Information processing device, information processing computer program, and information processing method

Abstract

To provide an information processing device capable of reducing a load of a hardware by releasing detection accuracy in detection processing, when a processing amount of the detection processing is low.SOLUTION: An information processing device comprises: a collection part by which processing is executed using a hardware common to a detection part that executes detection processing for controlling a vehicle, as the collection part for collecting information about controlling of the vehicle for machine learning; an estimation part for estimating a degree of a processing amount of the detection processing at the detection part, based on at least a piece of information among vehicle information representing a state of the vehicle, environment information representing an environment around the vehicle, and geographical information representing geographical information including a current position of the vehicle; and a determination part for determining that detection accuracy at the detection part is more released when it is estimated that the processing amount of the detection processing at the detection part is low than when it is estimated that the processing amount of the detection processing at the detection part is high.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, an information processing computer program, and an information processing method.

Background Art

[0002] An automatic control device that controls a vehicle performs vehicle control using a machine-learned identifier. When the vehicle is automatically driven, the identifier is used in detection processing for detecting the environment around the vehicle and control processing for generating a signal for controlling the vehicle.

[0003] When the vehicle is manually driven, a signal for controlling the vehicle is generated based on the driver's operation. The machine-learned identifier can further improve the accuracy of identification by being educated. Therefore, the automatic control device collects information when the vehicle is manually driven. The collected data is used as teacher data and used for further improvement of the identifier.

[0004] For example, Patent Document 1 proposes acquiring precise data by collecting data at a sampling interval that varies according to the traveling speed of the vehicle.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] The automatic control device performs detection processing and control processing for detecting the environment around the vehicle in the same manner as when the vehicle is automatically driven even when the vehicle is manually driven.

[0007] In an automatic control device, a load for performing detection processing and a load for collecting data related to vehicle control for machine learning occur, and it is required to reduce the load on the automatic control device.

[0008] The amount of detection processing changes according to the environment around the vehicle. For example, when the environment around the vehicle is not complex, the amount of detection processing is lower than when the environment around the vehicle is complex. Even when the vehicle is in autonomous driving, when the amount of detection processing is low, since the environment around the vehicle is not complex, it is considered that even if the detection accuracy of the detection processing is relaxed, the impact on vehicle safety is small. Also, when the vehicle is in manual driving, since the vehicle is controlled based on the driver's operation, even if the detection accuracy of the detection processing is relaxed, there is no impact on vehicle safety.

[0009] Therefore, an object of the present disclosure is to provide an information processing apparatus that can reduce the hardware load by relaxing the detection accuracy in detection processing when the amount of detection processing is low.

Means for Solving the Problem

[0010] (1) According to one embodiment, an information processing apparatus is provided. This information processing apparatus includes a collection unit that collects information related to vehicle control for machine learning, and the collection unit is processed using the same hardware as a detection unit that executes detection processing for controlling the vehicle. Based on at least one of vehicle information representing the state of the vehicle, environment information representing the environment around the vehicle, and terrain information representing the terrain including the current position of the vehicle, an estimation unit that estimates the degree of the amount of detection processing in the detection unit, and when it is estimated by the estimation unit that the amount of detection processing in the detection unit is low, a determination unit that determines to relax the detection accuracy in the detection unit more than when it is estimated that the amount of detection processing in the detection unit is high.

[0011] (2) In the information processing apparatus according to (1), the vehicle information includes information representing the degree of operation of the vehicle, and it is preferable that the estimation unit estimates that the processing amount of the detection processing in the detection unit is lower when the operation of the vehicle is slow than when the operation of the vehicle is fast.

[0012] (3) In the information processing apparatus according to (1) or (2), the environment information includes information representing the degree of complexity of the environment around the vehicle, and it is preferable that the estimation unit estimates that the processing amount of the detection processing in the detection unit is lower when the environment around the vehicle is not complex than when the environment around the vehicle is complex.

[0013] (4) In the information processing apparatus according to any one of (1) to (3), the environment information includes information representing the degree of complexity of the environment around the vehicle, and it is preferable that the estimation unit estimates that the processing amount of the detection processing in the detection unit is lower when the environment around the vehicle is not complex than when the environment around the vehicle is complex.

[0014] (5) In the information processing apparatus according to any one of (1) to (4), it is preferable that the determination unit determines to relax the detection accuracy in the detection unit by increasing the period for executing the detection processing in the detection unit.

[0015] (6) In the information processing apparatus according to any one of (1) to (5), it is preferable that the determination unit determines to relax the detection accuracy in the detection unit by reducing the number of sensors through which the detected information is input to the detection unit.

[0016] (7) In the information processing apparatus according to any one of (1) to (6), it is preferable that the determination unit determines to relax the detection accuracy in the detection unit by reducing the area in the image detected by the detection unit.

[0017] (8) In the information processing apparatus according to any one of (1) to (7), it is preferable that the determination unit determines to relax the detection accuracy in the detection unit by reducing the distance of the trajectory of the moving object estimated in the detection unit.

[0018] (9) In any of the information processing apparatuses (1) to (8), the detection unit includes a first detection unit capable of independently executing a detection process, and a second detection unit having a lower detection accuracy than the first detection unit and a lower hardware load during operation. When the determination unit estimates that the processing amount of the detection process in the detection unit estimated by the estimation unit is low, it is preferable that the second detection unit is used. When the determination unit estimates that the processing amount of the detection process in the detection unit estimated by the estimation unit is high, it is preferable that the first detection unit is used.

[0019] (10) In any of the information processing apparatuses (1) to (9), when the determination unit determines to relax the detection accuracy in the detection unit, it is preferable that the collection unit increases the processing amount of collecting information related to the control of the vehicle for machine learning.

[0020] (11) According to another embodiment, an information processing computer program is provided. This information processing computer program collects information related to the control of the vehicle for machine learning, collects information using the same hardware as the detection unit that executes the detection process for controlling the vehicle, and estimates the degree of the processing amount of the detection process in the detection unit based on at least one of vehicle information representing the state of the vehicle, environmental information representing the environment around the vehicle, and terrain information representing the terrain including the current position of the vehicle. When it is estimated that the processing amount of the detection process in the detection unit is low, it is determined to relax the detection accuracy in the detection unit more than when it is estimated that the processing amount of the detection process in the detection unit is high, and causes the processor to execute a process including this.

[0021] (12) According to another embodiment, an information processing method is provided. This information processing method is that an information processing apparatus collects information related to the control of a vehicle for machine learning, collects the information using the same hardware as a detection unit that executes a detection process for controlling the vehicle, and based on at least one of vehicle information representing the state of the vehicle, environmental information representing the environment around the vehicle, and terrain information representing the terrain including the current position of the vehicle, estimates the degree of the processing amount of the detection process in the detection unit, and when it is estimated that the processing amount of the detection process in the detection unit is low, determines to relax the detection accuracy in the detection unit more than when it is estimated that the processing amount of the detection process in the detection unit is high, and executes a process including this.

Advantages of the Invention

[0022] The information processing apparatus according to the present invention can reduce the hardware load by relaxing the detection accuracy in the detection process when the processing amount of the detection process is low, so that the collection process of the collection unit can be improved.

Brief Description of the Drawings

[0023]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Embodiments for Carrying Out the Invention

[0024] FIG. 1 is a diagram for explaining an outline of the operation of the control device 12 of the present embodiment. Hereinafter, an outline of the operation of the control device 12 of the present embodiment will be explained with reference to FIG. 1.

[0025] Vehicle 10 has a control device 12. Vehicle 10 supports manual driving and autonomous driving. When Vehicle 10 is manually driven, the control device 12 outputs a manual steering signal for controlling the steering device 13, a manual driving signal for controlling the driving device 14 such as an engine or a motor, and a manual braking signal for controlling the braking device 15 based on the operations of the steering wheel, accelerator pedal, and brake pedal by the driver.

[0026] Also, when Vehicle 10 is autonomously driven, the control device 12 inputs vehicle information representing the state of Vehicle 10, environmental information representing the environment around Vehicle 10, terrain information representing the terrain including the current position of Vehicle 10, etc., and outputs an automatic steering signal for controlling the steering device 13, an automatic driving signal for controlling the driving device 14 such as an engine or a motor, and an automatic braking signal for controlling the braking device 15.

[0027] For example, the control device 12 executes a detection process for detecting objects around Vehicle 10 based on the environmental information. Also, the control device 12 executes a control process for generating an automatic steering signal, an automatic driving signal, and an automatic braking signal based on vehicle information such as the speed of Vehicle 10, the objects detected by the detection process, and the map information included in the terrain information.

[0028] When Vehicle 10 is manually driven, the control device 12 performs a collection process for collecting manual driving data for training a machine-learned discriminator. The control device 12 collects vehicle information, environmental information, and terrain information, and manual steering signals, manual driving signals, and manual braking signals, and generates driving information. The driving information is sent to an external server (not shown) and used as teacher data for training the discriminator.

[0029] Also, the control device 12 executes the detection process and the control process even when Vehicle 10 is manually driven. However, when Vehicle 10 is manually driven, the automatic steering signal, automatic driving signal, and automatic braking signal generated by the control process are not output to the steering device 13, the driving device 14, and the braking device 15.

[0030] In the control device 12, the detection process and the collection process are executed using common hardware. For example, the detection process and the collection process are executed using a common processor.

[0031] In the case of manual driving, the control device 12 executes an estimation process for estimating the degree of the processing amount of the detection process based on at least one of vehicle information, environment information, and terrain information.

[0032] When it is estimated that the processing amount of the detection process is low, the control device 12 executes a determination process for determining to relax the detection accuracy in the detection process more than when it is estimated that the processing amount of the detection process is high.

[0033] Since the detection accuracy is relaxed, the load on the hardware is reduced. As a result, it becomes possible to increase the processing amount of the collection process.

[0034] For example, the control device 12 uses the speed of the vehicle 10 as vehicle information. When the speed of the vehicle 10 is less than a predetermined reference speed, the control device 12 relaxes the detection accuracy in the detection process by increasing the period of executing the detection process more than when the speed of the vehicle 10 is equal to or higher than the reference speed.

[0035] When the vehicle 10 is manually driven, the vehicle 10 is controlled based on the driver's operation. Therefore, even if the detection accuracy of the detection process is relaxed, it does not affect the safety of the vehicle 10. The automatic steering signal, automatic drive signal, and automatic brake signal generated based on the detection result are not used for controlling the vehicle 10.

[0036] As described above, the control device 12 of the present embodiment can reduce the load on the hardware by relaxing the detection accuracy in the detection process when the processing amount of the detection process is low, so that the collection process can be improved.

[0037] Next, the vehicle 10 in which the control device 12 is mounted will be described below with reference to FIG. 2. FIG. 2 is a hardware configuration diagram of the vehicle 10 in which the control device 12 of the present embodiment is mounted.

[0038] The vehicle 10 includes a steering wheel 1a, an accelerator pedal 1b, a brake pedal 1c, a communication device 2, a sensor group 3, a positioning information receiver 4, a navigation device 5, a map information storage device 11, a control device 12, a steering device 13, a driving device 14, a braking device 15, etc.

[0039] The steering wheel 1a, the accelerator pedal 1b, the brake pedal 1c, the communication device 2, the sensor group 3, the positioning information receiver 4, the navigation device 5, the map information storage device 11, the control device 12, the steering device 13, the driving device 14, and the braking device 15 are communicably connected via an in-vehicle network 16 compliant with a standard such as a controller area network.

[0040] The steering wheel 1a is operated by the driver and outputs a signal representing the steering angle of the steering wheel 1a to the control device 12 etc. via the in-vehicle network 16.

[0041] The accelerator pedal 1b is operated by the driver and outputs a signal representing the operation amount of the accelerator pedal 1b to the control device 12 etc. via the in-vehicle network 16.

[0042] The brake pedal 1c is operated by the driver and outputs a signal representing the operation amount of the brake pedal 1c to the control device 12 etc. via the in-vehicle network 16.

[0043] The communication device 2 has an interface circuit for connecting the control device 12 etc. to a communication network (not shown) via a macrocell base station (not shown). The control device 12 etc. can communicate with an external server connected to the communication network via the communication device 2.

[0044] The sensor group 3 has a plurality of sensors for detecting vehicle information, environmental information, and terrain information. For example, as sensors for detecting vehicle information, the sensor group 3 has a speed sensor for detecting information representing the speed of the vehicle 10, an acceleration sensor for detecting acceleration, an angular velocity sensor for detecting angular velocity, and the like.

[0045] As sensors for detecting environmental information, the sensor group 3 has a front camera, a rear camera, a LiDAR sensor, a millimeter-wave radar sensor, an ultrasonic sensor, and the like. The front camera acquires an image representing the environment in a predetermined range in front of the vehicle 10. The rear camera acquires an image representing the environment in a predetermined range behind the vehicle 10. The LiDAR sensor acquires reflected wave information representing laser reflectors around the vehicle 10. The millimeter-wave radar sensor acquires reflected wave information representing millimeter-wave reflectors around the vehicle 10. The ultrasonic sensor acquires reflected wave information representing ultrasonic reflectors around the vehicle 10.

[0046] The front camera and the millimeter-wave radar for detecting the environment in front of the vehicle 10 are an example of a front sensor for detecting the environmental information in front of the vehicle 10. The rear camera and the millimeter-wave radar for detecting the environment behind the vehicle 10 are an example of a rear sensor for detecting the environmental information behind the vehicle 10. The millimeter-wave radar and the LiDAR sensor for detecting the environment on the side of the vehicle 10 are an example of a side sensor for detecting the environmental information on the side of the vehicle. The ultrasonic sensor is an example of a surrounding sensor for detecting the environmental information in the vicinity around the vehicle.

[0047] Sensors for detecting environmental information such as the front camera and the LiDAR sensor are also used as sensors for acquiring terrain information representing the road around the vehicle 10.

[0048] The sensor group 3 outputs the information detected by the sensors to the control device 12 and the like via the in-vehicle network 16.

[0049] The positioning information receiver 4 outputs positioning information representing the current position of the vehicle 10. For example, the positioning information receiver 4 can be a GNSS receiver. Each time the positioning information receiver 4 acquires positioning information at a predetermined reception cycle, it outputs the positioning information and the positioning information acquisition time at which the positioning information was acquired to the navigation device 5, the map information storage device 11, and the like.

[0050] Based on the navigation map information, the destination position of the vehicle 10, and the positioning information representing the current position of the vehicle 10 input from the positioning information receiver 4, the navigation device 5 generates a navigation route from the current position of the vehicle 10 to the destination position. When the destination position is newly set, or when the current position of the vehicle 10 deviates from the navigation route, etc., the navigation device 5 newly generates the navigation route of the vehicle 10. Each time the navigation device 5 generates a navigation route, it outputs the navigation route to the control device 12 and the like via the in-vehicle network 16.

[0051] The map information storage device 11 stores wide-area map information of a relatively wide range (for example, a range of 10 km by 10 km to 30 km by 30 km) including the current position of the vehicle 10. This map information has high-precision map information including three-dimensional information of the road surface, road speed limits, road curvatures, road feature objects such as lane dividing lines on the road, information representing the types and positions of structures, and the like. In the map information, one lane is represented as a series of a plurality of lane links. Further, the map information includes the type of road. The type of road indicates whether the road is an expressway for automobiles or a general road.

[0052] The map information storage device 11 receives wide-area map information from an external server (not shown) via a macro cell base station (not shown) by wireless communication via the communication device 2 mounted on the vehicle 10 according to the current position of the vehicle 10, and stores it in the storage device. Each time the map information storage device 11 inputs positioning information from the positioning information receiver 4, it refers to the stored wide-area map information and outputs the map information of a relatively narrow area (for example, in the range of 100 m square to 10 km square) including the current position represented by the positioning information to the control device 12 etc. via the in-vehicle network 16. The map information is an example of terrain information.

[0053] The control device 12 executes detection processing, control processing, collection processing, estimation processing, and determination processing. For this purpose, the control device 12 includes a communication interface (IF) 21, a memory 22, and a processor 23. The communication interface 21, the memory 22, and the processor 23 are connected via a signal line 24. The communication interface 21 has an interface circuit for connecting the control device 12 to the in-vehicle network 16.

[0054] The memory 22 is an example of a storage unit, and has, for example, a volatile semiconductor memory and a non-volatile semiconductor memory. Then, the memory 22 stores a computer program of an application and various data used in the information processing executed by the processor 23.

[0055] All or part of the functions of the control device 12 are functional modules realized by, for example, a computer program operating on the processor 23. The processor 23 includes a detection unit 231, a control unit 232, a collection unit 233, an estimation unit 234, and a determination unit 235. Alternatively, the functional module of the processor 23 may be a dedicated arithmetic circuit provided in the processor 23. The processor 23 includes one or more CPUs (Central Processing Units) and their peripheral circuits. The processor 23 may further include other arithmetic circuits such as a logic arithmetic unit, a numerical arithmetic unit, or a graphic processing unit. The control device 12 is, for example, an electronic control unit (ECU).

[0056] The detection unit 231 performs detection processing to detect the state of the vehicle 10 such as speed, acceleration, and angular velocity based on vehicle information. Further, the detection unit 231 performs detection processing to obtain the current position and orientation of the vehicle 10 based on terrain information and environmental information. The detection unit 231 performs detection processing to detect the objects around the vehicle 10 based on environmental information. The objects include moving objects such as other vehicles or pedestrians, and structures such as guardrails. The detection unit 231 tracks the detected moving objects to obtain the trajectories and speeds of the objects. In the detection processing, a machine-learned discriminator may be used.

[0057] Further, the detection unit 231 performs detection processing to detect road features such as lane dividing lines, signs, or traffic lights based on environmental information. Further, the detection unit 231 performs detection processing to detect a construction section based on environmental information.

[0058] The detection unit 231 performs detection processing to detect the roads around the vehicle 10 based on terrain information.

[0059] When the vehicle 10 is under autonomous driving, the control unit 232 performs control processing to generate a driving lane plan representing the planned driving lane on which the vehicle 10 is to travel based on the current position of the vehicle 10, the navigation route, vehicle information, environmental information, and terrain information. Further, the control unit 232 performs control processing to generate a driving plan representing the planned driving trajectory of the vehicle 10 up to a predetermined time (for example, 5 seconds) ahead based on the driving lane plan.

[0060] When the vehicle 10 is under autonomous driving, the control unit 232 controls each part of the vehicle 10 based on the driving plan. The control unit 232 generates an automatic steering signal for controlling the steering device 13 that controls the steering wheel of the vehicle 10 based on the driving plan. The control unit 232 generates an automatic driving signal for controlling the driving device 14 such as the engine or motor of the vehicle 10 based on the driving plan. The control unit 232 generates an automatic braking signal for controlling the braking device 15 of the vehicle 10 based on the driving plan. The control unit 232 outputs the automatic steering signal, the automatic driving signal, or the automatic braking signal to the steering device 13, the driving device 14, and the braking device 15 via the in-vehicle network 16. In the above-described control processing, a machine-learned discriminator may be used.

[0061] When the vehicle 10 is under manual driving, the control unit 232 generates a manual steering signal, a manual driving signal, and a manual braking signal based on information representing the steering angle of the steering wheel 1a, information representing the operation amount of the accelerator pedal 1b, and information representing the operation amount of the brake pedal 1c. The control unit 232 outputs the manual steering signal, the manual driving signal, and the manual braking signal to the steering device 13, the driving device 14, or the braking device 15 via the in-vehicle network 16. The manual steering signal, the manual driving signal, and the manual braking signal are also notified to the collection unit 233.

[0062] Even when the vehicle 10 is under autonomous driving, the control unit 232 generates the automatic steering signal, the automatic driving signal, and the automatic braking signal and notifies them to the collection unit 233. However, the automatic steering signal, the automatic driving signal, and the automatic braking signal are not output to the steering device 13, the driving device 14, or the braking device 15.

[0063] The detection unit 231 and the collection unit 233 execute processing using common hardware. The detection unit 231 and the collection unit 233 share at least some hardware (such as a processor, memory, etc.). Therefore, reducing the processing amount of the detection unit 231 makes it possible to increase the processing amount in the collection unit 233.

[0064] In this embodiment, the detection unit 231, the control unit 232, the collection unit 233, the estimation unit 234, and the determination unit 235 execute processing using common hardware.

[0065] In FIG. 2, the map information storage device 11 and the control device 12 are described as separate devices, but these devices may be configured as one device.

[0066] FIG. 3 is an example of an operation flowchart of information processing of the control device 12. Hereinafter, the information processing of the control device 12 will be described with reference to FIG. 3. The control device 12 executes information processing according to the operation flowchart shown in FIG. 3 at an information processing time having a predetermined cycle.

[0067] First, the estimation unit 234 acquires vehicle information, environment information, and terrain information (step S101). The vehicle information, environment information, and terrain information are input to the control device 12 via the in-vehicle network 16.

[0068] Next, the estimation unit 234 estimates the degree of the processing amount of the detection processing in the detection unit 231 based on at least one of the vehicle information, the environment information, and the terrain information (step S102). The estimation unit 234 may estimate the degree of the processing amount of the detection processing in the detection unit 231 based on a plurality of the vehicle information, the environment information, and the terrain information. The estimation processing of the estimation unit 234 will be described later.

[0069] Next, the determination unit 235 determines whether the processing amount of the detection processing in the detection unit 231 estimated by the estimation unit 234 is estimated to be low or high (step S103).

[0070] When it is estimated by the estimation unit 234 that the processing amount of the detection process is low (step S103 - Yes), the determination unit 235 determines to relax the detection accuracy in the detection unit 231 more than when it is estimated that the processing amount of the detection process in the detection unit 231 is high (step S104), and ends the series of processes.

[0071] When it is determined by the determination unit 235 to relax the detection accuracy in the detection unit 231, the detection unit 231 relaxes the detection accuracy from the default accuracy and executes the detection process.

[0072] On the other hand, when it is estimated by the estimation unit 234 that the processing amount of the detection process is high (not low) (step S103 - No), the determination unit 235 determines not to relax the detection accuracy in the detection unit 231 (step S105), and ends the series of processes. The determination process by the determination unit 235 will be described later.

[0073] When it is determined by the determination unit 235 not to relax the detection accuracy in the detection unit 231, the detection unit 231 executes the detection process with the default detection accuracy.

[0074] Next, the estimation process by the above-described estimation unit 234 will be described below.

[0075] The vehicle information includes information representing the degree of operation of the vehicle 10. For example, the vehicle information includes the speed of the vehicle 10, the change amount of the steering angle per unit time, and the number of braking times per unit time, etc.

[0076] The estimation unit 234 estimates that when the operation of the vehicle 10 is slow, the processing amount of the detection process in the detection unit 231 is lower than when the operation of the vehicle 10 is fast.

[0077] When the speed of the vehicle 10 is less than the reference speed, the estimation unit 234 estimates that the processing amount of the detection process in the detection unit 231 is lower than when the speed of the vehicle 10 is greater than or equal to the reference speed. When the speed of the vehicle 10 is less than the reference speed, since there are few changes in the environment included in the field of view of the sensor, it is estimated that the number of objects and road features detected by the detection unit 231 per unit time is small. As the speed of the vehicle 10, the average speed of the vehicle 10 over a recent predetermined period can be used.

[0078] When the amount of change in the steering angle per unit time is less than the reference angle, the estimation unit 234 estimates that the processing amount of the detection process in the detection unit 231 is lower than when the amount of change in the steering angle per unit time is greater than or equal to the reference angle. When the amount of change in the steering angle per unit time is less than the reference angle, since there are few changes in the environment represented in the field of view of the sensor, it is estimated that the number of objects and road features detected by the detection unit 231 per unit time is small.

[0079] When the number of braking times per unit time is less than the reference value, the estimation unit 234 estimates that the processing amount of the detection process in the detection unit 231 is lower than when the number of braking times per unit time is greater than or equal to the reference value. When the number of braking times per unit time is less than the reference value, it is estimated that there are few objects around the vehicle 10 or the vehicle 10 is traveling on a straight road. Therefore, when the number of braking times per unit time is less than the reference value, since there are few changes in the environment represented in the field of view of the sensor, it is estimated that the number of objects and road features detected by the detection unit 231 per unit time is small.

[0080] The environmental information includes information representing the degree of complexity of the environment around the vehicle 10. For example, the environmental information includes the number of moving objects around the vehicle 10, the number of road features around the vehicle 10, and the presence or absence of a construction section around the vehicle 10.

[0081] When the environment around the vehicle 10 is not complex, the estimation unit 234 estimates that the processing amount of the detection process in the detection unit 231 is lower than when the environment around the vehicle 10 is complex.

[0082] When the number of moving objects around the vehicle 10 is less than the reference value, the estimation unit 234 estimates that the processing amount of the detection process in the detection unit 231 is lower than when the number of moving objects around the vehicle 10 is greater than or equal to the reference value. When the number of moving objects around the vehicle 10 is less than the reference value, it is estimated that the number of moving objects detected by the detection unit 231 per unit time is small. For example, when the vehicle 10 is traveling on a road without traffic congestion, the estimation unit 234 estimates that the processing amount of the detection process in the detection unit 231 is lower than when the vehicle 10 is traveling on a congested road.

[0083] When the number of road feature objects around the vehicle 10 is less than the reference value, the estimation unit 234 estimates that the processing amount of the detection process in the detection unit 231 is lower than when the number of road feature objects around the vehicle 10 is greater than or equal to the reference value. When the number of road feature objects around the vehicle 10 is less than the reference value, it is estimated that the number of road feature objects detected by the detection unit 231 per unit time is small.

[0084] When there is no construction section around the vehicle 10, the estimation unit 234 estimates that the processing amount of the detection process in the detection unit 231 is lower than when there is a construction section around the vehicle 10. When there is a construction section around the vehicle 10, since there are displays of lane restrictions associated with the construction, etc., it is estimated that the number of road feature objects detected by the detection unit 231 per unit time is large.

[0085] The terrain information includes information representing the degree of complexity of the terrain including the current position of the vehicle 10. For example, the terrain information includes the number of lane links around the vehicle 10, the curvature of the road, and the type of the road, etc.

[0086] When the terrain around the vehicle 10 is not complex, the estimation unit 234 estimates that the processing amount of the detection process in the detection unit 231 is lower than when the terrain around the vehicle 10 is complex.

[0087] When the number of lane links around the vehicle 10 is less than the reference value, the estimation unit 234 estimates that the processing amount of the detection process in the detection unit 231 is lower than when the number of lane links around the vehicle 10 is greater than or equal to the reference value. Near a road or intersection with a large number of lanes, the number of lane links around the vehicle 10 increases. When the number of lane links around the vehicle 10 is less than the reference value, it is estimated that the number of road features detected by the detection unit 231 per unit time is small.

[0088] When the curvature of the road around the vehicle 10 is less than the reference value, the estimation unit 234 estimates that the processing amount of the detection process in the detection unit 231 is lower than when the curvature of the road around the vehicle 10 is greater than or equal to the reference value. When the curvature of the road around the vehicle 10 is less than the reference value, it is estimated that the vehicle 10 is traveling on a relatively straight road. When the vehicle 10 is traveling on a relatively straight road, since there are few changes in the environment represented in the sensor's field of view, it is estimated that there are few changes in the objects and road features detected by the detection unit 231 per unit time.

[0089] When the vehicle 10 is traveling on an expressway, the estimation unit 234 estimates that the processing amount of the detection process in the detection unit 231 is lower than when the vehicle 10 is not traveling on an expressway (when traveling on a general road). On a general road, since there are many moving objects such as pedestrians and bicycles, stationary objects such as parked vehicles, and road features such as signs, it is estimated that the detection process is more than on an expressway.

[0090] The estimation unit 234 may estimate the degree of the processing amount of the detection process in the detection unit 231 based on a plurality of pieces of information among the vehicle information, the environment information, and the terrain information. For example, using a table in which the speed of the vehicle 10, which is vehicle information, the number of moving objects, which is environment information, and the degree of the processing amount of the detection process are associated, the degree of the processing amount of the detection process is estimated based on the speed and the number of moving objects.

[0091] Next, the determination process by the above-described determination unit 235 will be described below.

[0092] For example, the determination unit 235 determines to relax the detection accuracy in the detection unit 231 by increasing the period for executing the detection process in the detection unit 231. The determination unit 235 increases the period for the front camera or the rear camera to acquire an image. Further, the determination unit 235 increases the period for the LiDAR sensor, the millimeter-wave radar sensor, or the ultrasonic sensor to acquire reflected wave information.

[0093] Also, the determination unit 235 may determine to relax the detection accuracy in the detection unit 231 by reducing the number of sensors through which the detected information is input to the detection unit 231.

[0094] As described above, the sensor group 3 includes a front sensor, a surrounding sensor, a rear sensor, and a side sensor. Priorities are set for these sensors. The priority of the front sensor is 1, the priority of the surrounding sensor is 2, the priority of the rear sensor is 3, and the priority of the side sensor is 1. The higher the numerical value, the higher the priority.

[0095] The number of sensors through which the detected information is input to the detection unit 231 may be reduced so that only the information detected by the sensor with a higher priority is input to the detection unit 231. For example, when relaxing the detection accuracy in the detection unit 231, only the information detected by the front sensor may be input to the detection unit 231.

[0096] Also, the determination unit 235 may determine to relax the detection accuracy in the detection unit 231 by reducing the area within the image detected by the detection unit 231. Here, examples of the image include an image acquired by the front camera or the rear camera. For example, the detection accuracy in the detection unit 231 may be relaxed by setting only the central area of the image as the detection target.

[0097] Further, the determination unit 235 may determine to relax the detection accuracy in the detection unit 231 by reducing the distance of the trajectory of the moving object estimated in the detection unit 231. For example, the process of estimating the trajectory of the moving object up to 5 seconds ahead may be changed to the process of estimating the trajectory of the moving object up to 2 seconds ahead.

[0098] FIG. 4 is a diagram for explaining another determination process of the determination unit. In this example, the detection unit 231 includes a first detection unit 2311 and a second detection unit 2312 that has a lower detection accuracy than the first detection unit 2311 and a lower hardware load during operation. Each of the first detection unit 2311 and the second detection unit 2312 can independently execute a detection process.

[0099] The first detection unit 2311 and the second detection unit 2312 have an identifier including a deep neural network (DNN), but the number of intermediate layers of the DNN in the first detection unit 2311 is larger than that of the DNN in the second detection unit 2312.

[0100] When the determination unit 235 estimates that the processing amount of the detection process in the detection unit 231 estimated by the estimation unit 234 is low, the determination unit 235 determines to use the second detection unit 2312, and when the determination unit 235 estimates that the processing amount of the detection process in the detection unit 231 estimated by the estimation unit 234 is high, the determination unit 235 determines to use the first detection unit 2311.

[0101] FIG. 5 is an example of an operation flowchart of the collection process of the control device 12. Hereinafter, the collection process of the control device 12 will be described with reference to FIG. 5. When the vehicle 10 is manually driven, the control device 12 executes a collection process according to the operation flowchart shown in FIG. 5 at a collection process time having a predetermined cycle.

[0102] First, the collection unit 233 acquires vehicle information, environment information, and terrain information (step S201). The vehicle information, environment information, and terrain information are input to the control device 12 via the in-vehicle network 16. The collection unit 233 may collect the vehicle information, environment information, and terrain information over a predetermined collection period in a lump.

[0103] Next, the collection unit 233 acquires a manual steering signal, a manual drive signal, a manual braking signal, an automatic steering signal, an automatic drive signal, and an automatic braking signal (step S202). The manual steering signal, the manual drive signal, the manual braking signal, the automatic steering signal, the automatic drive signal, and the automatic braking signal are notified from the control unit 232 to the collection unit 233. In step S202, the manual steering signal, the manual drive signal, the manual braking signal, the automatic steering signal, the automatic drive signal, and the automatic braking signal over a predetermined collection period may be collected.

[0104] The process of step S202 may be performed before step S201. Also, the process of step S202 may be performed simultaneously with step S201.

[0105] Also, in step S201 or S202, when it is determined by the determination unit 235 to relax the detection accuracy in the detection unit 231, the collection unit 233 may increase the amount of processing for collecting information related to the control of the vehicle 10 for machine learning. For example, the collection unit 233 may shorten the period for executing the collection process. Thereby, teacher data with high resolution can be obtained.

[0106] Next, the collection unit 233 transmits driving information including vehicle information, environment information, terrain information, a manual steering signal, a manual drive signal, a manual braking signal, an automatic steering signal, an automatic drive signal, and an automatic braking signal to an external server via the communication device 2 (step S203), and ends a series of processes. In the external server, the vehicle information, the environment information, the terrain information, the manual steering signal, the manual drive signal, and the manual braking signal are used as teacher data for training an identifier. Also, the automatic steering signal, the automatic drive signal, and the automatic braking signal may be used to compare the current automatic driving control with the manual driving by comparing them with the manual steering signal, the manual drive signal, and the manual braking signal.

[0107] In the collection process shown in FIG. 5, the driving information collected during a predetermined collection period may be summarized and transmitted to the server. Also, in the collection process shown in FIG. 5, the operation signals of the steering wheel 1a, accelerator pedal 1b, and brake pedal 1c by the driver may be collected and transmitted to the server.

[0108] The collection process shown in FIG. 5 may also be performed when the vehicle 10 is in autonomous driving. For example, when the vehicle 10 is in autonomous driving, the driver may intervene in the operation. The operation data in which the driver intervenes during autonomous driving can also be collected as teacher data.

[0109] According to the determination device of the present embodiment described in detail above, when the processing amount of the detection process in the detection unit is low, the hardware load can be reduced by relaxing the detection accuracy in the detection unit, so that the collection process of the collection unit can be improved.

[0110] In the present invention, the information processing device, information processing computer program, and information processing method of the above-described embodiment can be appropriately changed as long as they do not depart from the gist of the present invention. Further, the technical scope of the present invention is not limited to those embodiments, and extends to the invention described in the claims and its equivalents.

[0111] For example, the collection unit may collect information acquired from other vehicles around the vehicle via a communication device. Examples of such information include events in front of the traveling direction of the vehicle. Examples of the event include traffic jam information or another vehicle stopped ahead of a blind curve. The collection unit may also transmit the information acquired by other vehicles to an external server.

Explanation of Signs

[0112] 1a Steering wheel 1b Accelerator pedal 1c Brake pedal 2 Communication device 3 Sensor group 4 Positioning information receiver 5 Navigation device 10 Vehicle 11 Map information storage device 12 Control device 21 Communication interface 22 Memory 23 Processor 231 Detection unit 232 Control unit 233 Collection unit 234 Estimation unit 235 Decision unit 24 Signal line 13 Steering device 14 Driving device 15 Braking device 16 In-vehicle network

Claims

1. A collection unit that collects information related to the control of a vehicle for machine learning, the collection unit being processed using hardware common to a detection unit that executes a detection process for controlling the vehicle, an estimation unit that estimates the degree of the processing amount of the detection process in the detection unit based on at least one of vehicle information representing the state of the vehicle, environment information representing the environment around the vehicle, and terrain information representing the terrain including the current position of the vehicle, a determination unit that, when it is estimated by the estimation unit that the processing amount of the detection process in the detection unit is low, determines to relax the detection accuracy in the detection unit more than when it is estimated that the processing amount of the detection process in the detection unit is high, An information processing apparatus comprising the above, characterized in that.

2. The vehicle information includes information representing the degree of operation of the vehicle, and the estimation unit estimates that the processing amount of the detection process in the detection unit is lower when the operation of the vehicle is slow than when the operation of the vehicle is fast. The information processing apparatus according to claim 1.

3. The environment information includes information representing the degree of complexity of the environment around the vehicle, and the estimation unit estimates that the processing amount of the detection process in the detection unit is lower when the environment around the vehicle is not complex than when the environment around the vehicle is complex. The information processing apparatus according to claim 1.

4. The terrain information includes information representing the degree of complexity of the terrain including the current position of the vehicle, and the estimation unit estimates that the processing amount of the detection process in the detection unit is lower when the terrain including the current position of the vehicle is not complex than when the terrain including the current position of the vehicle is complex. The information processing apparatus according to claim 1.

5. The determination unit according to claim 1 determines to relax the detection accuracy in the detection unit by increasing the period for executing the detection process in the detection unit.

6. The determination unit according to claim 1 determines to relax the detection accuracy in the detection unit by reducing the number of sensors to which the detected information is input to the detection unit.

7. The determination unit according to claim 1 determines to relax the detection accuracy in the detection unit by reducing the area in the image detected by the detection unit.

8. The information processing apparatus according to claim 1, wherein the determination unit determines to relax the detection accuracy in the detection unit by reducing the distance of the trajectory of the moving object estimated in the detection unit.

9. The detection unit includes a first detection unit capable of independently executing a detection process, and a second detection unit having a lower detection accuracy than the first detection unit and a lower hardware load during operation. When it is estimated that the processing amount of the detection process in the detection unit estimated by the estimation unit is low, the determination unit uses the second detection unit. When it is estimated that the processing amount of the detection process in the detection unit estimated by the estimation unit is high, the determination unit determines to use the first detection unit. The information processing apparatus according to claim 1.

10. When it is determined by the determination unit to relax the detection accuracy in the detection unit, the collection unit increases the processing amount of collecting information related to the control of the vehicle for machine learning. The information processing apparatus according to any one of claims 1 to 9.

11. Collecting information related to the control of the vehicle for machine learning, collecting information using hardware common to a detection unit that executes a detection process for controlling the vehicle, estimating the degree of the processing amount of the detection process in the detection unit based on at least one of vehicle information representing the state of the vehicle, environmental information representing the environment around the vehicle, and terrain information representing the terrain including the current position of the vehicle, when it is estimated that the processing amount of the detection process in the detection unit is low, determining to relax the detection accuracy in the detection unit more than when it is estimated that the processing amount of the detection process in the detection unit is high, causing a processor to execute a process including the above. A computer program for information processing, characterized by this.

12. An information processing apparatus collecting information related to the control of the vehicle for machine learning, collecting information using hardware common to a detection unit that executes a detection process for controlling the vehicle, estimating the degree of the processing amount of the detection process in the detection unit based on at least one of vehicle information representing the state of the vehicle, environmental information representing the environment around the vehicle, and terrain information representing the terrain including the current position of the vehicle, when it is estimated that the processing amount of the detection process in the detection unit is low, determining to relax the detection accuracy in the detection unit more than when it is estimated that the processing amount of the detection process in the detection unit is high, An information processing method characterized by executing a process including

Images