation data acquired when the vehicle 2 is driven under environmental conditions similar to the above-mentioned environmental conditions (the same environmental classification). The comparison unit 332 then outputs the comparison result (e.g., the degree of deviation) between the manual driving operation data and the similar driving operation data to the accumulation determination unit 333. The comparison between the manual driving operation data and the similar driving operation data is performed in the same manner as the comparison between the manual driving operation data and the system driving operation data.

[0078] On the other hand, when the determination unit 337 determines that the similar driving operation data is not stored in the server storage unit 32, i.e., when the determination unit 337 inputs a signal indicating that the similar driving operation data is not stored, the system operation generation unit 331 generates system driving operation data. The system operation generation unit 331 then inputs the generated system driving operation data to the comparison unit 332. In this case, the comparison unit 332 compares the manual driving operation data with the system driving operation data. Therefore, in this embodiment, the comparison unit 332 compares the manual driving operation data with the similar driving operation data and the system driving operation.

[0079] In this manner, in this embodiment, if similar driving operation data is stored in the server storage unit 32, the similar driving operation data is used. As a result, the deviation is calculated based on a comparison with driving operation data manually driven by, for example, a good driver. This makes it possible to more appropriately determine the data to be stored in the learning database 322 as learning data.

[0080] The data storage system 1 according to the first embodiment and the data storage system 1 according to the second embodiment may be used in combination. In this case, the data storage system 1 according to the first embodiment and the data storage system 1 according to the second embodiment may be used depending on the conditions. For example, if the manual driving operation data is data obtained when driving on a motorway, system driving operation data may be used as the comparative driving operation data, and if the manual driving operation data is data obtained when driving on an ordinary road, similar driving operation data may be used as the comparative driving operation data. This makes it possible to appropriately determine the data to be stored in the training database 322 as training data for both motorway and ordinary roads when autonomous driving by the system is only possible on motorway.

[0081] Next, the flow of data storage processing by the data storage system 1 according to the second embodiment will be described with reference to Fig. 9. Fig. 9 is a flowchart showing the flow of data storage processing according to the second embodiment. The data storage processing shown in Fig. 9 is executed by the server processor 33. Note that steps S31, S32, S34, and S36 to S42 in Fig. 9 are similar to steps S11 to S20 in Fig. 7, respectively, and therefore will not be described again.

[0082] 9, after the classification unit 336 performs classification in step S32, the determination unit 337 determines whether similar driving operation data is stored in the server storage unit 32 for the environment classification obtained by the classification unit 336 (step S33). If it is determined in step S33 that similar driving operation data is not stored in the server storage unit 32, the system operation generation unit 331 generates system driving operation data as comparative driving operation data (step S34). On the other hand, if it is determined in step S33 that similar driving operation data is stored in the server storage unit 32, the determination unit 337 sets the similar driving operation data as comparative driving operation data to be compared with the manual driving operation data (step S35).

[0083] Third embodiment Next, a data storage system 1 according to a third embodiment will be described with reference to Figures 10 and 11. The configuration and control of the data storage system 1 according to the third embodiment are basically the same as those of the data storage system 1 according to the first embodiment. The following will focus on the differences from the configuration and control of the data storage system 1 according to the first embodiment. Note that in the third embodiment, some of the control of the data storage system 1 according to the first embodiment is changed, but some of the control of the data storage system 1 according to the second embodiment may also be changed.

[0084] In the data accumulation system 1 according to the first embodiment, driver status data that indicates the driver's status is not used to determine whether to accumulate the manual driving operation data as is. In contrast, in the data accumulation system 1 according to the third embodiment, driver status data that indicates the driver's status is used in addition to the degree of discrepancy between the manual driving operation data and the comparative driving operation data to determine whether to accumulate the manual driving operation data as is.

[0085] Fig. 10 is a functional block diagram of the server processor 33 according to the third embodiment. As shown in Fig. 10, the server processor 33 according to the third embodiment includes a driver state determination unit 338 in addition to the units included in the server processor 33 according to the first embodiment.

[0086] The driver state determination unit 338 determines the driver's state based on the output of the driver state sensor 24, which detects the driver's state. In this embodiment, the driver state determination unit 338 detects the driver's state represented by each image data by, for example, sequentially inputting image data of images captured by the driver monitor camera 241 into a classifier. In particular, in this embodiment, the driver state determination unit 338 determines whether the driver's state is a concentrated driving state in which the driver is concentrating on driving, or a distracted driving state in which the driver is not concentrating on driving. For example, when the driver is looking away from the wheel, when the driver has their eyes closed, when the driver's posture is poor, or when the driver's hands are off the steering wheel, the driver's driving is determined to be distracted driving. The classifier is, for example, a convolutional neural network (CNN) having multiple convolution layers connected in series from the input side to the output side. Note that in this embodiment, the driver monitor camera 241 is used as the driver state sensor 24, but other sensors, such as a grip sensor that detects whether the driver is gripping the steering wheel, may be used.

[0087] Furthermore, in this embodiment, the accumulation determination unit 333 determines whether or not to accumulate the manual driving operation data as learning data in the learning database 322 based on the driver's state determined by the driver state determination unit 338 in addition to the comparison result by the comparison unit 332. Specifically, for example, if the driver state determination unit 338 determines that the driver's driving state is a distracted driving state, the accumulation determination unit 333 determines not to accumulate the manual driving operation data as learning data in the learning database 322 regardless of the comparison result by the comparison unit 332. On the other hand, for example, if the driver state determination unit 338 determines that the driver's driving state is a concentrated driving state, the accumulation determination unit 333 determines whether or not to accumulate the manual driving operation data as learning data in the learning database 322 based on the comparison result by the comparison unit 332.

[0088] According to this embodiment, whether or not to store the manual driving operation data as learning data is determined based on the driver's state determined by the driver state determination unit 338 in addition to the comparison result by the comparison unit 332. This prevents manual driving operation data when the driver's driving state is inappropriate (for example, when the driver's driving state is absentminded) from being stored as learning data, thereby enabling appropriate learning data to be stored in the learning database 322.

[0089] At least one of the data storage system 1 according to the first embodiment and the data storage system 1 according to the second embodiment may be used in combination with the data storage system 1 according to the third embodiment. In this case, the data storage system 1 according to the first embodiment, the data storage system 1 according to the second embodiment, and the data storage system 1 according to the third embodiment may be used depending on the conditions.

[0090] Next, the flow of data storage processing by the data storage system 1 according to the third embodiment will be described with reference to Fig. 11. Fig. 11 is a flowchart showing the flow of data storage processing according to the third embodiment. The data storage processing shown in Fig. 11 is executed by the server processor 33. Note that steps S51 to S55 and S58 to S61 in Fig. 11 are similar to steps S11 to S15 and S17 to S20 in Fig. 7, respectively, and therefore will not be described again.

[0091] 11, when the accumulation reference value corresponding to the environment classification is identified in step S55, the driver state determination unit 338 determines the state of the driver based on the output of the driver state sensor 24 (step S56). Next, the accumulation determination unit 333 determines whether the deviation calculated in step S54 is equal to or less than the accumulation reference value and whether the driver is in a concentrated driving state (step S57).

[0092] If it is determined in step S57 that the deviation is equal to or less than the accumulation reference value and that the driver is in a concentrated driving state, the accumulation determination unit 333 determines that the manual driving operation data is to be learned (step S58).On the other hand, if it is determined in step S57 that the deviation is greater than the accumulation reference value or that the driver is in a distracted driving state, the accumulation determination unit 333 determines whether the correction data accumulation condition is met (step S59).

[0093] 12, the server processor 33 may include an identification unit 339 that identifies an area on a road where the driver's state is likely to become unsuitable for driving (for example, a state of careless driving) based on multiple determination results by the driver state determination unit 338. The determination results of the driver state determination unit 338 regarding various types of driving are input to the identification unit 339, as well as environmental data corresponding to the driving at that time (particularly, data on the area in which the vehicle 2 was traveling). Then, based on such input data, the identification unit 339 identifies an area where the driver's state is likely to become a state of careless driving (hereinafter also referred to as a "careless driving area"). The identification unit 339 outputs information on the careless driving area to the standard setting unit 335.

[0094] The standard setting unit 335 sets the accumulation standard based on the distracted driving area in addition to the environment classification. In particular, the standard setting unit 335 sets the accumulation standard for the distracted driving area so that manual driving operation data is less likely to be collected. Therefore, the standard setting unit 335 sets a small accumulation standard value for the distracted driving area. As a result, the accumulation determination unit 333 determines whether or not to accumulate manual driving operation data as is in the learning database 322 using different accumulation standards for each area on the road. This prevents manual driving operation data when the driver's driving state is inappropriate (for example, when the driver's driving state is distracted) from being accumulated as learning data, thereby enabling appropriate learning data to be accumulated in the learning database 322.

[0095] Although preferred embodiments according to the present disclosure have been described above, the present disclosure is not limited to these embodiments, and various modifications and changes can be made within the scope of the claims. [Explanation of symbols]

[0096] 1. Data storage system 2 vehicles 3 Server 32 Server storage unit 322 Learning Database 33 Server Processors

Claims

1. A data storage device that stores learning data for vehicle driving operations in a database, a comparison unit that compares manual driving operation data acquired when the vehicle is being manually driven under predetermined environmental conditions with comparative driving operation data including at least one of system driving operation data generated under the assumption that the vehicle is being driven autonomously by a system of the vehicle under the same environmental conditions as the predetermined environmental conditions and similar driving operation data when the vehicle is driven under environmental conditions similar to the predetermined environmental conditions; A data storage device having an accumulation decision unit that decides whether to store the manual driving operation data as is in the database as learning data based on the comparison result between the manual driving operation data and the comparison driving operation data by the comparison unit.

2. The data storage device of claim 1, wherein the learning data of the driving operation is used to construct a machine learning model through predetermined learning, and autonomous driving by the vehicle system is performed using a model different from the machine learning model constructed through the predetermined learning.

3. a classification unit that classifies environmental conditions when the vehicle is traveling; a determination unit that determines whether or not similar driving operation data when the vehicle was driven under environmental conditions similar to the environmental conditions under which the vehicle traveled is stored in a storage unit based on the classification result by the classification unit, 3. The data storage device according to claim 1, wherein the comparison unit uses the similar driving operation data as the comparison driving operation data when the determination unit determines that similar driving operation data under similar environmental conditions is stored in the memory unit.

4. The data storage device according to claim 3 , wherein the classification unit classifies the environmental conditions under which the vehicle traveled based on at least one of geographical conditions, time conditions, weather conditions, and conditions related to traffic participants around the vehicle.

5. The data storage device according to claim 1 or 2, wherein the storage decision unit further decides whether to modify at least a portion of the manual driving operation data that has been determined not to be stored in the database as the learning data and then store the modified data in the database as the learning data.

6. a data storage unit that stores data in a database based on the decision of the storage decision unit; 6. The data storage device of claim 5, wherein when the storage decision unit determines to modify at least a portion of the manual driving operation data and store it as the learning data, the data storage unit modifies at least a portion of the manual driving operation data to the corresponding comparison driving operation data and stores it in the database as the learning data.

7. 6. The data storage device according to claim 5, wherein the storage decision unit determines whether to modify a portion of the manual driving operation data and store it in the database as the learning data for each environmental condition under which the vehicle travels.

8. The system further includes a standard setting unit that sets an accumulation standard based on an average deviation between the manual driving operation data and the comparative driving operation data for each environmental condition, 3. The data storage device according to claim 1, wherein the storage decision unit decides whether to store the manual driving operation data as is in the database as the learning data based on the storage criteria corresponding to the environmental conditions under which the vehicle was traveling.

9. a driver state determination unit that determines the state of the driver based on an output of a sensor that detects the state of the driver; 3. The data storage device according to claim 1, wherein the storage determination unit determines whether or not to store the manual driving operation data as the learning data in the database based on the driver's state determined by the driver state determination unit.

10. The vehicle driving control system further includes an identification unit that identifies an area where the driver's condition is likely to become unsuitable for driving based on a plurality of determination results by the driver condition determination unit, 10. The data storage device according to claim 9, wherein the storage determination unit determines whether or not to store the manual driving operation data as is in the database as the learning data, based on a different storage standard for each of the areas.

11. 1. A data storage method executed by a processor for storing learning data of a vehicle driving operation in a database, the method comprising: comparing manual driving operation data acquired when the vehicle is being manually driven under predetermined environmental conditions with comparative driving operation data including at least one of system driving operation data generated under the assumption that the vehicle is being driven autonomously by a system of the vehicle under the same environmental conditions as the predetermined environmental conditions and similar driving operation data when the vehicle is driven under environmental conditions similar to the predetermined environmental conditions; and determining whether or not to store the manual driving operation data as is in the database as the learning data based on the comparison result between the manual driving operation data and the comparative driving operation data.

12. A data storage program for storing learning data of vehicle driving operations in a database, comparing manual driving operation data acquired when the vehicle is being manually driven under predetermined environmental conditions with comparative driving operation data including at least one of system driving operation data generated under the assumption that the vehicle is being driven autonomously by a system of the vehicle under the same environmental conditions as the predetermined environmental conditions and similar driving operation data when the vehicle is driven under environmental conditions similar to the predetermined environmental conditions; determining whether to store the manual driving operation data as is in the database as the learning data based on a comparison result between the manual driving operation data and the comparative driving operation data; A data storage program that causes a computer to execute the following.

Images