JP7615938B2 - Information processing server, processing method for information processing server, and program - Google Patents

Description

The present invention relates to an information processing server, a processing method for the information processing server, and a program.

Regarding the processing of information related to vehicle travel, for example, JP 2020-052607 A is known. This publication discloses a system that associates and stores unstable behavior information at a behavior occurrence position where the target vehicle exhibits unstable behavior with a determination result as to whether the unstable behavior is caused by the driver.

JP 2020-052607 A

When assessing unstable behavior as described above, it is considered to use a vehicle judgment model generated from various information about the target vehicle. However, as time passes after the vehicle judgment model is generated, the vehicle judgment model may deviate from the actual market environment, and the judgment accuracy may decrease. On the other hand, uniformly increasing the update frequency of the vehicle judgment model would lead to an increase in processing load and operating costs.

One aspect of the present invention is an information processing server that obtains target vehicle data including the driving state of a target vehicle and position information of the target vehicle on a map from a plurality of target vehicles, the information processing server including: an unstable behavior position recognition unit that recognizes an unstable behavior position, which is a position on the map where the target vehicle has become unstable, based on the target vehicle data; a model generation unit that generates a behavior discrimination model from target vehicle data of the target vehicle at the unstable behavior position for each of a plurality of set areas set on the map; an assistance target vehicle determination unit that determines whether or not there is an assistance target vehicle that is a target vehicle that is estimated to become unstable at the unstable behavior position, based on the unstable behavior position, the target vehicle data, and the behavior discrimination model corresponding to the set area including the unstable behavior position; a vehicle assistance unit that provides vehicle assistance to the assistance target vehicle when it is determined that there is an assistance target vehicle by the assistance target vehicle determination unit; a statistical value calculation unit that calculates a vehicle traffic statistical value in the set area or an unstable behavior statistical value in the set area based on the target vehicle data; The system includes a model update necessity determination unit that determines whether or not it is necessary to update the behavior discrimination model for a set area based on changes in traffic statistics or changes in unstable behavior statistics since the behavior discrimination model was generated in the set area, and a model updating unit that updates the behavior discrimination model for the set area when the model update necessity determination unit determines that it is necessary to update the behavior discrimination model for the set area, wherein unstable behavior is behavior of the target vehicle that makes the driving of the target vehicle unstable, and unstable behavior includes slipping of the target vehicle, and the support target vehicle determination unit determines whether or not there is a candidate vehicle that is a target vehicle that will reach the unstable behavior position within a certain period of time, based on the unstable behavior position and position information of the target vehicle included in the target vehicle data, and when it is determined that there is such a candidate vehicle, the support target vehicle determination unit determines whether or not there is an support target vehicle that is a candidate vehicle that will cause unstable behavior at the unstable behavior position, by using a behavior discrimination model corresponding to the set area including the unstable behavior position, based on the target vehicle data of the candidate vehicle .

In an information processing server according to one aspect of the present invention, a vehicle traffic statistical value in a set area is calculated based on target vehicle data, and a behavior discrimination model in the set area is determined to need to be updated based on a change in the vehicle traffic statistical value since the behavior discrimination model was generated in the set area. Alternatively, the information processing server calculates an unstable behavior statistical value in the set area based on target vehicle data, and a behavior discrimination model in the set area is determined to need to be updated based on a change in the unstable behavior statistical value since the behavior discrimination model was generated in the set area. Therefore, compared to a case where a model is updated periodically without considering the situation in the set area, the information processing server can update the behavior discrimination model at an appropriate timing according to the situation in the set area. In addition, the information processing server does not make a determination based only on the latest vehicle traffic statistical value or unstable behavior statistical value, but determines whether or not a model update is necessary on a ratio basis by focusing on the change since the behavior discrimination model was generated in the set area, so that the behavior discrimination model can be updated taking into consideration the change in the situation in the set area since the behavior discrimination model was generated.

In an information processing server according to one embodiment of the present invention, the model update necessity determination unit may determine that updating of the behavior discrimination model of a set area is necessary if an amount of change in the latest vehicle traffic statistics value of the set area relative to the vehicle traffic statistics value when the behavior discrimination model was generated in the set area is equal to or greater than a first model update threshold, or if an amount of change in the latest unstable behavior statistics value of the set area relative to the unstable behavior statistics value when the behavior discrimination model was generated in the set area is equal to or greater than a second model update threshold.
According to this information processing server, when the amount of change in the latest vehicle traffic statistics for a set area relative to the vehicle traffic statistics when a behavior discrimination model was generated in the set area is equal to or greater than a first model update threshold, or when the amount of change in the latest unstable behavior statistics for a set area relative to the unstable behavior statistics when a behavior discrimination model was generated in the set area is equal to or greater than a second model update threshold, it is determined that an update of the behavior discrimination model for the set area is necessary. Therefore, by setting the threshold, the behavior discrimination model can be updated at an appropriate timing according to the situation in the set area.

In an information processing server according to one embodiment of the present invention, the statistical value calculation unit may calculate a vehicle traffic calculation value for each divided area obtained by dividing a set area into a plurality of areas based on the target vehicle data, and calculate a vehicle traffic statistical value for the set area from the number of divided areas in which a difference between the vehicle traffic calculation value when a behavior discrimination model was generated in the set area and the latest vehicle traffic calculation value is equal to or greater than a third model update threshold, or may calculate an unstable behavior calculation value for each divided area obtained by dividing the set area into a plurality of areas based on the target vehicle data, and calculate an unstable behavior statistical value for the set area from the number of divided areas in which a difference between the unstable behavior calculation value when a behavior discrimination model was generated in the set area and the latest unstable behavior calculation value is equal to or greater than a fourth model update threshold.
According to this information processing server, the statistical values of the set area are calculated based on the conditions of each divided area into which the set area is divided, so that even if extreme values occur in part of the set area, they can be gradually reflected in the statistical values.

In an information processing server according to one embodiment of the present invention, the statistical value calculation unit may calculate a vehicle traffic calculation value for each road in a set area based on target vehicle data, and calculate a vehicle traffic statistical value for the set area from the number of roads for which the difference between the vehicle traffic calculation value when the behavior discrimination model was generated in the set area and the latest vehicle traffic calculation value is equal to or greater than a fifth model update threshold, or may calculate an unstable behavior calculation value for each road in the set area based on the target vehicle data, and calculate an unstable behavior statistical value for the set area from the number of roads for which the difference between the unstable behavior calculation value when the behavior discrimination model was generated in the set area and the latest unstable behavior calculation value is equal to or greater than a fourth model update threshold.
According to this information processing server, the statistical values for the set area are calculated based on the conditions of each road within the set area, so that even if extreme values occur on some roads within the set area, this can be gradually reflected in the statistical values.

Another aspect of the present invention is a processing method of an information processing server for obtaining target vehicle data including driving conditions of the target vehicles and position information of the target vehicles on a map from a plurality of target vehicles, the method including an unstable behavior position recognition step for recognizing an unstable behavior position, which is a position on the map where the target vehicle has become unstable, based on the target vehicle data; a model generation step for generating a behavior discrimination model from the target vehicle data of the target vehicle at the unstable behavior position for each set area set on the map; an assistance target vehicle determination step for determining whether or not there is an assistance target vehicle, which is a target vehicle that is estimated to become unstable at the unstable behavior position, based on the unstable behavior position, the target vehicle data, and the behavior discrimination model corresponding to the set area including the unstable behavior position; a vehicle assistance step for providing vehicle assistance to the assistance target vehicle when it is determined that there is an assistance target vehicle in the assistance target vehicle determination step; a statistical value calculation step for calculating vehicle traffic statistics in the set area or unstable behavior statistics in the set area based on the target vehicle data; and and a model updating step of updating the behavior discrimination model for the set area when it is determined in the model updating necessity determination step that it is necessary to update the behavior discrimination model for the set area. The unstable behavior is behavior of the target vehicle that makes the traveling of the target vehicle unstable, and the unstable behavior includes slipping of the target vehicle. In the support target vehicle determination step, it is determined whether or not there is a candidate vehicle that is a target vehicle that will reach the unstable behavior position within a certain time based on the unstable behavior position and position information of the target vehicle included in the target vehicle data, and when it is determined that the candidate vehicle exists, it is determined whether or not there is an support target vehicle that is a candidate vehicle that will cause unstable behavior at the unstable behavior position, based on the target vehicle data of the candidate vehicle, by using a behavior discrimination model corresponding to the set area including the unstable behavior position.

In another aspect of the present invention, the processing method of the information processing server calculates a vehicle traffic statistical value in a set area based on target vehicle data, and determines whether or not the behavior discrimination model of the set area needs to be updated based on the change in the vehicle traffic statistical value since the behavior discrimination model was generated in the set area. Alternatively, in this processing method of the information processing server, an unstable behavior statistical value in a set area is calculated based on target vehicle data, and determines whether or not the behavior discrimination model of the set area needs to be updated based on the change in the unstable behavior statistical value since the behavior discrimination model was generated in the set area. Therefore, according to this processing method of the information processing server, the behavior discrimination model can be updated at an appropriate timing according to the situation of the set area, compared to a case where the model is updated periodically without considering the situation of the set area. In addition, in this processing method of the information processing server, the determination is not made based only on the latest vehicle traffic statistical value or unstable behavior statistical value, but the necessity of model update is determined on a ratio basis by focusing on the change since the behavior discrimination model was generated in the set area, so that the behavior discrimination model can be updated taking into consideration the change in the situation of the set area since the behavior discrimination model was generated.

Yet another aspect of the present invention is a program for operating an information processing server which obtains target vehicle data including driving conditions of the target vehicles and position information of the target vehicles on a map from a plurality of target vehicles, the program including an unstable behavior position recognition unit which recognizes an unstable behavior position which is a position on the map where the target vehicle has become unstable based on the target vehicle data, a model generation unit which generates a behavior discrimination model from the target vehicle data of the target vehicle at the unstable behavior position for each set area set on the map, an assistance target vehicle determination unit which determines whether or not there is an assistance target vehicle which is a target vehicle which is estimated to become unstable at the unstable behavior position based on the unstable behavior position, the target vehicle data, and the behavior discrimination model corresponding to the set area including the unstable behavior position, a vehicle assistance unit which provides vehicle assistance to the assistance target vehicle when it is determined that the assistance target vehicle exists by the assistance target vehicle determination unit, and a statistical value calculation unit which calculates a vehicle traffic statistical value in the set area or an unstable behavior statistical value in the set area based on the target vehicle data, and a vehicle traffic discrimination model corresponding to the set area including the unstable behavior position. and a model update necessity determination unit which determines whether or not it is necessary to update the behavior discrimination model for the set area based on a change in the statistical value or a change in the unstable behavior statistical value since the behavior discrimination model was generated in the set area, and operates the information processing server as a model update unit which updates the behavior discrimination model for the set area when the model update necessity determination unit determines that it is necessary to update the behavior discrimination model for the set area , wherein unstable behavior is behavior of the target vehicle that makes the driving of the target vehicle unstable, and unstable behavior includes slippage of the target vehicle, and the assisted vehicle determination unit determines whether or not there is a candidate vehicle that is a target vehicle that will reach the unstable behavior position within a certain period of time, based on the unstable behavior position and position information of the target vehicle included in the target vehicle data, and when it is determined that there is such a candidate vehicle, it determines whether or not there is an assisted vehicle that is a candidate vehicle that will cause unstable behavior at the unstable behavior position, based on the target vehicle data of the candidate vehicle, by using a behavior discrimination model corresponding to the set area which includes the unstable behavior position .

In a program according to yet another aspect of the present invention, a vehicle traffic statistical value in a set area is calculated based on target vehicle data, and based on the change in the vehicle traffic statistical value since the behavior discrimination model was generated in the set area, it is determined whether or not the behavior discrimination model in the set area needs to be updated. Alternatively, in this program, an unstable behavior statistical value in the set area is calculated based on target vehicle data, and based on the change in the unstable behavior statistical value since the behavior discrimination model was generated in the set area, it is determined whether or not the behavior discrimination model in the set area needs to be updated. Therefore, according to this program, compared to a case where the model is updated periodically without considering the situation in the set area, it is possible to update the behavior discrimination model at an appropriate timing according to the situation in the set area. In addition, this program does not make a determination based only on the latest vehicle traffic statistical value or unstable behavior statistical value, but focuses on the change since the behavior discrimination model was generated in the set area and determines whether or not the model needs to be updated on a ratio basis, so that the behavior discrimination model can be updated taking into consideration the change in the situation in the set area since the behavior discrimination model was generated.

According to each aspect of the present invention, the behavior discrimination model can be updated at an appropriate timing according to the situation in the set area.

FIG. 2 is a diagram illustrating an information processing server and a target vehicle according to an embodiment. FIG. 1 is a diagram illustrating an example of information processing. FIG. 2 is a block diagram showing an example of a configuration of a target vehicle. FIG. 2 is a block diagram showing an example of a configuration of an information processing server. 1A is a diagram for explaining an example of a continuous occurrence state, and FIG. 1B is a diagram for explaining an example of a non-continuous occurrence state. FIG. 11 is a diagram for explaining an example of a set area and a divided area. FIG. 2 is a diagram for explaining an example of a road. 13 is a flowchart showing an example of an unstable behavior position storage process. 10A is a flowchart illustrating an example of a behavior discrimination model generation process, and FIG. 10B is a flowchart illustrating an example of a vehicle assistance execution process. 13 is a flowchart illustrating an example of a model updating process. 13A is a flowchart showing an example of a statistical value calculation process using a divided area, and FIG. 13B is a flowchart showing another example of a statistical value calculation process using a road.

The following describes an embodiment of the present invention with reference to the drawings.

Figure 1 is a diagram showing an information processing server 10 and a target vehicle 2 according to the first embodiment. As shown in Figure 1, the information processing server 10 is communicably connected to target vehicles 2 (2A to 2Z) via a network N. The network N is a wireless communication network. The target vehicle 2 refers to a vehicle that is a target of information collection by the information processing server 10. The target vehicle 2 includes vehicles that are the target of support for which various types of support are provided by the information processing server 10. When describing the target vehicles 2 individually, the target vehicles 2A to 2Z are used.

Figure 2 is a diagram for explaining an example of information processing. As shown in Figure 2, when the target vehicle 2A slips due to road surface freezing or the like, the target vehicle 2A transmits target vehicle data including the unstable behavior position D, which is the position where the slip occurred, to the information processing server 10. The information processing server 10 notifies the target vehicle 2B traveling behind the target vehicle 2A of information on the unstable behavior position. This enables the target vehicle 2B to suppress the occurrence of slip of the target vehicle 2B at the unstable behavior position D. The unstable behavior position will be described in detail later.

[Target vehicle configuration]
First, the configuration of the target vehicle 2 will be described. The target vehicle 2 is assigned an ID (vehicle identification number) for identifying the vehicle. The target vehicle 2 may be one, two or more, several tens or more, or several hundred or more. The target vehicles 2 do not need to be vehicles having the same configuration, and may be of different models. The target vehicle 2 may be an autonomous vehicle having an autonomous driving function, or may be a vehicle without an autonomous driving function.

The target vehicle 2 will be described below with reference to FIG. 3. FIG. 3 is a block diagram showing an example of the configuration of the target vehicle 2. Here, the target vehicle 2 will be described as an autonomous vehicle.

As shown in FIG. 3, the target vehicle 2 is equipped with an autonomous driving ECU 30. The autonomous driving ECU 30 is an electronic control unit having a CPU, ROM, RAM, etc. In the autonomous driving ECU 30, for example, a program stored in the ROM is loaded into the RAM, and the program loaded into the RAM is executed by the CPU to realize various functions. The autonomous driving ECU 30 may be composed of multiple electronic units.

The autonomous driving ECU 30 is connected to a GPS [Global Positioning System] receiver 21, an external sensor 22, an internal sensor 23, a driving operation detector 24, a map database 25, a communication unit 26, an HMI [Human Machine Interface] 27, and an actuator 28.

The GPS receiver 21 measures the position of the target vehicle 2 (e.g., the latitude and longitude of the target vehicle 2) by receiving signals from three or more GPS satellites. The GPS receiver 21 transmits the measured position information of the target vehicle 2 to the autonomous driving ECU 30.

The external sensor 22 is a detection device that detects the external environment of the target vehicle 2. The external sensor 22 includes at least one of a camera and a radar sensor.

The camera is an imaging device that captures images of the external environment of the target vehicle 2. The camera is provided behind the windshield of the target vehicle 2 and captures images of the area in front of the vehicle. The camera transmits imaging information related to the external environment of the target vehicle 2 to the autonomous driving ECU 30. The camera may be a monocular camera or a stereo camera.

The radar sensor is a detection device that detects objects around the target vehicle 2 using radio waves (e.g., millimeter waves) or light. Radar sensors include, for example, millimeter wave radar or LIDAR (Light Detection and Ranging). The radar sensor detects objects by transmitting radio waves or light to the vicinity of the target vehicle 2 and receiving the radio waves or light reflected by the objects. The radar sensor transmits information about the detected objects to the autonomous driving ECU 30. Objects include fixed objects such as guardrails and buildings, as well as moving objects such as pedestrians, bicycles, and other vehicles. The external sensor 22 may include an outside air temperature sensor that detects the outside air temperature of the target vehicle 2. The external sensor 22 may include a light sensor that detects the brightness outside.

The internal sensor 23 is a detection device that detects the state of the target vehicle 2. The internal sensor 23 includes a vehicle speed sensor, an acceleration sensor, and a yaw rate sensor as sensors that detect the driving state of the target vehicle 2. The vehicle speed sensor is a detector that detects the speed of the target vehicle 2. As the vehicle speed sensor, a wheel speed sensor that is provided on the wheels of the target vehicle 2 or on a drive shaft that rotates integrally with the wheels and detects the rotational speed of each wheel can be used. The vehicle speed sensor transmits the detected vehicle speed information (wheel speed information) to the autonomous driving ECU 30.

The acceleration sensor is a detector that detects the acceleration of the target vehicle 2. The acceleration sensor includes, for example, a longitudinal acceleration sensor that detects the longitudinal acceleration of the target vehicle 2. The acceleration sensor may include a lateral acceleration sensor that detects the lateral acceleration of the target vehicle 2. The acceleration sensor transmits, for example, acceleration information of the target vehicle 2 to the autonomous driving ECU 30. The yaw rate sensor is a detector that detects the yaw rate (rotational angular velocity) around the vertical axis of the center of gravity of the target vehicle 2. A gyro sensor, for example, can be used as the yaw rate sensor. The yaw rate sensor transmits the detected yaw rate information of the target vehicle 2 to the autonomous driving ECU 30.

The internal sensor 23 detects at least one of the tire pressure, the wiper operation state, and the lighting state as the vehicle state of the target vehicle 2. The tire pressure is the air pressure of the tires of the target vehicle 2. The wiper operation state may include not only whether the wipers are operating, but also the operating speed of the wipers. The lighting state includes the lighting state of the turn signal. The lighting state may include whether the headlights are on or off, and whether the fog lights are on or off.

The internal sensor 23 may also detect the brake pressure of a hydraulic brake system from a brake pressure sensor as the vehicle state of the target vehicle 2, and may detect the on/off state of a driving assistance system (for example, a vehicle stability control system described below). The internal sensor 23 may also detect the load state of each wheel from a wheel load sensor as the vehicle state of the target vehicle 2. Additionally, the internal sensor 23 may have a fault detection unit that detects various faults in the target vehicle 2.

The driving operation detection unit 24 detects the operation of the operation unit of the target vehicle 2 by the driver. The driving operation detection unit 24 includes, for example, a steering sensor, an accelerator sensor, and a brake sensor. The operation unit of the target vehicle 2 is a device into which the driver inputs operations for driving the vehicle. The operation unit of the target vehicle 2 includes at least one of the steering unit, the accelerator operation unit, and the brake operation unit. The steering unit is, for example, a steering wheel. The steering unit is not limited to being wheel-shaped, and may be configured to function as a handle. The accelerator operation unit is, for example, an accelerator pedal. The brake operation unit is, for example, a brake pedal. The accelerator operation unit and the brake operation unit do not necessarily have to be pedals, and may be configured to allow the driver to input acceleration or deceleration.

The steering sensor detects the amount of operation of the steering unit by the driver. The amount of operation of the steering unit includes the steering angle. The amount of operation of the steering unit may include the steering torque. The accelerator sensor detects the amount of operation of the accelerator operation unit by the driver. The amount of operation of the accelerator operation unit includes, for example, the amount of depression of the accelerator pedal. The brake sensor detects the amount of operation of the brake operation unit by the driver. The amount of operation of the brake operation unit includes, for example, the amount of depression of the brake pedal. The brake sensor may be configured to detect the master cylinder pressure of the hydraulic brake system. The amount of operation of the accelerator operation unit and the brake operation unit may include the depression speed. The driving operation detection unit 24 transmits operation amount information regarding the detected amount of operation by the driver to the autonomous driving ECU 30.

The map database 25 is a database that stores map information. The map database 25 is formed, for example, in a storage device such as an HDD mounted on the target vehicle 2. The map information includes road position information, road shape information (e.g., curvature information), intersection and branch point position information, and the like. The map information may include traffic regulation information such as legal speed limits associated with the position information. The map information may include landmark information used to recognize the position of the target vehicle 2 on the map. The landmarks may include lane markings, traffic lights, guard rails, road markings, and the like. The map database 25 may be configured in a server (not limited to the information processing server 10) that can communicate with the target vehicle 2.

The communication unit 26 is a communication device that controls wireless communication with the outside of the target vehicle 2. It transmits and receives various information via the network N. The communication unit 26 transmits various information to the information processing server 10 in response to a signal from the autonomous driving ECU 30.

The HMI 27 is an interface for inputting and outputting information between the autonomous driving ECU 30 and the driver or passengers. The HMI 27 is equipped with, for example, a display and a speaker provided in the vehicle cabin. The HMI 27 outputs images from the display and sounds from the speaker in response to control signals from the autonomous driving ECU 30.

The actuators 28 are devices used to control the target vehicle 2. The actuators 28 include at least a drive actuator, a brake actuator, and a steering actuator. The drive actuator controls the amount of air supplied to the engine (throttle opening) in response to a control signal from the automatic driving ECU 30, thereby controlling the driving force of the target vehicle 2. If the target vehicle 2 is a hybrid vehicle, in addition to the amount of air supplied to the engine, a control signal from the automatic driving ECU 30 is input to a motor serving as a power source to control the driving force. If the target vehicle 2 is an electric vehicle, a control signal from the automatic driving ECU 30 is input to a motor serving as a power source to control the driving force. The motor serving as a power source in these cases constitutes the actuator 28.

The brake actuator controls the brake system in response to a control signal from the autonomous driving ECU 30, and controls the braking force applied to the wheels of the target vehicle 2. For example, a hydraulic brake system can be used as the brake system. The steering actuator controls the drive of the assist motor, which controls the steering torque of the electric power steering system, in response to a control signal from the autonomous driving ECU 30. In this way, the steering actuator controls the steering torque of the target vehicle 2.

Next, the functional configuration of the autonomous driving ECU 30 will be described. As shown in FIG. 3, the autonomous driving ECU 30 has a target vehicle data acquisition unit 31, a route generation unit 32, and an autonomous driving control unit 33. Note that some of the functions of the autonomous driving ECU 30 described below may be executed in a server (not limited to the information processing server 10) capable of communicating with the target vehicle 2.

The target vehicle data acquisition unit 31 acquires target vehicle data, which is data related to the target vehicle 2. The target vehicle data includes position information of the target vehicle 2 on a map and the driving state of the target vehicle 2. The target vehicle data may include the external environment of the target vehicle 2, and may include the route along which the target vehicle 2 is driven. The target vehicle data may include driving operation information by the driver of the target vehicle 2 and the vehicle state of the target vehicle 2. The target vehicle data acquisition unit 31 transmits the acquired target vehicle data to the information processing server 10.

The target vehicle data acquisition unit 31 has a vehicle position acquisition unit 31a, an external environment recognition unit 31b, a driving state recognition unit 31c, a driving operation information acquisition unit 31d, and a vehicle state recognition unit 31e.

The vehicle position acquisition unit 31a acquires the position information of the target vehicle 2 on the map based on the position information of the GPS receiver unit 21 and the map information of the map database 25. The vehicle position acquisition unit 31a may also acquire the position information of the target vehicle 2 by SLAM [Simultaneous Localization and Mapping] technology using the target information included in the map information of the map database 25 and the detection results of the external sensor 22. The vehicle position acquisition unit 31a may recognize the lateral position of the target vehicle 2 relative to the lane (the position of the target vehicle 2 in the lane width direction) from the positional relationship between the lane markings and the target vehicle 2, and include it in the position information. The vehicle position acquisition unit 31a may also acquire the position information of the target vehicle 2 on the map by other well-known methods.

The external environment recognition unit 31b recognizes the external environment of the target vehicle 2 based on the detection results of the external sensor 22. The external environment includes the relative positions of surrounding objects with respect to the target vehicle 2. The external environment may include the relative speed and movement direction of surrounding objects with respect to the target vehicle 2. The external environment may include types of objects such as other vehicles, pedestrians, and bicycles. The types of objects can be identified by well-known techniques such as pattern matching. The external environment may include the results of lane marking recognition (white line recognition) around the target vehicle 2. The external environment may include the outside temperature and weather.

The driving state recognition unit 31c recognizes the driving state of the target vehicle 2 based on the detection result of the internal sensor 23. The driving state includes the vehicle speed and yaw rate of the target vehicle 2. The driving state may include the acceleration of the target vehicle 2. Specifically, the driving state recognition unit 31c recognizes the vehicle speed of the target vehicle 2 based on vehicle speed information from a vehicle speed sensor. The driving state recognition unit 31c recognizes the acceleration of the target vehicle 2 based on vehicle speed information from an acceleration sensor. The driving state recognition unit 31c recognizes the orientation of the target vehicle 2 based on yaw rate information from a yaw rate sensor.

The driving operation information acquisition unit 31d acquires driving operation information of the target vehicle 2 based on the detection result of the driving operation detection unit 24. The driving operation information includes, for example, at least one of the driver's accelerator operation amount, brake operation amount, and steering amount.

If the target vehicle 2 has a personal authentication function, the driving operation information acquisition unit 31d stores the driving operation history for each personally authenticated driver. The driving operation history may be associated with the external environment and driving state of the target vehicle 2. The autonomous driving ECU 30 does not necessarily need to have the driving operation information acquisition unit 31d. In this case, the driving operation detection unit 24 is also not required.

The vehicle state recognition unit 31e recognizes the vehicle state of the target vehicle 2 based on the detection result of the internal sensor 23. The vehicle state may include tire pressure. The vehicle state may include a wiper operation state, a lighting state, and a malfunction state of the target vehicle 2. The autonomous driving ECU 30 does not necessarily have to have the vehicle state recognition unit 31e.

The trajectory generation unit 32 generates a trajectory to be used for the autonomous driving of the target vehicle 2. The trajectory generation unit 32 generates a trajectory for autonomous driving based on a preset driving route, map information, the position of the target vehicle 2 on the map, the external environment of the target vehicle 2, and the driving state of the target vehicle 2.

The driving route is the route along which the target vehicle 2 travels during autonomous driving. The course generation unit 32 determines the driving route for autonomous driving based on, for example, the destination, map information, and the position of the target vehicle 2 on the map. The driving route may be set by a well-known navigation system. The destination may be set by an occupant of the target vehicle 2, or may be automatically proposed by the autonomous driving ECU 30 or a navigation system, etc.

The course includes the path along which the vehicle travels in autonomous driving and the vehicle speed profile during autonomous driving. The route is the planned trajectory along which the vehicle is traveling in autonomous driving on the travel route. The route can be, for example, data on the change in steering angle of the target vehicle 2 according to the position on the travel route (steering angle profile). The position on the travel route is, for example, a set vertical position set at a predetermined interval (for example, 1 m) in the traveling direction of the travel route. The steering angle profile is data in which a target steering angle is associated with each set vertical position.

The path generation unit 32 generates a path for the vehicle to travel based on, for example, the travel route, map information, the external environment of the target vehicle 2, and the travel state of the target vehicle 2. For example, the path generation unit 32 generates a path so that the target vehicle 2 passes through the center (center in the lane width direction) of the lane included in the travel route.

In place of the steering angle profile, a steering torque profile in which a target steering torque is associated with each set longitudinal position may be used. Also, in place of the steering angle profile, a lateral position profile in which a target lateral position is associated with each set longitudinal position may be used. The target lateral position is a target position in the width direction of the lane. In this case, the set longitudinal position and the target lateral position may be set together as a single position coordinate.

The vehicle speed profile is data in which a target vehicle speed is associated with each set longitudinal position, for example. The set longitudinal position may be set based on the vehicle's travel time rather than distance. The set longitudinal position may be set as the position the vehicle will reach in 1 second, or the position the vehicle will reach in 2 seconds.

The route generation unit 32 generates a vehicle speed profile based on speed-related information, such as the legal speed included in the route and map information. Instead of the legal speed, a preset speed for a position or section on the map may be used. The route generation unit 32 generates a route for automated driving from the route and the vehicle speed profile. Note that the route generation method in the route generation unit 32 is not limited to the above-mentioned method, and other well-known methods may be adopted.

When the path generation unit 32 receives a notification from the information processing server 10 to change the driving path to avoid the unstable behavior position, the path generation unit 32 changes the path of the target vehicle 2 so as not to pass through the unstable behavior position. The path generation unit 32 may change the driving route (road on which the target vehicle 2 is traveling) or may change the path of the target vehicle 2 so as to move away from the unstable behavior position in the road width direction by a certain distance or more within the same road.

When the course generation unit 32 receives notification of unstable behavior position information and stable driving data from the information processing server 10, the course generation unit 32 generates a course based on the unstable behavior position information and stable driving data so that unstable behavior does not occur in the target vehicle 2. The unstable behavior position information and stable driving data will be described in detail later. The course generation unit 32 generates a course so that the driving state of the target vehicle 2 near the unstable behavior position approaches the driving state of the stable driving data. When the course generation unit 32 receives a stable driving instruction instead of a notification of stable driving data from the information processing server 10, the course generation unit 32 generates a course in accordance with the stable driving instruction. The stable driving instruction will be described in detail later.

The automatic driving control unit 33 executes automatic driving of the target vehicle 2. The automatic driving control unit 33 executes automatic driving of the target vehicle 2 based on, for example, the external environment of the target vehicle 2, the driving state of the target vehicle 2, and the route generated by the route generation unit 32. The automatic driving control unit 33 executes automatic driving of the target vehicle 2 by sending a control signal to the actuator 28.

When the autonomous driving control unit 33 receives an instruction to cancel autonomous driving from the information processing server 10, it cancels autonomous driving at the unstable behavior location. The autonomous driving control unit 33 notifies the driver via the HMI 27 of the transition to manual driving. After notifying the driver, the autonomous driving control unit 33 cancels autonomous driving and transitions to manual driving by the driver. The autonomous driving control unit 33 also notifies the driver of the unstable behavior location information received from the information processing server 10 via the HMI 27.

The target vehicle 2 does not necessarily have to be an autonomous vehicle. In this case, the ECU of the target vehicle 2 does not have to have the route generation unit 32 and the autonomous driving control unit 33. The ECU of the target vehicle 2 only needs to have an information provision unit that can notify the driver of unstable behavior position information and the like through the HMI 27. The ECU of the target vehicle 2 may have a driving assistance unit that, when receiving notification of stable driving data from the information processing server 10, provides driving assistance so that the driving state of the target vehicle 2 in the vicinity of the unstable behavior position approaches the driving state of the stable driving data. The method of driving assistance is not particularly limited, and may be information provision to the driver or vehicle driving control.

[Configuration of information processing server]
The information processing server 10 is provided in a facility such as an information management center, and is configured to be able to communicate with the target vehicle 2. Fig. 4 is a block diagram showing an example of the configuration of the information processing server 10. The information processing server 10 shown in Fig. 4 is configured as a general computer including a processor 11, a storage unit 12, a communication unit 13, and a user interface 14.

The processor 11, for example, operates an operating system to control the information processing server 10. The processor 11 is an arithmetic unit such as a CPU [Central Processing Unit] that includes a control device, an arithmetic unit, a register, etc. The processor 11 controls the memory unit 12, the communication unit 13, and the user interface 14. The memory unit 12 is configured to include at least one of a memory and a storage. The memory is a recording medium such as a ROM [Read Only Memory] or a RAM [Random Access Memory]. The storage is a recording medium such as a HDD [Hard Disk Drive].

The communication unit 13 is a communication device for communicating via the network N. A network device, a network controller, a network card, etc. may be used as the communication unit 13. The user interface 14 is a device including an output device such as a display and a speaker, and an input device such as a touch panel. Note that the information processing server 10 does not necessarily need to be installed in a facility, and may be mounted on a moving object such as a vehicle or a ship.

The information processing server 10 is also connected to a memory database 15. The memory database 15 is a database for storing unstable behavior location information and the like. The memory database 15 can be configured in the same manner as a well-known database in a HDD. The memory database 15 may also be provided in a facility away from the information processing server 10.

Next, the functional configuration of the processor 11 will be described. As shown in FIG. 4, the processor 11 has a target vehicle data recognition unit 11a, an unstable behavior position recognition unit 11b, a situation determination unit 11c, a storage processing unit 11d, a model generation unit 11e, an assistance target vehicle determination unit 11f, a vehicle assistance unit 11g, a statistical value calculation unit 11h, a model update necessity determination unit 11j, and a model update unit 11k.

The target vehicle data recognition unit 11a recognizes (acquires) target vehicle data transmitted from the target vehicle 2. The target vehicle data includes the position information of the target vehicle 2 on a map and the driving state of the target vehicle 2. The target vehicle data may include the external environment of the target vehicle 2, and may include the route along which the target vehicle 2 is driving. The target vehicle data recognition unit 11a may acquire the target vehicle data in association with a set area described below.

The unstable behavior position recognition unit 11b recognizes an unstable behavior position, which is a position on a map where the target vehicle 2 has become unstable, based on the target vehicle data acquired by the target vehicle data recognition unit 11a. Unstable behavior is vehicle behavior that makes the vehicle's driving unstable. Unstable behavior includes, for example, slipping. Unstable behavior may include sudden deceleration or a sudden change in steering angle. Unstable behavior may include lane departure of the target vehicle 2, or excessive approach of the target vehicle 2 to an object.

First, the determination of unstable behavior will be described. The unstable behavior position recognition unit 11b determines whether the target vehicle 2 has become unstable based on the target vehicle data. The unstable behavior position recognition unit 11b determines that the target vehicle 2 has become slippery, which is an unstable behavior, based on at least one of, for example, the acceleration (longitudinal acceleration and lateral acceleration) detected by the acceleration sensor, the wheel speed of each wheel detected by the wheel speed sensor, the yaw rate detected by the yaw rate sensor, the steering angle of the driver detected by the steering sensor, the amount of braking operation of the driver detected by the brake sensor, and the brake pressure of the brake pressure sensor. The master cylinder pressure of the hydraulic brake system may be used instead of the amount of braking operation of the brake sensor.

The unstable behavior position recognition unit 11b may use the activation conditions of a well-known antilock brake system [ABS: Antilock Brake System] to determine slippage. For example, an antilock brake system compares the wheel speed of each wheel with the estimated vehicle speed and activates when a wheel that is thought to be locked is identified. The estimated vehicle speed may be determined from the wheel speed of each wheel before slippage occurs, or may be determined from the change in acceleration before slippage occurs.

The unstable behavior position recognition unit 11b may use the activation conditions of a well-known vehicle stability control system [VSC: Vehicle Stability Control] or the activation conditions of a well-known traction control system [TRC: Traction Control System] to determine slippage. Traction control may also be activated when a spinning wheel is identified by comparing the wheel speed of each wheel with the estimated vehicle speed. The unstable behavior position recognition unit 11b may determine slippage of the target vehicle 2 using other well-known methods.

The unstable behavior position recognition unit 11b may determine whether the target vehicle 2 has suddenly decelerated as an unstable behavior based on the deceleration detected by the acceleration sensor. In this case, the unstable behavior position recognition unit 11b determines that the target vehicle 2 has suddenly decelerated when, for example, the absolute value of the deceleration becomes equal to or greater than the sudden deceleration threshold. The sudden deceleration threshold is a threshold value that is set in advance. In the following explanation, the threshold value refers to a threshold value that is set in advance.

The unstable behavior position recognition unit 11b may determine whether or not a sudden change in steering angle has occurred in the target vehicle 2 as an unstable behavior based on the yaw rate detected by the yaw rate sensor. In this case, the unstable behavior position recognition unit 11b determines that a sudden change in steering angle has occurred in the target vehicle 2 when, for example, the yaw rate is equal to or greater than a steering angle change threshold. Note that the tire turning angle may be used instead of the yaw rate.

When the turn signal is not on, the unstable behavior position recognition unit 11b may determine whether the target vehicle 2 has deviated from the lane, which is an unstable behavior, based on the lateral position of the target vehicle 2 or the external environment of the target vehicle 2. In this case, the unstable behavior position recognition unit 11b may determine lane departure from the lateral position of the target vehicle 2, for example. Alternatively, the unstable behavior position recognition unit 11b may determine lane departure when it recognizes from the external environment of the target vehicle 2 that the target vehicle 2 has crossed a lane dividing line.

The unstable behavior position recognition unit 11b may determine whether the target vehicle 2 has come too close to an object as an unstable behavior based on the traveling state of the target vehicle 2 and the external environment of the target vehicle 2. In this case, the unstable behavior position recognition unit 11b determines that the target vehicle 2 has come too close to an object when the vehicle speed of the target vehicle 2 is equal to or greater than the vehicle speed threshold and the time to collision [TTC: Time To Collision] between the target vehicle 2 and the object is equal to or less than the TTC threshold, because the behavior is not unstable when the target vehicle 2 is traveling at a low speed even if the distance to the object is small. Instead of the time to collision, the time headway [THW: Time Headway] or distance may be used.

The determination of whether the target vehicle 2 has become unstable may be performed each time target vehicle data is obtained, or may be performed all at once at regular intervals or for a regular period of time. The determination of whether the target vehicle 2 has become unstable may be performed while the target vehicle 2 is stopped.

Next, the recognition of the unstable behavior position will be described. The unstable behavior position is the position on the map of the target vehicle 2 when the target vehicle 2 becomes unstable. When it is determined that the target vehicle 2 becomes unstable, the unstable behavior position recognition unit 11b recognizes the unstable behavior position.

The unstable behavior position recognition unit 11b recognizes the unstable behavior position based on the position information on the map of the target vehicle 2 when it is determined that the target vehicle 2 has become unstable. The unstable behavior position is recognized separately for each lane. When the unstable behavior is a lane departure, the unstable behavior position may be the position on the driving lane before the lane departure, or may be the position on the dividing line.

The unstable behavior position may be recognized as a section or area, rather than a point on a map. In a case where the target vehicle 2 slides while slipping, the unstable behavior position recognition unit 11b may determine the start position of the slip as the unstable behavior position, or may recognize all sections through which the target vehicle 2 has moved while being determined to be slipping as unstable behavior positions. The area may be a range within a certain distance from the slipping target vehicle 2, or may be the region or area in which the target vehicle 2 is traveling. The same applies to other unstable behaviors.

The situation determination unit 11c determines whether the unstable behavior position is a continuous occurrence situation or a discontinuous occurrence situation based on the presence or absence of unstable behavior of multiple target vehicles 2 at the unstable behavior position recognized by the unstable behavior position recognition unit 11b.

The situation determination unit 11c determines whether the target vehicle 2 has passed an unstable behavior position, for example, based on the target vehicle data recognized by the target vehicle data recognition unit 11a and the unstable behavior position recognized by the unstable behavior position recognition unit 11b. When the situation determination unit 11c determines that the target vehicle 2 has passed an unstable behavior position, it determines whether the unstable behavior position is in a continuous or discontinuous state based on the presence or absence of unstable behavior of the target vehicle 2. Note that the situation determination unit 11c may perform the above determination by batch processing multiple target vehicle data for a certain period of time.

A continuous occurrence situation is a situation in which unstable behavior occurs continuously. In the case of a continuous occurrence situation, it can be considered that the possibility that unstable behavior has occurred due to individual vehicle factors of the target vehicle 2 is low, and the possibility that unstable behavior has occurred due to external factors such as the road environment is high. A discontinuous situation is a situation that is not a continuous occurrence situation. In the case of a discontinuous situation, it can be considered that the possibility that unstable behavior has occurred due to individual vehicle factors of the target vehicle 2 is high. When the situation determination unit 11c does not determine that the unstable behavior position is a continuous occurrence situation, it determines that the unstable behavior position is a discontinuous situation.

Figure 5(a) is a diagram for explaining an example of a consecutive occurrence situation. As shown in Figure 5(a), as an example, when two target vehicles 2A and 2B exhibit unstable behavior consecutively at unstable behavior position D, the situation determination unit 11c determines that the unstable behavior position is a consecutive occurrence situation. Figure 5(b) is a diagram for explaining an example of a discontinuous situation. Even if the target vehicle 2A exhibits unstable behavior at unstable behavior position D, the situation determination unit 11c may determine that the unstable behavior position is a discontinuous situation when the following target vehicle 2B passes by without exhibiting unstable behavior.

The situation determined as a continuous occurrence situation is not limited to the situation in FIG. 5(a). The situation determination unit 11c may determine that the unstable behavior position D is in a continuous occurrence situation when three target vehicles 2A to 2C have exhibited unstable behavior in succession. The situation determination unit 11c may determine that the unstable behavior position D is in a continuous occurrence situation when four or more target vehicles 2 have exhibited unstable behavior in succession. The situation determination unit 11c may determine that the unstable behavior position D is in a continuous occurrence situation when all of the multiple target vehicles 2 that pass through the unstable behavior position D within a certain time period have exhibited unstable behavior.

The situation determination unit 11c may determine that the unstable behavior position D is in a continuous occurrence state even if there is one target vehicle 2 that did not exhibit unstable behavior, if unstable behavior occurs in the target vehicles 2 before and after it. Specifically, the situation determination unit 11c may determine that the unstable behavior position D is in a continuous occurrence state even if the middle target vehicle 2B of the three target vehicles 2A to 2C passes the unstable behavior position D without exhibiting unstable behavior, but if the target vehicle 2A and the target vehicle 2C exhibit unstable behavior. Alternatively, the situation determination unit 11c may determine that the unstable behavior position D is in a continuous occurrence state even if there are multiple target vehicles 2 that did not exhibit unstable behavior, if the number of target vehicles 2 that exhibited unstable behavior within a certain period of time is equal to or greater than a threshold value.

The memory processing unit 11d stores unstable behavior position information related to the unstable behavior position recognized by the unstable behavior position recognition unit 11b in the memory database 15. When a judgment is made by the situation judgment unit 11c, the memory processing unit 11d may store the unstable behavior position and the judgment result of the situation judgment unit 11c in association with each other in the memory database 15.

The memory processing unit 11d also associates the unstable behavior position information with the target vehicle data of the target vehicle 2 that has traveled through the unstable behavior position, and stores the associated information in the memory database 15. Note that the memory processing unit 11d does not necessarily need to store the target vehicle data of all target vehicles 2 that have traveled through the unstable behavior position. The memory processing unit 11d may also associate the unstable driving data and/or stable driving data acquired by the model generation unit 11e, which will be described later, with the unstable behavior position and store the associated information in the memory database 15.

The model generation unit 11e generates a behavior discrimination model from the target vehicle data of the target vehicle 2 that has become unstable at the unstable behavior position for each of the multiple set areas set on the map. A set area is an area arbitrarily set on the map. A set area may be, for example, a region or prefecture (Tokyo or Osaka), or may be a state or county. A set area may be a city, town, village, or ward. A set area may be an area of any shape set on the map (for example, a rectangular area). Note that multiple set areas do not overlap with each other.

The behavior discrimination model is a vehicle behavior model used to determine (estimate) whether the target vehicle 2 will exhibit unstable behavior at the unstable behavior position. The behavior discrimination model includes at least one of the following feature parameters: vehicle speed, longitudinal acceleration, steering angle, yaw rate, operation duration of the ABS [Antilock Brake System], operation state of driving assistance (such as a vehicle stability control system), and maximum steering angle before the unstable behavior. The behavior discrimination model may include feature parameters related to the surrounding environment, such as outside air temperature, estimated road friction, road gradient, wiper operation state, and lighting state, and may include time or time period as a feature parameter. The behavior discrimination model may be generated to include the median value of feature parameters such as the vehicle speed of a large number of target vehicles 2 that have exhibited unstable behavior, or each feature parameter may be obtained by a predetermined calculation process. Various well-known methods for generating vehicle behavior models can be used to generate the behavior discrimination model.

The behavior discrimination model may be generated in different models depending on the judgment result of the situation judgment unit 11c. The behavior discrimination model may be generated in two ways: a model corresponding to an unstable behavior position in a continuous occurrence situation and a model corresponding to an unstable behavior position in a discontinuous situation.

Different behavior discrimination models may be generated depending on the location attributes of the unstable behavior location. The location attributes may distinguish between expressways and general roads as different attributes, and between narrow roads and general roads as different attributes depending on the road width. Various attributes that can be determined from map information can be used as location attributes.

Different behavior discrimination models may be generated depending on individual vehicle attributes such as vehicle type. Individual vehicle attributes may distinguish between large vehicles, medium-sized vehicles, and small vehicles as different attributes, or between freight vehicles and passenger cars as different attributes. Various attributes that can be determined from vehicle specifications can be used as individual vehicle attributes.

The behavior discrimination model includes at least one of an unstable behavior discrimination model and a stable behavior discrimination model. The unstable behavior discrimination model is a vehicle behavior model generated from the target vehicle data of a target vehicle 2 that has become unstable at the unstable behavior position. The stable behavior discrimination model is a vehicle behavior model generated from the target vehicle data of a target vehicle 2 that has not become unstable at the unstable behavior position.

The model generation unit 11e acquires unstable driving data for each set area from the target vehicle data of the target vehicle 2 that has become unstable at an unstable behavior position, for example. The unstable driving data is data related to the driving state of the target vehicle 2 in the past when unstable behavior occurred at the unstable behavior position. The unstable driving data can be, for example, data on the driving state of the target vehicle 2 from a position a certain distance before the unstable behavior position to the unstable behavior position. The unstable driving data may be data on the driving state of the target vehicle 2 for a certain time until the unstable behavior occurs. The unstable driving data may include the driving trajectory of the target vehicle 2. The model generation unit 11e may generate an unstable behavior discrimination model for the set area using the unstable driving data for each set area.

The model generation unit 11e may acquire stable driving data of the target vehicle 2 that passed the unstable behavior position without unstable behavior for each set area. The stable driving data is data related to the driving state of the target vehicle 2 in the past when the target vehicle 2 passed the unstable behavior position in a discontinuous situation without unstable behavior. The stable driving data may be data of the driving state of the target vehicle 2 from a position a certain distance before the unstable behavior position to the unstable behavior position, for example. The stable driving data may be data of the driving state of the target vehicle 2 for a certain time until the unstable behavior position is reached. The stable driving data may include the driving trajectory of the target vehicle 2. The model generation unit 11e may generate a stable behavior discrimination model for the set area using the stable driving data for each set area.

The support target vehicle determination unit 11f determines whether or not a support target vehicle exists based on the unstable behavior position recognized by the unstable behavior position recognition unit 11b and the target vehicle data recognized by the target vehicle data recognition unit 11a. The support target vehicle is a target vehicle 2 that is estimated to become unstable at the unstable behavior position within a certain time. The certain time may be 5 minutes or 10 minutes. The certain time may be less than 5 minutes or 30 minutes or more. The certain time is not particularly limited and may be set arbitrarily by the server administrator.

Specifically, the support target vehicle determination unit 11f determines whether or not there is a candidate vehicle, which is a target vehicle 2 that will reach the unstable behavior position within a certain time, based on the unstable behavior position and the position information of the target vehicle 2 included in the target vehicle data. A candidate vehicle is a vehicle that is a candidate for support. The traveling direction of the target vehicle 2 may be recognized based on temporal changes in the position information, or may be recognized from the travel direction of the lane included in the map information by matching the position information with the map information.

When the support target vehicle determination unit 11f determines that there is a candidate vehicle that will reach an unstable behavior position within a certain time, it estimates whether the candidate vehicle will exhibit unstable behavior at the unstable behavior position.

The support target vehicle determination unit 11f estimates whether the candidate vehicle will exhibit unstable behavior at the unstable behavior position using a behavior discrimination model generated by the model generation unit 11e based on the target vehicle data of the candidate vehicle acquired by the target vehicle data recognition unit 11a. The support target vehicle determination unit 11f makes a judgment using a behavior discrimination model (e.g., an unstable behavior discrimination model and/or a stable behavior discrimination model) in a set area that includes the unstable behavior position. Note that the support target vehicle determination unit 11f may estimate whether the candidate vehicle will exhibit unstable behavior at the unstable behavior position without using a behavior discrimination model.

The support target vehicle determination unit 11f may select a model to be used based on the determination result of the situation determination unit 11c, the location attributes of the unstable behavior position (whether it is a motorway or a general road), and the individual vehicle attributes of the candidate vehicle (whether it is a freight vehicle or a passenger car). The support target vehicle determination unit 11f may further select a model taking into consideration the surrounding environment, such as the outside temperature and weather information.

The support target vehicle determination unit 11f determines whether the driving state of the candidate vehicle is closer to an unstable behavior discrimination model or a stable behavior discrimination model. The support target vehicle determination unit 11f determines whether any parameter of the candidate vehicle, such as the vehicle speed, acceleration, or yaw rate, is closer to the value of the feature parameter included in the unstable behavior discrimination model or the value of the feature parameter included in the stable behavior discrimination model. If the support target vehicle determination unit 11f determines that the driving state of the candidate vehicle is closer to the unstable behavior discrimination model, it estimates that the candidate vehicle will exhibit unstable behavior at the unstable behavior position. If the support target vehicle determination unit 11f estimates that the candidate vehicle will exhibit unstable behavior at the unstable behavior position, it determines that a support target vehicle exists.

The support target vehicle determination unit 11f may estimate whether the candidate vehicle will exhibit unstable behavior at the unstable behavior position by using a neural network (deep learning network) that has been trained to reflect the behavior discrimination model. One example of a neural network is CNN (Convolutional neural networks). There is no particular limitation on the learning method of the neural network, and well-known methods such as backpropagation can be used. Note that the support target vehicle determination unit 11f may employ well-known machine learning methods other than neural networks (e.g., support vector machines and boosting). It is sufficient that the behavior discrimination model is used in some way to determine the support target vehicle.

When the support target vehicle determination unit 11f determines that a support target vehicle exists, the vehicle support unit 11g performs vehicle support to prevent the support target vehicle from becoming unstable at the unstable behavior position. The vehicle support unit 11g performs various vehicle support for the target vehicle 2 via the communication unit 13. The vehicle support unit 11g may be configured to be able to perform vehicle support for the support target vehicle in real time from the occurrence of unstable behavior.

As vehicle assistance, the vehicle assistance unit 11g performs at least one of the following: notification of unstable behavior position information, which is information regarding the unstable behavior position; assistance in changing the driving route; notification of stable driving data regarding the driving state of the target vehicle 2 that did not exhibit unstable behavior at the unstable behavior position; and stable driving instructions that bring the driving state of the assisted target vehicle closer to the driving state of the stable driving data.

The vehicle assistance unit 11g notifies the assisted vehicle of unstable behavior position information, for example. The assisted vehicle can use the unstable behavior position information to prevent unstable behavior. The vehicle assistance unit 11g may also instruct the assisted vehicle to change its driving route to avoid the unstable behavior position. The assisted vehicle avoids the occurrence of unstable behavior at the unstable behavior position by changing the route (road) or the driving route within the road.

The vehicle assistance unit 11g may notify the assisted vehicle of stable driving data related to the driving state of the target vehicle 2 that did not exhibit unstable behavior at the unstable behavior position. The assisted vehicle can suppress unstable behavior by referring to the stable driving data.

The vehicle assistance unit 11g may issue a stable driving instruction to the assisted vehicle so that the driving state of the assisted vehicle approaches the driving state of the stable driving data. In this case, the vehicle assistance unit 11g can suppress the occurrence of unstable behavior at the unstable behavior position by bringing the driving state of the assisted vehicle closer to the driving state of the stable driving data.

When the assisted vehicle is in autonomous driving, the vehicle assistance unit 11g may instruct the cancellation of autonomous driving at the unstable behavior position together with the notification of the unstable behavior position information. By canceling autonomous driving and shifting to manual driving by the driver, the assisted vehicle can avoid the assisted vehicle becoming unstable at the unstable behavior position while still in autonomous driving.

The statistical value calculation unit 11h calculates statistical values for the set area based on the target vehicle data recognized by the target vehicle data recognition unit 11a (position information on the map of the target vehicle 2 and data on the driving state of the target vehicle 2). The statistical value calculation unit 11h repeats the calculation of statistical values at regular intervals. The regular interval is, for example, one month or two months. The regular interval is not particularly limited, and may be one week, two weeks, or six months.

The statistics include vehicle traffic statistics and unstable behavior statistics. First, we will explain the case of vehicle traffic statistics.

The vehicle traffic statistical value is a value related to the passage of the target vehicle 2 within the set area. As an example, the vehicle traffic statistical value can be the number of times the target vehicle 2 passes through the set area during a certain period of time. The statistical value calculation unit 11h may calculate the vehicle traffic statistical value by distinguishing the number of times each vehicle passes, or may calculate the vehicle traffic statistical value by distinguishing the number of times each driver passes.

Alternatively, the statistical value calculation unit 11h may calculate the vehicle traffic statistical value based on the conditions of divided areas obtained by dividing the set area into multiple areas. A divided area is an area obtained by dividing the set area into multiple areas. FIG. 6 is a diagram showing an example of a set area and divided areas. In FIG. 6, Tokyo is shown as set area A, and each ward is shown as divided areas B1 to B4. Note that FIG. 6 shows only some of the divided areas. The divided areas do not need to be cities, wards, towns, and villages, and can be areas of any shape within one set area. The divided areas do not overlap with each other. There may be blank areas within the set area that are not set as divided areas.

The statistical value calculation unit 11h calculates a vehicle traffic calculation value for each divided area based on the target vehicle data. The vehicle traffic calculation value for each divided area is a value related to the passage of the target vehicle 2 within the divided area. The vehicle traffic calculation value for each divided area can be, for example, the number of times the target vehicle 2 passes through the divided area in a certain period of time. The statistical value calculation unit 11h may calculate the vehicle traffic calculation value taking into account the number of times each vehicle passes, or may calculate the vehicle traffic calculation value taking into account the number of times each driver passes. Note that the certain period of time may be a different period for each calculation target, or may be common regardless of the calculation target.

The statistical value calculation unit 11h counts the number of divided areas in which the vehicle traffic calculation value is equal to or greater than the first divided area threshold. The first divided area threshold is a preset threshold value. Any value can be adopted as the first divided area threshold. The first divided area threshold may be set to a common value within one set area, or may be set to a different value for each divided area.

The statistical value calculation unit 11h calculates the vehicle traffic statistical value of the set area from the number (total number) of divided areas in the set area whose vehicle traffic calculation value is equal to or greater than the first divided area threshold. The statistical value calculation unit 11h may also calculate the number of divided areas in the set area whose vehicle traffic calculation value is equal to or greater than the first divided area threshold as the vehicle traffic statistical value.

The statistical value calculation unit 11h may use a predetermined arithmetic formula to calculate a vehicle traffic statistical value from the number of divided areas in which the vehicle traffic calculation value is equal to or greater than the first divided area threshold. The statistical value calculation unit 11h may use an evaluation value preset for each divided area to calculate the vehicle traffic statistical value as the total evaluation value of each divided area in the set area in which the vehicle traffic calculation value is equal to or greater than the first divided area threshold. The evaluation value is preset according to, for example, the position or area of each divided area. The evaluation value may be set to a larger value as the area of the divided area increases.

The statistical value calculation unit 11h may calculate the vehicle traffic statistical value based on the condition of the roads in the set area instead of the divided area. FIG. 7 is a diagram for explaining an example of a road. FIG. 7 shows roads C1 to C5 in the divided area B1. Roads C1 to C5 are also roads in the set area A in FIG. 6. It is noted that the divided area does not necessarily have to be set.

The statistical value calculation unit 11h calculates a vehicle traffic calculation value for each road based on the target vehicle data. Each road may be recognized as a different road at regular intervals, or may be recognized as being divided at any intersection or the like. The vehicle traffic calculation value for each road is a value related to the passage of the target vehicle 2 on each road. The vehicle traffic calculation value for each road can be, for example, the number of times the target vehicle 2 passes through each road in a certain period of time.

The statistical value calculation unit 11h counts the number of roads whose vehicle traffic calculation value is equal to or greater than the first road threshold. The first road threshold is a preset threshold value. Any value can be adopted as the first road threshold. The first road threshold may be set to a common value within one set area, or may be set to a different value for each road.

The statistical value calculation unit 11h calculates the vehicle traffic statistical value of the set area from the number (total number) of roads in the set area whose vehicle traffic calculation value is equal to or greater than the first road threshold. The statistical value calculation unit 11h may directly calculate the number of roads whose vehicle traffic calculation value is equal to or greater than the first road threshold as the vehicle traffic statistical value.

The statistical value calculation unit 11h may use a predetermined arithmetic formula to calculate a vehicle traffic statistical value from the number of roads whose vehicle traffic calculation value is equal to or greater than the first road threshold. The statistical value calculation unit 11h may use an evaluation value preset for each road to calculate a vehicle traffic statistical value by adding up the evaluation values of roads whose vehicle traffic calculation value is equal to or greater than the first road threshold within the set area. The evaluation value is preset according to, for example, the position or length of each road. The evaluation value may be set to a larger value as the length of the road increases.

Furthermore, the statistical value calculation unit 11h may calculate the statistical value in two stages using the divided areas and the roads. For example, the statistical value calculation unit 11h calculates the vehicle traffic calculation value of the divided area from the number of target vehicles 2 passing through each road in the divided area, and calculates the vehicle traffic statistical value of the set area from the vehicle traffic calculation value of the divided area.

The statistical value calculation unit 11h calculates the number of target vehicles 2 passing through each road in a certain period of time based on the target vehicle data (see FIG. 7). The statistical value calculation unit 11h calculates the vehicle traffic calculation value of the divided area from the number of target vehicles 2 passing through each road in the divided area in a certain period of time. For example, the statistical value calculation unit 11h can calculate the number of roads in the divided area on which the number of target vehicles 2 passing through is equal to or greater than the first calculation threshold as the vehicle traffic calculation value of the divided area. The first calculation threshold is a threshold value that is set in advance. Any value can be adopted as the first calculation threshold. The first calculation threshold may be set to a common value within one divided area, or may be set to a different value for each road.

The statistical value calculation unit 11h may use a predetermined arithmetic formula to calculate the vehicle traffic calculation value of the divided area from the number of roads on which the number of passing target vehicles 2 is equal to or greater than the first calculation threshold. The statistical value calculation unit 11h may use an evaluation value preset for each road to calculate the total evaluation value of roads within the divided area on which the vehicle traffic calculation value is equal to or greater than the first calculation threshold as the vehicle traffic calculation value of the divided area.

The statistical value calculation unit 11h calculates a vehicle traffic statistical value of a set area from the vehicle traffic calculation values of multiple divided areas within the set area. For example, the statistical value calculation unit 11h can calculate the number of divided areas within the set area whose vehicle traffic statistical value is equal to or greater than the second statistical threshold as the vehicle traffic statistical value of the set area. The second statistical threshold is a threshold value that is set in advance. Any value can be adopted as the second statistical threshold. A common value may be set for the second statistical threshold within one set area, or a different value may be set for each divided area.

The statistical value calculation unit 11h may use a predetermined arithmetic formula to calculate the vehicle traffic statistical value of the set area from the number of divided areas in which the vehicle traffic calculation value is equal to or greater than the second statistical threshold. The statistical value calculation unit 11h may use an evaluation value preset for each divided area to calculate the vehicle traffic statistical value as the total evaluation value of each divided area in the set area in which the vehicle traffic calculation value is equal to or greater than the second statistical threshold.

The statistical value calculation unit 11h may calculate the number of target vehicles 2 passing on the road by vehicle type, or may calculate the number of target vehicles 2 passing on the road by driver type.

Next, the case of unstable behavior statistics will be described. The unstable behavior statistics is a value related to the occurrence of unstable behavior within a set area. The unstable behavior statistics can be, for example, the number of times the unstable behavior position is recognized within a set area in a certain period of time or the number of times unstable behavior occurs. Even if unstable behavior occurs multiple times at the same position, the number of times the unstable behavior position is recognized may be one. The statistical value calculation unit 11h may directly calculate the unstable behavior statistics from the target vehicle data, or may calculate the unstable behavior statistics from the recognition result of the unstable behavior position recognition unit 11b based on the target vehicle data. Note that the statistical value calculation unit 11h may calculate the unstable behavior statistics by distinguishing the occurrence of unstable behavior for each vehicle, or may calculate the unstable behavior statistics by distinguishing the occurrence of unstable behavior for each driver.

Alternatively, the statistical value calculation unit 11h may calculate the unstable behavior statistical value taking into account the conditions of the divided areas obtained by dividing the set area into a number of areas (see FIG. 6). The statistical value calculation unit 11h calculates the unstable behavior calculation value for each divided area based on the target vehicle data. The unstable behavior calculation value for each divided area is a value related to the unstable behavior of the target vehicle 2 within the divided area in a certain period of time. The unstable behavior calculation value for each divided area may be, for example, the number of times the unstable behavior position is recognized within the divided area in a certain period of time or the number of times the unstable behavior occurs. The statistical value calculation unit 11h may calculate the unstable behavior calculation value taking into account the number of times each vehicle passes, or may calculate the unstable behavior calculation value taking into account the number of times each driver passes.

The statistical value calculation unit 11h counts the number of divided areas in which the unstable behavior calculation value is equal to or greater than the second divided area threshold. The second divided area threshold is a preset threshold value. Any value can be adopted as the second divided area threshold. The second divided area threshold may be set to a common value within one set area, or may be set to a different value for each divided area.

The statistical value calculation unit 11h calculates the unstable behavior statistical value of the set area from the number (total number) of divided areas in the set area in which the unstable behavior calculation value is equal to or greater than the second divided area threshold. The statistical value calculation unit 11h may calculate, for example, the number of divided areas in the set area in which the unstable behavior calculation value is equal to or greater than the second divided area threshold as the unstable behavior statistical value.

The statistical value calculation unit 11h may use a predetermined arithmetic formula to calculate the unstable behavior statistical value from the number of divided areas in which the unstable behavior calculation value is equal to or greater than the second divided area threshold. The statistical value calculation unit 11h may use an evaluation value preset for each divided area to calculate the unstable behavior statistical value as the total evaluation value of each divided area in the set area in which the unstable behavior calculation value is equal to or greater than the second divided area threshold. The evaluation value may be the same as that for the vehicle traffic statistical value.

The statistical value calculation unit 11h may calculate the unstable behavior statistical value based on the road conditions within the set area instead of the divided areas. The statistical value calculation unit 11h calculates the unstable behavior calculation value for each road based on the target vehicle data. Each road may be recognized as a different road at regular intervals, or may be recognized by dividing it at any intersection or the like. The unstable behavior calculation value for each road is a value related to the unstable behavior for each road in a certain period of time. The unstable behavior calculation value for each road may be, for example, the number of times the unstable behavior position is recognized for each road in a certain period of time, or the number of times unstable behavior occurs.

The statistical value calculation unit 11h counts the number of roads for which the unstable behavior calculation value is equal to or greater than the second road threshold. The second road threshold is a preset threshold value. Any value can be adopted as the second road threshold. The second road threshold may be set to a common value within one set area, or may be set to a different value for each road.

The statistical value calculation unit 11h calculates the unstable behavior statistical value of the set area from the number (total number) of roads in the set area whose unstable behavior calculation value is equal to or greater than the second road threshold. The statistical value calculation unit 11h may directly calculate the number of roads whose unstable behavior calculation value is equal to or greater than the second road threshold as the unstable behavior statistical value.

The statistical value calculation unit 11h may use a predetermined arithmetic formula to calculate the unstable behavior statistical value from the number of roads whose unstable behavior calculation values are equal to or greater than the second road threshold. The statistical value calculation unit 11h may use an evaluation value preset for each road to calculate the unstable behavior statistical value as the total evaluation value of roads whose unstable behavior calculation values are equal to or greater than the second road threshold within the set area. The evaluation value may be the same as that of the vehicle traffic statistical value.

Furthermore, the statistical value calculation unit 11h may calculate the statistical value in two stages using the divided area and the road. For example, the statistical value calculation unit 11h calculates an unstable behavior calculation value for the divided area from the unstable behavior count for each road in the divided area, and calculates the unstable behavior statistical value for the set area from the unstable behavior calculation value for the divided area (see FIG. 7). The unstable behavior count for each road may be, for example, the number of times the unstable behavior position is recognized for each road in a certain period of time or the number of times unstable behavior occurs.

The statistical value calculation unit 11h calculates the number of unstable behaviors for each road in a certain period of time based on the target vehicle data. The statistical value calculation unit 11h calculates an unstable behavior calculation value for the divided area from the number of unstable behaviors for each road in the divided area. The statistical value calculation unit 11h can calculate, for example, the number of roads in the divided area whose number of unstable behaviors is equal to or greater than the second calculation threshold as the unstable behavior calculation value for the divided area. The second calculation threshold is a threshold value that is set in advance. Any value can be adopted as the second calculation threshold. The second calculation threshold may be set to a common value within one divided area, or may be set to a different value for each road.

The statistical value calculation unit 11h may use a predetermined arithmetic formula to calculate the unstable behavior calculation value of the divided area from the number of roads whose unstable behavior count is equal to or greater than the second calculation threshold. The statistical value calculation unit 11h may use an evaluation value preset for each road to calculate the total evaluation values of roads whose unstable behavior calculation value is equal to or greater than the second calculation threshold within the divided area as the unstable behavior calculation value of the divided area.

The statistical value calculation unit 11h calculates an unstable behavior statistical value of a set area from the unstable behavior calculation values of multiple divided areas within the set area. For example, the statistical value calculation unit 11h can calculate the number of divided areas within the set area in which the unstable behavior calculation value is equal to or greater than the second statistical threshold as the unstable behavior statistical value of the set area. The second statistical threshold is a threshold value that is set in advance. Any value can be adopted as the second statistical threshold. A common value may be set for the second statistical threshold within one set area, or a different value may be set for each divided area.

The statistical value calculation unit 11h may use a predetermined arithmetic formula to calculate the unstable behavior statistical value of the set area from the number of divided areas in which the unstable behavior calculation value is equal to or greater than the second statistical threshold. The statistical value calculation unit 11h may use an evaluation value preset for each divided area to calculate the unstable behavior statistical value as the total of the evaluation values of each divided area in the set area in which the unstable behavior calculation value is equal to or greater than the second statistical threshold.

The model update necessity determination unit 11j determines whether or not it is necessary to update the behavior discrimination model for the set area based on the change in the vehicle traffic statistics since the behavior discrimination model was generated in the set area, or the change in the unstable behavior statistics since the behavior discrimination model was generated in the set area. The model update necessity determination unit 11j determines whether or not it is necessary to update the model based on a ratio that takes into account the change since the behavior discrimination model was generated in the set area.

The model update necessity determination unit 11j determines that it is necessary to update the behavior discrimination model of a set area, for example, when the amount of change in the latest vehicle traffic statistics value of the set area relative to the vehicle traffic statistics value when the behavior discrimination model was generated in the set area is equal to or greater than a first model update threshold. The first model update threshold is a threshold value set in advance. A different value may be set for the first model update threshold for each set area. The first model update threshold may be changed according to the time elapsed since the model update. The first model update threshold may be changed to a smaller value as the time elapsed since the model update becomes longer.

The vehicle traffic statistical value when the behavior discrimination model is generated in the set area is the vehicle traffic statistical value at the timing when the behavior discrimination model is generated by the model generation unit 11e, among the vehicle traffic statistical values repeatedly calculated at regular intervals. The amount of change is an absolute value. The model update necessity determination unit 11j determines that the behavior discrimination model needs to be updated if the amount of change is equal to or greater than the first model update threshold, regardless of whether the vehicle traffic statistical value when the behavior discrimination model is generated in the set area or the latest vehicle traffic statistical value is the larger value.

Alternatively, the model update necessity determination unit 11j determines that the behavior discrimination model of the set area needs to be updated when the amount of change in the latest unstable behavior statistical value of the set area relative to the unstable behavior statistical value when the behavior discrimination model was generated in the set area is equal to or greater than the second model update threshold. The second model update threshold can be set in the same way as the first model update threshold. The unstable behavior statistical value when the behavior discrimination model was generated in the set area is also similar to the vehicle traffic statistical value. The model update necessity determination unit 11j may determine whether or not the model needs to be updated based on the change in the statistical value from a neural network or the like.

When the model update necessity determination unit 11j determines that it is necessary to update the behavior discrimination model for the set area, the model update unit 11k updates the behavior discrimination model for the set area. The model update unit 11k updates the behavior discrimination model based on the target vehicle data corresponding to the set area after the behavior discrimination model is generated in the set area. The update of the behavior discrimination model can be performed in the same manner as the model generation by the model generation unit 11e.

The model update unit 11k may update the behavior discrimination model using only the target vehicle data after the behavior discrimination model has been generated, or may update the behavior discrimination model using both the target vehicle data before the behavior discrimination model has been generated and the target vehicle data after the behavior discrimination model has been generated.

[program]
The program causes the processor 11 of the information processing server 10 to function (operate) as a target vehicle data recognition unit 11a, an unstable behavior position recognition unit 11b, a situation determination unit 11c, a storage processing unit 11d, a model generation unit 11e, an assistance target vehicle determination unit 11f, a vehicle assistance unit 11g, a statistical value calculation unit 11h, a model update necessity determination unit 11j, and a model update unit 11k. The program is provided by a non-transitory recording medium such as a ROM or a semiconductor memory. The program may also be provided via communication such as a network.

[Processing method of information processing server]
Next, a processing method of the information processing server 10 according to the present embodiment will be described with reference to the drawings. Fig. 8 is a flowchart showing an example of an unstable behavior position storage process. The unstable behavior position storage process is executed, for example, when the information processing server 10 manages the target vehicle 2.

As shown in FIG. 8, in S10, the information processing server 10 recognizes the target vehicle data using the target vehicle data recognition unit 11a (target vehicle data recognition step). The target vehicle data recognition unit 11a recognizes the target vehicle data transmitted from the target vehicle 2. The target vehicle data includes the position information of the target vehicle 2 on the map and the driving state of the target vehicle 2. Thereafter, the information processing server 10 proceeds to S12.

In S12, the information processing server 10 determines whether the target vehicle 2 has become unstable using the unstable behavior position recognition unit 11b (unstable behavior determination step). The unstable behavior position recognition unit 11b determines whether the target vehicle 2 has become unstable based on the target vehicle data. The unstable behavior position recognition unit 11b determines the presence or absence of unstable behavior by determining whether the target vehicle 2A has become slipping as an unstable behavior from the wheel speed of each wheel of the target vehicle 2, for example. If the information processing server 10 has not determined that the target vehicle 2 has become unstable (S12: NO), it ends the storage process of the current unstable behavior position information. If the information processing server 10 has determined that the target vehicle 2 has become unstable (S12: YES), it proceeds to S14.

In S14, the information processing server 10 recognizes the unstable behavior position by the unstable behavior position recognition unit 11b (unstable behavior position recognition step). The unstable behavior position recognition unit 11b recognizes the unstable behavior position based on the position information on the map of the target vehicle 2 when it is determined that the target vehicle 2 has become unstable. After that, the information processing server 10 proceeds to S16.

In S16, the information processing server 10 performs storage processing of the unstable behavior position by the storage processing unit 11d (storage processing step). The storage processing unit 11d stores the unstable behavior position recognized by the unstable behavior position recognition unit 11b in the storage database 15. When a determination is made by the situation determination unit 11c, the storage processing unit 11d may associate the unstable behavior position with the determination result of the situation determination unit 11c and store them in the storage database 15. The storage processing unit 11d may associate the unstable behavior position information with the target vehicle data of the target vehicle 2 that traveled through the unstable behavior position and store them in the storage database 15.

Figure 9 (a) is a flowchart showing an example of a behavior discrimination model generation process. The behavior discrimination model generation process is executed, for example, when a setting area is set.

As shown in FIG. 9(a), the information processing server 10 recognizes target vehicle data in the set area by the target vehicle data recognition unit 11a in S20 (target vehicle data recognition step). The target vehicle data recognition unit 11a recognizes the target vehicle data by wireless communication with each target vehicle 2.

In S22, the information processing server 10 generates a behavior discrimination model for the set area by the model generation unit 11e (model generation step). For each of the multiple set areas set on the map, the model generation unit 11e generates a behavior discrimination model from the target vehicle data of the target vehicle 2 that has become unstable at the unstable behavior position.

Figure 9 (b) is a flowchart showing an example of the vehicle assistance execution process. The vehicle assistance execution process is executed periodically, for example, for each target vehicle 2 that is traveling.

As shown in FIG. 9(b), in step S30, the information processing server 10 recognizes target vehicle data in the set area using the target vehicle data recognition unit 11a (target vehicle data recognition step).

In S32, the information processing server 10 determines whether or not a support target vehicle exists by using the support target vehicle determination unit 11f, using a behavior discrimination model corresponding to the set area (support target vehicle determination step). The support target vehicle determination unit 11f determines whether or not a support target vehicle exists by using a behavior discrimination model corresponding to the set area, based on target vehicle data such as the target vehicle 2 heading toward the unstable behavior position within the set area. If the information processing server 10 determines that a support target vehicle exists (S32: YES), it proceeds to S34. If the information processing server 10 does not determine that a support target vehicle exists (S32: NO), it ends this processing.

In S34, the information processing server 10 executes vehicle assistance for the assisted vehicle by the vehicle assistance unit 11g (vehicle assistance step). As vehicle assistance, the vehicle assistance unit 11g performs at least one of the following: notification of unstable behavior position information, which is information on the unstable behavior position; driving route change assistance; notification of stable driving data on the driving state of the target vehicle 2 that did not become unstable at the unstable behavior position; and stable driving instructions for bringing the driving state of the assisted vehicle closer to the driving state of the stable driving data. Thereafter, the information processing server 10 ends this processing.

Figure 10 is a flowchart showing an example of a model update process. The model update process is performed periodically. The model update process may be performed each time a statistical value of a set area is calculated.

As shown in FIG. 10, in S40, the information processing server 10 calculates the statistical values (vehicle traffic statistical values or unstable behavior statistical values) of the set area by the statistical value calculation unit 11h (statistical value calculation step). The statistical value calculation unit 11h calculates the statistical values based on the target vehicle data. The statistical value calculation unit 11h calculates the vehicle traffic statistical values or the unstable behavior statistical values. The statistical value calculation unit 11h may calculate the unstable behavior statistical values from the recognition results of the unstable behavior positions recognized based on the target vehicle data.

In S42, the information processing server 10 determines whether or not the behavior discrimination model of the set area needs to be updated based on the change in the statistical value since the model refinement, using the model update necessity determination unit 11j (model update necessity determination step). For example, when the amount of change in the latest vehicle traffic statistics value of the set area relative to the vehicle traffic statistics value when the behavior discrimination model was generated in the set area is equal to or greater than a first model update threshold, the model update necessity determination unit 11j determines that the behavior discrimination model of the set area needs to be updated. Alternatively, when the amount of change in the latest unstable behavior statistics value of the set area relative to the unstable behavior statistics value when the behavior discrimination model was generated in the set area is equal to or greater than a second model update threshold, the model update necessity determination unit 11j determines that the behavior discrimination model of the set area needs to be updated.

If the information processing server 10 determines that the behavior discrimination model for the set area needs to be updated (S42: YES), it proceeds to S44. If the information processing server 10 determines that the behavior discrimination model for the set area needs to be updated (S42: NO), it ends the current process.

In step S44, the information processing server 10 updates the behavior discrimination model for the set area using the model update unit 11k (model update step). The model update unit 11k updates the behavior discrimination model based on the target vehicle data corresponding to the set area after the behavior discrimination model has been generated for the set area. Thereafter, the information processing server 10 ends this processing.

Fig. 11(a) is a flowchart showing an example of a statistical value calculation process using a divided area. As shown in Fig. 11(a), the information processing server 10 calculates, as S50, a vehicle traffic calculation value or an unstable behavior calculation value for each divided area in a set area by the statistical value calculation unit 11h (divided area calculation step). The vehicle traffic calculation value for each divided area may be the number of times the target vehicle 2 passes through the divided area in a certain period of time. The unstable behavior calculation value for each divided area may be the number of times the unstable behavior position is recognized in the divided area in a certain period of time or the number of times unstable behavior occurs.

As S52, the information processing server 10 calculates a statistical value (vehicle traffic statistical value or unstable behavior statistical value) for the set area by the statistical value calculation unit 11h (statistical value calculation step). The statistical value calculation unit 11h calculates a vehicle traffic statistical value for the set area including the divided area from the vehicle traffic calculation value for each divided area. Alternatively, the statistical value calculation unit 11h calculates an unstable behavior statistical value for the set area including the divided area from the unstable behavior calculation value for each divided area. Thereafter, the information processing server 10 ends this process.

Fig. 11(b) is a flowchart showing an example of a statistical value calculation process using roads. As shown in Fig. 11(b), in S60, the information processing server 10 calculates a vehicle traffic calculation value or an unstable behavior calculation value for each road in the set area by the statistical value calculation unit 11h (road calculation step). The vehicle traffic calculation value for each road may be the number of times the target vehicle 2 passes through the road in a certain period of time. The unstable behavior statistical value for each road may be the number of times the unstable behavior position on the road is recognized or the number of times unstable behavior occurs in a certain period of time.

As S62, the information processing server 10 calculates the statistical value (vehicle traffic statistical value or unstable behavior statistical value) for the set area by the statistical value calculation unit 11h (statistical value calculation step). The statistical value calculation unit 11h calculates the vehicle traffic statistical value for the set area from the vehicle traffic calculation value for each road. Alternatively, the statistical value calculation unit 11h calculates the unstable behavior statistical value for the set area from the unstable behavior calculation value for each road. Thereafter, the information processing server 10 ends this process.

The information processing server 10 (and the processing method and program of the information processing server 10) according to the present embodiment described above calculates vehicle traffic statistics in a set area based on target vehicle data, and determines whether or not the behavior discrimination model for the set area needs to be updated based on changes in the vehicle traffic statistics since the behavior discrimination model was generated in the set area. Alternatively, the information processing server 10 calculates unstable behavior statistics in a set area based on target vehicle data, and determines whether or not the behavior discrimination model for the set area needs to be updated based on changes in the unstable behavior statistics since the behavior discrimination model was generated in the set area.

Therefore, according to the information processing server 10, the behavior discrimination model can be updated at an appropriate timing according to the situation of the set area, compared to a case where the model is updated periodically without considering the situation of the set area. As a result, it is possible to prevent the learned behavior discrimination model from deviating from the market environment. In addition, it is possible to reduce the load and operating costs of the information processing server 10, compared to a case where the model is updated periodically with high frequency.

Furthermore, this information processing server 10 does not make a judgment based only on the latest vehicle traffic statistics or unstable behavior statistics, but instead focuses on changes since the behavior discrimination model was generated in the set area and judges whether or not the model needs to be updated on a ratio basis, so that the behavior discrimination model can be updated taking into account changes in the situation in the set area since the behavior discrimination model was generated (such as market changes).

In addition, according to the information processing server 10, if the amount of change in the latest vehicle traffic statistics for the set area relative to the vehicle traffic statistics when the behavior discrimination model was generated in the set area is equal to or greater than a first model update threshold, or if the amount of change in the latest unstable behavior statistics for the set area relative to the unstable behavior statistics when the behavior discrimination model was generated in the set area is equal to or greater than a second model update threshold, it is determined that the behavior discrimination model for the set area needs to be updated. Therefore, by setting the threshold, the behavior discrimination model can be updated at an appropriate timing according to the situation in the set area.

In addition, the information processing server 10 can calculate statistical values for a set area based on the conditions of each divided area into which the set area is divided, or the conditions of each road within the set area. Even if extreme values occur in part of the set area, they can be gradually reflected in the statistical values.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above-mentioned embodiments. The present invention can be implemented in various forms, including the above-mentioned embodiments, with various modifications and improvements based on the knowledge of those skilled in the art.

The information processing server 10 does not necessarily have to have the situation determination unit 11c. The information processing server 10 may generate a behavior discrimination model based on data determined by another server.

2, 2A-2Z... target vehicle, 10... information processing server, 11... processor, 11b... unstable behavior position recognition unit, 11e... model generation unit, 11f... assistance target vehicle determination unit, 11g... vehicle assistance unit, 11h... statistical value calculation unit, 11j... model update necessity determination unit, 11k... model update unit, 12... storage unit, 13... communication unit, 14... user interface, 15... storage database, A... setting area, B1-B4... division area, C1-C5... road, D... unstable behavior position.

Claims (6)

An information processing server that obtains target vehicle data including a driving state of the target vehicle and position information of the target vehicle on a map from a plurality of target vehicles,
an unstable behavior position recognition unit that recognizes an unstable behavior position, which is a position on a map where the target vehicle has become unstable, based on the target vehicle data;
a model generation unit that generates a behavior discrimination model from the target vehicle data of the target vehicle at the unstable behavior position for each of a plurality of set areas that are set on a map;
an assistance target vehicle determination unit that determines whether or not there is an assistance target vehicle that is the target vehicle that is estimated to have an unstable behavior at the unstable behavior location, based on the unstable behavior location, the target vehicle data, and the behavior discrimination model corresponding to the set area including the unstable behavior location; and
a vehicle assistance unit that provides vehicle assistance to the support target vehicle when the support target vehicle determination unit determines that the support target vehicle exists;
a statistical value calculation unit that calculates a vehicle traffic statistical value in the set area or an unstable behavior statistical value in the set area based on the target vehicle data;
a model update necessity determination unit that determines whether or not the behavior discrimination model of the set area needs to be updated based on a change in the vehicle traffic statistics from when the behavior discrimination model was generated in the set area or a change in the unstable behavior statistics from when the behavior discrimination model was generated in the set area; and
a model updating unit that updates the behavior discrimination model of the set area when the model update necessity determining unit determines that the behavior discrimination model of the set area needs to be updated;
Equipped with
The unstable behavior is a behavior of the target vehicle that makes the traveling of the target vehicle unstable,
the unstable behavior of the target vehicle includes a slip of the target vehicle;
The assistance target vehicle determination unit determines whether or not a candidate vehicle, which is the target vehicle that will reach the unstable behavior location within a certain period of time, exists based on the unstable behavior location and the position information of the target vehicle included in the target vehicle data, and if it is determined that the candidate vehicle exists, determines whether or not the assistance target vehicle, which is the candidate vehicle that will cause the unstable behavior at the unstable behavior location, exists based on the target vehicle data of the candidate vehicle, by using the behavior discrimination model corresponding to the set area that includes the unstable behavior location . The information processing server according to claim 1, wherein the model update necessity determination unit determines that the behavior discrimination model of the set area needs to be updated if the amount of change in the latest vehicle traffic statistics of the set area relative to the vehicle traffic statistics when the behavior discrimination model was generated in the set area is equal to or greater than a first model update threshold, or if the amount of change in the latest unstable behavior statistics of the set area relative to the unstable behavior statistics when the behavior discrimination model was generated in the set area is equal to or greater than a second model update threshold. the statistical value calculation unit calculates a vehicle traffic calculation value for each divided area obtained by dividing the set area into a plurality of divided areas based on the target vehicle data, and calculates the vehicle traffic statistical value for the set area from the number of divided areas in which a difference between the vehicle traffic calculation value at the time when the behavior discrimination model was generated in the set area and the latest vehicle traffic calculation value is equal to or greater than a first divided area threshold value.
Alternatively, the information processing server according to claim 1 or 2 calculates an unstable behavior calculation value for each divided area obtained by dividing the set area into a plurality of areas based on the target vehicle data, and calculates the unstable behavior statistical value for the set area from the number of divided areas in which the difference between the unstable behavior calculation value when the behavior discrimination model was generated in the set area and the latest unstable behavior calculation value is equal to or greater than a second divided area threshold value. the statistical value calculation unit calculates a vehicle traffic calculation value for each road in the set area based on the target vehicle data, and calculates a vehicle traffic statistical value for the set area from the number of roads for which a difference between the vehicle traffic calculation value at the time when the behavior discrimination model was generated in the set area and the latest vehicle traffic calculation value is equal to or greater than a first road threshold value;
Alternatively, the information processing server according to claim 1 or 2 calculates an unstable behavior calculation value for each road within the set area based on the target vehicle data, and calculates an unstable behavior statistical value for the set area from the number of roads for which the difference between the unstable behavior calculation value when the behavior discrimination model was generated in the set area and the latest unstable behavior calculation value is equal to or greater than a second road threshold. A processing method of an information processing server for obtaining target vehicle data including a driving state of the target vehicles and position information of the target vehicles on a map from a plurality of target vehicles, the method comprising:
an unstable behavior position recognition step of recognizing an unstable behavior position, which is a position on a map where the target vehicle has become unstable, based on the target vehicle data;
a model generation step of generating a behavior discrimination model from the target vehicle data of the target vehicle at the unstable behavior position for each of a plurality of set areas set on a map;
an assistance target vehicle determination step of determining whether or not there is an assistance target vehicle, which is the target vehicle that is estimated to have an unstable behavior at the unstable behavior location, based on the unstable behavior location, the target vehicle data, and the behavior discrimination model corresponding to the set area including the unstable behavior location;
a vehicle assistance step of providing vehicle assistance to the support target vehicle when it is determined in the support target vehicle determination step that the support target vehicle exists;
a statistical value calculation step of calculating a vehicle traffic statistical value in the set area or an unstable behavior statistical value in the set area based on the target vehicle data;
a model update necessity determination step of determining whether or not the behavior discrimination model of the set area needs to be updated based on a change in the vehicle traffic statistics from when the behavior discrimination model was generated in the set area or a change in the unstable behavior statistics from when the behavior discrimination model was generated in the set area;
a model updating step of updating the behavior discrimination model of the set area when it is determined in the model update necessity determining step that the update of the behavior discrimination model of the set area is necessary;
Including,
The unstable behavior is a behavior of the target vehicle that makes the traveling of the target vehicle unstable,
the unstable behavior includes slippage of the target vehicle;
A processing method of an information processing server in which, in the assistance target vehicle determination step, it is determined whether or not a candidate vehicle, which is the target vehicle that will reach the unstable behavior position within a certain period of time, exists based on the unstable behavior position and the position information of the target vehicle included in the target vehicle data, and if it is determined that the candidate vehicle exists, it is determined whether or not the assistance target vehicle, which is the candidate vehicle that will cause the unstable behavior at the unstable behavior position, exists based on the target vehicle data of the candidate vehicle, by using the behavior discrimination model corresponding to the set area that includes the unstable behavior position . A program for operating an information processing server that obtains target vehicle data including a driving state of the target vehicle and position information of the target vehicle on a map from a plurality of target vehicles,
an unstable behavior position recognition unit that recognizes an unstable behavior position, which is a position on a map where the target vehicle has become unstable, based on the target vehicle data;
a model generation unit that generates a behavior discrimination model from the target vehicle data of the target vehicle at the unstable behavior position for each of a plurality of set areas that are set on a map;
an assistance target vehicle determination unit that determines whether or not there is an assistance target vehicle, which is the target vehicle that is estimated to have an unstable behavior at the unstable behavior location, based on the unstable behavior location, the target vehicle data, and the behavior discrimination model corresponding to the set area including the unstable behavior location;
a vehicle assistance unit that provides vehicle assistance to the support target vehicle when the support target vehicle determination unit determines that the support target vehicle exists;
a statistical value calculation unit that calculates a vehicle traffic statistical value in the set area or an unstable behavior statistical value in the set area based on the target vehicle data;
a model update necessity determination unit that determines whether or not the behavior discrimination model of the set area needs to be updated based on a change in the vehicle traffic statistics from when the behavior discrimination model was generated in the set area or a change in the unstable behavior statistics from when the behavior discrimination model was generated in the set area;
and operating the information processing server as a model update unit that updates the behavior discrimination model of the set area when the model update necessity determination unit determines that the behavior discrimination model of the set area needs to be updated ;
The unstable behavior is a behavior of the target vehicle that makes the traveling of the target vehicle unstable,
the unstable behavior includes slippage of the target vehicle;
The support target vehicle determination unit determines whether or not a candidate vehicle, which is the target vehicle that will reach the unstable behavior position within a certain period of time, exists based on the unstable behavior position and the position information of the target vehicle included in the target vehicle data, and if it is determined that the candidate vehicle exists, determines whether or not the support target vehicle, which is the candidate vehicle that will cause the unstable behavior at the unstable behavior position, exists based on the target vehicle data of the candidate vehicle, by using the behavior discrimination model corresponding to the set area that includes the unstable behavior position .

Images