JP2023013458A - Information processing server, processing method of information processing server, and program - Google Patents

Abstract

To provide an information processing server capable of updating a behavior discrimination model at an appropriate timing according to a situation of a set area.SOLUTION: An information processing server for obtaining target vehicle data from a plurality of target vehicles, comprises: an unstable behavior position recognition unit that recognizes an unstable behavior position of the target vehicle 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 set on a map; a statistical value calculation unit that calculates either a vehicle traffic statistical value in the set area or an unstable behavior statistical value in the set area; a model updating necessity determination unit that determines whether updating of the behavior discrimination model is necessary based on a change in the vehicle traffic statistical value or the unstable behavior statistical value since the behavior discrimination model was generated in the set area; and a model updating unit that updates the behavior discrimination model in the set area when it is determined that the behavior discrimination model is required to be updated.SELECTED DRAWING: Figure 10

Description

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

Japanese Unexamined Patent Application Publication No. 2020-052607, for example, is known regarding the processing of information related to travel of a vehicle. This publication discloses a system that stores the unstable behavior information at the behavior occurrence position where the target vehicle has become unstable behavior in association with the judgment result as to whether or not the unstable behavior is caused by the driver.

JP 2020-052607 A

By the way, it is considered to use a vehicle determination model generated from various types of information of the target vehicle for determination regarding unstable behavior as described above. However, when time elapses after the vehicle determination model is generated, the vehicle determination model may deviate from the actual market environment and the determination accuracy may decrease. On the other hand, uniformly increasing the update frequency of the vehicle determination model results in an increase in processing load and operation cost.

One aspect of the present invention is an information processing server that obtains target vehicle data including the driving state of the target vehicle and the position information of the target vehicle on a map from a plurality of target vehicles, the target vehicle An unstable behavior position recognition unit that recognizes the unstable behavior position, which is the position on the map where the behavior becomes unstable, and a target vehicle at the unstable behavior position for each set area set on the map Unstable at the unstable behavior position based on the model generation unit that generates the behavior discrimination model from the data, and the behavior discrimination model corresponding to the setting area containing the unstable behavior position, the target vehicle data, and the unstable behavior position. a support target vehicle determination unit that determines whether or not a support target vehicle, which is a target vehicle that is estimated to behave, exists or not; A vehicle support unit that provides vehicle support for the vehicle, a statistical value calculation unit that calculates vehicle traffic statistics in the set area or unstable behavior statistics in the set area based on the target vehicle data, and a behavior discrimination model in the set area Based on changes in vehicle traffic statistics from the time it was generated, or changes in unstable behavior statistics from the time the behavior discrimination model was generated in the set area, it is necessary to update the behavior discrimination model in that set area. When it is determined that the model update necessity determination unit that determines whether or not the model update necessity determination unit needs to update the behavior determination model of the setting area, the behavior determination model of the setting area and a model updating unit that performs updating.

In the information processing server according to one aspect of the present invention, the vehicle traffic statistics in the set area are calculated based on the target vehicle data, and based on changes in the vehicle traffic statistics from when the behavior discrimination model is generated in the set area. Then, it is determined whether or not the behavior discrimination model of the setting area needs to be updated. Alternatively, the information processing server calculates an unstable behavior statistic value in the set area based on the target vehicle data, and based on a change in the unstable behavior statistic value from when the behavior discrimination model is generated in the set area, It is determined whether or not the behavior discrimination model of the set area needs to be updated. Therefore, according to this information processing server, the behavior discrimination model can be updated at an appropriate timing according to the situation of the set area, compared to the case where the model is regularly updated without considering the situation of the set area. can be done. In addition, in this information processing server, instead of making a judgment based only on the latest vehicle traffic statistics or unstable behavior statistics, it focuses on changes from when the behavior discrimination model was generated in the set area and on a ratio basis. Since it is determined whether or not the model needs to be updated, the behavior discrimination model can be updated in consideration of the change in the situation of the set area since the behavior discrimination model was generated.

In the information processing server according to one aspect of the present invention, the model update necessity determination unit determines the amount of change in the latest vehicle traffic statistic value in the set area with respect to the vehicle traffic statistic value when the behavior discrimination model is generated in the set area. is greater than or equal to the first model update threshold, or the amount of change in the latest unstable behavior statistic value of the setting area with respect to the unstable behavior statistic value when the behavior discrimination model is generated in the setting area is the second If it is equal to or greater than the model update threshold, it may be determined that the behavior discrimination model of the set area needs to be updated.
According to this information processing server, the amount of change in the latest vehicle traffic statistic value in the set area with respect to the vehicle traffic statistic value when the behavior discrimination model was generated in the set area is greater than or equal to the first model update threshold, or When the amount of change in the latest unstable behavior statistic value of the set area with respect to the unstable behavior statistic value when the behavior discrimination model is generated in the area is greater than or equal to the second model update threshold, behavior discrimination of the set area Since it is determined that the model needs to be updated, the behavior discrimination model can be updated at an appropriate timing according to the situation of the set area by setting the threshold.

In the information processing server according to one aspect of the present invention, the statistic value calculation unit calculates the vehicle traffic calculation value for each divided area obtained by dividing the set area based on the target vehicle data, and the behavior discrimination model is calculated in the set area. Calculate the vehicle traffic statistic value of the set area from the number of divided areas in which the difference between the calculated vehicle traffic value at the time of generation and the latest calculated vehicle traffic value is equal to or greater than the third model update threshold, or the target vehicle Based on the data, calculate the unstable behavior calculation value for each divided area by dividing the setting area into multiple areas, and compare the unstable behavior calculation value and the latest unstable behavior calculation value when the behavior discrimination model is generated in the setting area. The unstable behavior statistic value of the set area may be calculated from the number of divided areas in which the difference is equal to or greater than the fourth model update threshold.
According to this information processing server, the setting area is divided into a plurality of areas, and the statistical values of the setting area are calculated based on the situation of each divided area. It can be reflected in statistics.

In the information processing server according to one aspect of the present invention, the statistical value calculation unit calculates the vehicle traffic calculation value for each road in the set area based on the target vehicle data, and when the behavior discrimination model is generated in the set area Calculate the vehicle traffic statistics of the set area from the number of roads where the difference between the calculated vehicle traffic value and the latest calculated vehicle traffic value is equal to or greater than the fifth model update threshold, or set based on the target vehicle data Calculate the unstable behavior calculated value for each road in the area, and the difference between the unstable behavior calculated value when the behavior discrimination model is generated in the set area and the latest unstable behavior calculated value is the fourth model update threshold The unstable behavior statistic value of the set area may be calculated from the above number of roads.
According to this information processing server, the statistical values of the set area are calculated based on the situation of each road in the set area. can be reflected in

Another aspect of the present invention is a processing method of an information processing server for obtaining target vehicle data including the driving state of the target vehicle and the position information of the target vehicle on a map from a plurality of target vehicles, the method comprising: an unstable behavior position recognition step for recognizing the unstable behavior position on the map where the target vehicle has become unstable; Based on the model generation step of generating a behavior discrimination model from the target vehicle data of the unstable behavior position, the target vehicle data, and the behavior discrimination model corresponding to the set area containing the unstable behavior position, the unstable behavior position A support target vehicle determination step for determining whether or not there is a support target vehicle that is a target vehicle that is estimated to exhibit unstable behavior in , and if it is determined that the support target vehicle exists in the support target vehicle determination step, A vehicle assistance step of providing vehicle assistance to a vehicle to be assisted; a statistical value calculation step of calculating vehicle traffic statistics in a set area or unstable behavior statistics in the set area based on target vehicle data; and behavior in the set area. Based on the change in vehicle traffic statistics from when the discrimination model was generated, or the change in unstable behavior statistics from when the behavior discrimination model was generated in the set area, the behavior discrimination model of the set area A model update necessity determination step for determining whether an update is necessary, and when it is determined in the model update necessity determination step that the behavior determination model of the setting area needs to be updated, the behavior of the setting area and a model update step for updating the discriminant model.

In the processing method of the information processing server according to another aspect of the present invention, the vehicle traffic statistical value in the set area is calculated based on the target vehicle data, and the vehicle traffic statistical value from when the behavior discrimination model is generated in the set area is calculated. Based on the change in , it is determined whether or not the behavior discrimination model of the set area needs to be updated. Alternatively, in the processing method of the information processing server, the unstable behavior statistic value in the set area is calculated based on the target vehicle data, and changes in the unstable behavior statistic value from when the behavior discrimination model is generated in the set area Based on this, it is determined whether or not the behavior discrimination model of the setting area needs to be updated. Therefore, according to this processing method of the information processing server, the behavior discrimination model is updated at an appropriate timing according to the situation of the setting area compared to the case where the model is updated periodically without considering the situation of the setting area. It can be performed. In addition, in the processing method of this information processing server, instead of making a determination based only on the latest vehicle traffic statistics or unstable behavior statistics, attention is paid to changes since the behavior discrimination model was generated in the set area. Since the need for model update is determined on a ratio basis, the behavior discrimination model can be updated in consideration of the change in the situation of the set area since the behavior discrimination model was generated.

Still another aspect of the present invention is a program for operating an information processing server for obtaining target vehicle data including the driving state of the target vehicle and positional information of the target vehicle on a map from a plurality of target vehicles, the program comprising: Based on, the unstable behavior position recognition unit that recognizes the unstable behavior position, which is the position on the map where the target vehicle has become unstable, and the object at the unstable behavior position for each set area set on the map A model generation unit that generates a behavior discrimination model from target vehicle data of a vehicle, an unstable behavior position based on a behavior discrimination model corresponding to a set area that includes the unstable behavior position, the target vehicle data, and the unstable behavior position A support target vehicle determination unit that determines whether or not a support target vehicle, which is a target vehicle that is estimated to exhibit unstable behavior in , exists, and if the support target vehicle determination unit determines that a support target vehicle exists, A vehicle support unit that provides vehicle support for the target vehicle, a statistical value calculation unit that calculates vehicle traffic statistics in the set area or unstable behavior statistics in the set area based on the target vehicle data, and a behavior discrimination model in the set area Based on the change in vehicle traffic statistics from when the is generated, or the change in unstable behavior statistics from when the behavior discrimination model was generated in the set area, the behavior discrimination model in the set area is updated. A model update necessity determination unit that determines whether it is necessary, and when the model update necessity determination unit determines that the behavior determination model of the set area needs to be updated, the behavior determination of the set area The information processing server is operated as a model update unit that updates the model.

In the program according to still another aspect of the present invention, the vehicle traffic statistics in the set area are calculated based on the target vehicle data, and based on changes in the vehicle traffic statistics from when the behavior discrimination model is generated in the set area. Then, it is determined whether or not the behavior discrimination model of the setting area needs to be updated. Alternatively, in this program, the unstable behavior statistic value in the set area is calculated based on the target vehicle data, and based on the change in the unstable behavior statistic value from when the behavior discrimination model is generated in the set area, the setting It is determined whether or not the area behavior discrimination model needs to be updated. Therefore, according to this program, the behavior discrimination model can be updated at an appropriate timing according to the situation of the set area, compared to the case where the model is regularly updated without considering the situation of the set area. . In addition, in this program, instead of making judgments based only on the latest vehicle traffic statistics or unstable behavior statistics, we focus on changes since the behavior discrimination model was generated in the set area and update the model on a ratio basis. Therefore, the behavior discrimination model can be updated in consideration of the change in the situation of the set area since the behavior discrimination model was generated.

According to each aspect of the present invention, it is possible to update the behavior determination model at an appropriate timing according to the situation of the set area.

It is a figure which shows the information processing server which concerns on one Embodiment, and a target vehicle. It is a figure for demonstrating an example of information processing. It is a block diagram which shows an example of a structure of a target vehicle. It is a block diagram which shows an example of a structure of an information processing server. (a) It is a figure for demonstrating an example of the continuous occurrence condition. (b) It is a figure for demonstrating an example of a discontinuous situation. It is a figure for demonstrating an example of a setting area and a division area. It is a figure for demonstrating an example of a road. 7 is a flowchart showing an example of unstable behavior position storage processing; (a) It is a flowchart which shows an example of a behavior discrimination|determination model generation process. (b) It is a flowchart which shows an example of a vehicle assistance execution process. 6 is a flowchart illustrating an example of model update processing; (a) It is a flowchart which shows an example of the statistical value calculation process using a division area. (b) It is a flowchart which shows the other example of the statistical value calculation process using a road.

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a diagram showing an information processing server 10 and a target vehicle 2 according to the first embodiment. As shown in FIG. 1, the information processing server 10 is connected to the target vehicle 2 (2A to 2Z) via the network N so as to be communicable. Network N is a wireless communication network. The target vehicle 2 means a vehicle for which the information processing server 10 collects information. The target vehicle 2 includes a support target vehicle for which various types of support are provided from the information processing server 10 . When the subject vehicles 2 are individually described, the subject vehicles 2A to 2Z are used.

FIG. 2 is a diagram for explaining an example of information processing. As shown in FIG. 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, for example, the target vehicle 2B traveling behind the target vehicle 2A of information on the unstable behavior position. As a result, it becomes possible to suppress the occurrence of slipping of the target vehicle 2B at the unstable behavior position D in the target vehicle 2B. The unstable behavior position will be described later in detail.

[Configuration of target vehicle]
First, the configuration of the target vehicle 2 will be described. An ID [identification] (vehicle identification number) for identifying the vehicle is assigned to the target vehicle 2 . The number of target vehicles 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 types. The target vehicle 2 may be an automatically driven vehicle having an automatic driving function, or may be a vehicle without an automatic driving function.

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

As shown in FIG. 3 , the target vehicle 2 has an automatic driving ECU 30 . The automatic driving ECU 30 is an electronic control unit having a CPU, ROM, RAM, and the like. The automatic driving ECU 30 implements various functions by, for example, loading programs stored in the ROM into the RAM and executing the programs loaded into the RAM by the CPU. The automatic driving ECU 30 may be composed of a plurality of electronic units.

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

The GPS receiver 21 measures the position of the target vehicle 2 (for example, 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 positional information of the target vehicle 2 to the automatic 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 imaging equipment that images the external environment of the target vehicle 2 . The camera is provided on the back side of the windshield of the target vehicle 2 and images the front of the vehicle. A camera transmits the imaging information regarding the external environment of the target vehicle 2 to automatic driving ECU30. The camera may be a monocular camera or a stereo camera.

A radar sensor is a detection device that detects objects around the target vehicle 2 using radio waves (for example, millimeter waves) or light. Radar sensors include, for example, millimeter wave radar or lidar [LIDAR: Light Detection and Ranging]. The radar sensor detects an object by transmitting radio waves or light around the target vehicle 2 and receiving radio waves or light reflected by the object. The radar sensor transmits detected object information to the automatic 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 temperature sensor that detects the outside temperature of the target vehicle 2 . External sensor 22 may include a light sensor that detects external brightness.

The internal sensor 23 is a detection device that detects the state of the target vehicle 2 . The internal sensors 23 include a vehicle speed sensor, an acceleration sensor, and a yaw rate sensor as sensors for detecting the running state of the target vehicle 2 . A vehicle speed sensor is a detector that detects the speed of the target vehicle 2 . As the vehicle speed sensor, it is possible to use a wheel speed sensor that is provided for a wheel of the target vehicle 2 or a drive shaft that rotates integrally with the wheel and that detects the rotation speed of each wheel. The vehicle speed sensor transmits detected vehicle speed information (wheel speed information) to the automatic driving ECU 30 .

The acceleration sensor is a detector that detects acceleration of the target vehicle 2 . The acceleration sensor includes, for example, a longitudinal acceleration sensor that detects longitudinal acceleration of the target vehicle 2 . The acceleration sensor may include a lateral acceleration sensor that detects lateral acceleration of the target vehicle 2 . The acceleration sensor transmits acceleration information of the target vehicle 2 to the automatic driving ECU 30, for example. 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 automatic driving ECU 30 .

The internal sensor 23 detects at least one of tire air pressure, wiper operating state, and lighting device state as the vehicle state of the target vehicle 2 . The tire air pressure is the tire air pressure of the target vehicle 2 . The wiper operation state may include not only the presence or absence of the wiper operation but also the wiper operation speed. The lighting state includes the lighting state of the direction indicator. The lighting device state may include whether or not the headlights are lit and whether or not the fog lamps are lit.

In addition, the internal sensor 23 may detect the brake pressure of the hydraulic brake system from a brake pressure sensor as the vehicle state of the target vehicle 2. may be detected. The internal sensor 23 may detect the load state of each wheel from a wheel load sensor as the vehicle state of the target vehicle 2 . In addition, the internal sensor 23 may have a failure detection section that detects various failures of 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 operating unit of the target vehicle 2 is a device with which the driver inputs operations for driving the vehicle. The operation units of the target vehicle 2 include at least one of a steering unit, an accelerator operation unit, and a brake operation unit. The steering section is, for example, a steering wheel. The steering section is not limited to being wheel-shaped, and may be configured to function as a steering wheel. The accelerator operation unit is, for example, an accelerator pedal. A brake operating 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 so that the driver can input acceleration or deceleration.

The steering sensor detects the amount of operation of the steering section by the driver. The steering angle is included in the operation amount of the steering unit. The amount of operation of the steering unit may include steering torque. The accelerator sensor detects the amount of operation of the accelerator operating 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 in the form of detecting master cylinder pressure of a hydraulic brake system. The operation amount 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 operation amount of the driver to the automatic driving ECU 30 .

The map database 25 is a database that stores map information. The map database 25 is formed in a storage device such as an HDD mounted on the target vehicle 2, for example. The map information includes road position information, road shape information (for example, curvature information), position information of intersections and branch points, and the like. The map information may include traffic regulation information such as legal speed limits associated with the location information. The map information may include target information used for recognizing the position of the target vehicle 2 on the map. Targets can include lane markings, traffic lights, guardrails, 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. FIG.

The communication unit 26 is a communication device that controls wireless communication with the outside of the target vehicle 2 . Various types of information are transmitted and received via the network N. The communication unit 26 transmits various types of information to the information processing server 10 according to signals from the automatic driving ECU 30 .

The HMI 27 is an interface for inputting and outputting information between the automatic driving ECU 30 and the driver or passenger. The HMI 27 includes, for example, a display and speakers provided inside the vehicle. The HMI 27 performs image output from the display and audio output from the speaker according to the control signal from the automatic driving ECU 30 .

The actuator 28 is a device used for controlling the target vehicle 2 . Actuators 28 include at least drive actuators, brake actuators, and steering actuators. The drive actuator controls the amount of air supplied to the engine (throttle opening) according to a control signal from the automatic operation ECU 30, thereby controlling the driving force of the target vehicle 2. When 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 the motor as the power source to control the driving force. When the target vehicle 2 is an electric vehicle, a control signal from the automatic driving ECU 30 is input to the motor as the power source to control the driving force. A motor as a power source in these cases constitutes the actuator 28 .

The brake actuator controls the brake system according to a control signal from the automatic driving ECU 30 and controls the braking force applied to the wheels of the target vehicle 2 . A hydraulic brake system, for example, can be used as the brake system. The steering actuator controls driving of an assist motor that controls steering torque in the electric power steering system according to a control signal from the automatic driving ECU 30 . Thereby, the steering actuator controls the steering torque of the target vehicle 2 .

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

The target vehicle data acquisition unit 31 acquires target vehicle data, which is data relating to the target vehicle 2 . The target vehicle data includes the position information of the target vehicle 2 on the map and the running 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 on which the target vehicle 2 travels. 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 31 a acquires position information of the target vehicle 2 on the map based on the position information of the GPS reception unit 21 and the map information of the map database 25 . In addition, the vehicle position acquisition unit 31a obtains 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 result of the external sensor 22. may be obtained. The vehicle position acquisition unit 31a may recognize the lateral position of the target vehicle 2 with respect 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. . In addition, the vehicle position acquisition unit 31a may acquire the position information of the target vehicle 2 on the map by a well-known method.

The external environment recognition unit 31 b recognizes the external environment of the target vehicle 2 based on the detection result 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 relative velocities and moving directions 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 type of object can be identified by well-known techniques such as pattern matching. The external environment may include the result of lane marking recognition (white line recognition) around the target vehicle 2 . The external environment may include the outside temperature and may include the weather.

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

The driving operation information acquisition unit 31d acquires the driving operation information of the target vehicle 2 based on the detection result of the driving operation detection unit 24. FIG. 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 acquiring unit 31d stores the driving operation history for each driver who has undergone personal authentication. The driving operation history may be associated with the external environment and driving state of the target vehicle 2 . The automatic driving ECU 30 does not necessarily need to have the driving operation information acquiring section 31d. In this case, the driving operation detection unit 24 is also unnecessary.

The vehicle state recognition unit 31 e recognizes the vehicle state of the target vehicle 2 based on the detection result of the internal sensor 23 . Vehicle conditions may include tire pressure. The vehicle state may include a wiper operating state, a lighting device state, and a failure state of the target vehicle 2 . The automatic driving ECU 30 does not necessarily need to have the vehicle state recognition section 31e.

The route generation unit 32 generates a route [trajectory] used for automatic driving of the target vehicle 2 . The route generating unit 32 generates a route for automatic 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. .

A travel route is a route along which the target vehicle 2 travels in automatic driving. The route generator 32 obtains a travel route for automatic driving based on, for example, the destination, map information, and the position of the target vehicle 2 on the map. A travel route may be set by a well-known navigation system. The destination may be set by the occupant of the target vehicle 2, or may be automatically suggested by the automatic driving ECU 30, the navigation system, or the like.

The course includes a route along which the vehicle travels in automatic driving and a vehicle speed profile in automatic driving. A route is a trajectory along which a vehicle under automatic operation is scheduled to travel on a travel route. The route can be, for example, data (steering angle profile) of the steering angle change of the target vehicle 2 according to the position on the travel route. The position on the travel route is, for example, set vertical positions set at predetermined intervals (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 route generation unit 32 generates a route along which the vehicle travels, 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. The route generation unit 32 generates a route, for example, so that the target vehicle 2 passes through the center of the lane included in the travel route (the center in the lane width direction).

A steering torque profile in which a target steering torque is associated with each set vertical position may be used instead of the steering angle profile. Further, instead 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 the target position in the width direction of the lane. In this case, the set vertical position and the target horizontal position may be set as one position coordinate.

The vehicle speed profile is, for example, data in which a target vehicle speed is associated with each set vertical position. Note that the set vertical position may be set based on the running time of the vehicle instead of the distance. The set vertical position may be set as the arrival position of the vehicle after 1 second or the arrival position of the vehicle after 2 seconds.

The route generator 32 generates a vehicle speed profile based on speed-related information such as legal speeds included in the route and map information. Instead of the statutory speed, a set speed preset for a position or section on the map may be used. A route generator 32 generates a route for automatic driving from the route and the vehicle speed profile. It should be noted that the route generation method in the route generation unit 32 is not limited to the content described above, and other well-known methods can be adopted.

When receiving a notification from the information processing server 10 to change the travel route to avoid the unstable behavior position, the route generator 32 changes the route of the target vehicle 2 so as not to pass through the unstable behavior position. The route generating unit 32 may change the travel route (road on which the target vehicle 2 travels), and may move the target vehicle 2 so that it is separated from the unstable behavior position in the road width direction by a certain distance or more on the same road. You may change your route.

When receiving the notification of the unstable behavior position information and the stable traveling data from the information processing server 10, the route generation unit 32 prevents the target vehicle 2 from exhibiting unstable behavior based on the unstable behavior position information and the stable traveling data. Create a path as follows. The unstable behavior position information and the stable running data will be described later in detail. The route generation unit 32 generates a route so that the running state of the target vehicle 2 approaches the running state of the stable running data in the vicinity of the unstable behavior position. When receiving a stable driving instruction from the information processing server 10 instead of the stable driving data, the route generating unit 32 generates a route according to the stable driving instruction. The stable running instruction will be described later in detail.

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

When the automatic driving control unit 33 receives an instruction to cancel the automatic driving from the information processing server 10, the automatic driving control unit 33 cancels the automatic driving at the unstable behavior position. The automatic driving control unit 33 notifies the driver via the HMI 27 of the transition to manual driving. After notifying the driver, the automatic driving control unit 33 cancels automatic driving and shifts to manual driving by the driver. Also, the automatic driving control unit 33 notifies the driver of the unstable behavior position information received from the information processing server 10 through the HMI 27 .

Note that the target vehicle 2 does not necessarily have to be an automatically driven vehicle. In this case, the ECU of the target vehicle 2 does not need to have the route generator 32 and the automatic driving controller 33 . The ECU of the target vehicle 2 only needs to have an information providing unit capable of notifying the driver of unstable behavior position information and the like through the HMI 27 . When the ECU of the target vehicle 2 receives the notification of the stable running data from the information processing server 10, the ECU of the target vehicle 2 provides driving assistance so that the running state of the target vehicle 2 approaches the running state of the stable running data in the vicinity of the unstable behavior position. You may have the driving assistance part which performs. The method of driving support is not particularly limited, and may be information provision to the driver or control of vehicle travel.

[Configuration of information processing server]
The information processing server 10 is provided in a facility such as an information management center, for example, 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. As shown in FIG. 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 controls the information processing server 10 by operating an operating system, for example. The processor 11 is an arithmetic unit such as a CPU [Central Processing Unit] including a control device, an arithmetic device, registers, and the like. The processor 11 controls the storage unit 12 , the communication unit 13 and the user interface 14 . The storage unit 12 includes at least one of memory and storage. The memory is a recording medium such as ROM [Read Only Memory] and RAM [Random Access Memory]. The storage is a recording medium such as an HDD [Hard Disk Drive].

The communication unit 13 is communication equipment for performing communication via the network N. FIG. A network device, a network controller, a network card, or the like can be used for 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 have to be installed in a facility, and may be installed in a mobile object such as a vehicle or ship.

The information processing server 10 is also connected to a storage database 15 . The storage database 15 is a database for storing unstable behavior position information and the like. The storage database 15 may have a configuration similar to well-known databases on HDDs. Note that the storage database 15 may be provided in a facility or the like remote from the information processing server 10 .

Next, a functional configuration of the processor 11 will be described. As shown in FIG. 4, the processor 11 includes a target vehicle data recognition unit 11a, an unstable behavior position recognition unit 11b, a situation determination unit 11c, a memory processing unit 11d, a model generation unit 11e, a support target vehicle determination unit 11f, a vehicle support It has a 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 11 a 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 the map and the running 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 on which the target vehicle 2 travels. The target vehicle data recognition unit 11a may acquire target vehicle data in association with a setting area described later.

The unstable behavior position recognition unit 11b recognizes the unstable behavior position, which is the position on the map where the target vehicle 2 exhibits unstable behavior, based on the target vehicle data acquired by the target vehicle data recognition unit 11a. Unstable behavior is behavior of the vehicle that makes the running of the vehicle unstable. Unstable behavior includes, for example, slip. Unstable behavior may include sudden deceleration or sudden steering angle changes. The unstable behavior may include lane departure of the subject vehicle 2 and may include excessive approach of the subject vehicle 2 to an object.

First, determination of unstable behavior will be described. The unstable behavior position recognition unit 11b determines whether or not the target vehicle 2 exhibits unstable behavior based on the target vehicle data. The unstable behavior position recognition unit 11b detects, 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, and the driving detected by the steering sensor. Based on at least one of the driver's steering angle, the driver's brake operation amount detected by the brake sensor, and the brake pressure of the brake pressure sensor, it is determined that the target vehicle 2 has slipped as unstable behavior. A master cylinder pressure of a hydraulic brake system may be used instead of the brake operation amount of the brake sensor.

The unstable behavior position recognizing section 11b may use an operation start condition of a well-known antilock brake system [ABS: Antilock Brake System] as a slip determination. For example, an anti-lock braking system operates when, as an example, the wheel speed of each wheel is compared to an estimated vehicle speed to identify a wheel that is believed to be locked. The estimated vehicle body speed may be obtained from the wheel speed of each wheel until the vehicle slips, or from the change in acceleration until the vehicle slips.

In addition, the unstable behavior position recognition unit 11b may use an operation start condition of a well-known vehicle stability control system [VSC: Vehicle Stability Control] as a slip determination, or a well-known traction control [TRC: Traction Control System]. may be used. Traction control can also be activated if the wheel speed of each wheel is compared to the estimated vehicle speed to identify a spinning wheel. The unstable behavior position recognition unit 11b may determine the slip of the target vehicle 2 by other well-known methods.

The unstable behavior position recognition unit 11b may determine whether or not the subject 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 is greater than or equal to the rapid deceleration threshold. The rapid deceleration threshold is a preset threshold value. A threshold value used in the following description means a preset threshold value.

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 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 the target vehicle 2 undergoes a sudden steering angle change when, for example, the yaw rate becomes equal to or greater than the steering angle change threshold. Note that the tire steering angle may be used instead of the yaw rate.

The unstable behavior position recognition unit 11b determines whether the target vehicle 2 has deviated from the lane as an unstable behavior based on the lateral position of the target vehicle 2 or the external environment of the target vehicle 2 when the direction indicator is not lit. It may be determined whether In this case, the unstable behavior position recognition unit 11b determines lane deviation from the lateral position of the target vehicle 2, for example. Alternatively, the unstable behavior position recognizing unit 11b may determine lane deviation when recognizing from the external environment of the target vehicle 2 that the target vehicle 2 has straddled a lane marking.

The unstable behavior position recognition unit 11b determines whether or not the target vehicle 2 has excessively approached an object as an unstable behavior based on the running state of the target vehicle 2 and the external environment of the target vehicle 2. good. In this case, if the target vehicle 2 is moving at a low speed, the unstable behavior position recognition unit 11b does not exhibit an unstable behavior even if the distance to the object is small. When the time to collision [TTC: Time To Collision] between 2 and the object is equal to or less than the TTC threshold value, it is determined that the target vehicle 2 has come too close to the object. A time headway (THW) or a distance may be used instead of the time to collision.

The determination as to whether or not the target vehicle 2 has become unstable may be performed each time the target vehicle data is obtained, or may be performed collectively at regular intervals or at regular intervals. The determination as to whether or not 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 of the target vehicle 2 on the map when the target vehicle 2 exhibits unstable behavior. The unstable behavior position recognition unit 11b recognizes the unstable behavior position when it is determined that the target vehicle 2 exhibits unstable behavior.

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

Note that the unstable behavior position may be recognized as a section or area instead of a point on the map. When the target vehicle 2 skids while slipping, the unstable behavior position recognizing unit 11b may set the slip starting position to the unstable behavior position, and the target vehicle 2 moves in a state where it is determined to be slipping. All sections may be recognized as unstable behavior positions. The area may be a range within a certain distance centering on the target vehicle 2 that has slipped, or may be an area or zone in which the target vehicle 2 is traveling. The same is true for other unstable behaviors.

Based on the presence or absence of unstable behavior of the plurality of target vehicles 2 at the unstable behavior position recognized by the unstable behavior position recognition unit 11b, the situation determination unit 11c determines whether or not the unstable behavior position is in a continuous occurrence state. Determine whether it is a continuous situation.

The situation determination unit 11c determines whether the target vehicle 2 has passed through the unstable behavior position based on, for example, 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. Determine whether or not When determining that the target vehicle 2 has passed through the unstable behavior position, the situation determination unit 11c determines whether the unstable behavior position is in a continuous occurrence state or discontinuously based on the presence or absence of the unstable behavior of the target vehicle 2. Determine if the situation is In addition, the situation determination unit 11c may perform the above determination by collectively processing a plurality of target vehicle data for each fixed period.

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 the unstable behavior has occurred due to individual vehicle factors of the target vehicle 2 is low, and the possibility that the unstable behavior has occurred due to external factors such as the road environment increases. can. A non-continuous situation is a situation that is not a continuous occurrence situation. In the case of discontinuous situations, it can be considered that the possibility of unstable behavior occurring due to individual vehicle factors of the target vehicle 2 increases. If the situation determination unit 11c does not determine that the unstable behavior position is in the continuous occurrence state, it determines that the unstable behavior position is in the non-continuous state.

FIG. 5(a) is a diagram for explaining an example of a continuous occurrence situation. As shown in FIG. 5( a ), as an example, when two target vehicles 2A and 2B continuously exhibit unstable behavior at the unstable behavior position D, the situation determination unit 11c determines that the unstable behavior Determine that the position is in a continuous occurrence situation. FIG. 5(b) is a diagram for explaining an example of a discontinuous situation. Even if the subject vehicle 2A behaves unstable at the unstable behavior position D, the situation determination unit 11c determines that when the following subject vehicle 2B passes without exhibiting unstable behavior, the unstable behavior position is It may be determined that the situation is discontinuous.

In addition, the situation determined as the continuous occurrence situation is not limited to the situation shown in FIG. 5(a). The situation determination unit 11c may determine that the unstable behavior position D is continuously occurring when the three target vehicles 2A to 2C continuously exhibit unstable behavior. The situation determination unit 11c may determine that the unstable behavior position D is continuously occurring when four or more target vehicles 2 continuously exhibit unstable behavior. When all of the plurality of target vehicles 2 passing through the unstable behavior position D within a certain period of time exhibit unstable behavior, the situation determination unit 11c determines that the unstable behavior position D is in a state of continuous occurrence. You may

Even if there is one target vehicle 2 that does not exhibit unstable behavior, the situation determination unit 11c determines that the unstable behavior position D is in a continuous occurrence situation when unstable behavior occurs in the target vehicle 2 before and after it. It may be determined that there is Specifically, even if the middle target vehicle 2B among the three target vehicles 2A to 2C does not exhibit unstable behavior and passes through the unstable behavior position D, the situation determination unit 11c determines that the target vehicle 2A and the target vehicle When 2C becomes an unstable behavior, it may be determined that the unstable behavior position D is in a state of continuous occurrence. Alternatively, even if there are a plurality of target vehicles 2 that did not exhibit unstable behavior, if the number of target vehicles 2 exhibiting unstable behavior within a certain period of time is equal to or greater than the threshold, the situation determination unit 11c It may be determined that the unstable behavior position D is continuously occurring.

The storage processing unit 11d causes the storage database 15 to store the unstable behavior position information related to the unstable behavior position recognized by the unstable behavior position recognition unit 11b. 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 when the determination is made by the situation determination unit 11c.

In addition, the storage processing unit 11d associates the unstable behavior position information with the target vehicle data of the target vehicle 2 that has traveled the unstable behavior position, and stores them in the storage database 15 . Note that the storage processing unit 11d does not necessarily have to store the target vehicle data of all the target vehicles 2 that have traveled through the unstable behavior position. The memory processing unit 11d may store the unstable running data and/or the stable running data acquired by the model generating unit 11e, which will be described later, in the storage database 15 in association with the unstable behavior position.

The model generation unit 11e generates a behavior discrimination model from the target vehicle data of the target vehicle 2 that exhibits unstable behavior at the unstable behavior position for each of a plurality of set areas set on the map. A set area is an area arbitrarily set on the map. The set area may be, for example, a region or a prefecture (Tokyo or Osaka), a state or a group. The set area may be a municipality. The set area may be an area of any shape set on the map (for example, a rectangular area). Note that the plurality of setting areas do not overlap each other.

A behavior determination model is a vehicle behavior model used for determining (estimating) whether or not the target vehicle 2 behaves unstable at an unstable behavior position. The behavior discrimination model includes vehicle speed, longitudinal acceleration, steering angle, yaw rate, ABS [Antilock Brake System] operation duration, driving support (vehicle stability control system, etc.) operation status, and maximum steering angle before unstable behavior. contains at least one feature parameter among The behavior discrimination model may include feature parameters related to the surrounding environment, such as outside temperature, estimated road friction, road gradient, wiper operation state, lighting device state, etc. Time or time period is included as feature parameters. may be The behavior discrimination model may be generated so as to include the median value of feature quantity parameters such as vehicle speed of a large number of target vehicles 2 exhibiting unstable behavior, or each feature quantity parameter may be obtained by predetermined arithmetic processing. good. Various well-known techniques related to vehicle behavior model generation can be used to generate the behavior discrimination model.

Different models may be generated as the behavior discrimination model according to the determination result of the situation determination unit 11c. As for the behavior discrimination model, 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 may be generated respectively.

Different behavior discrimination models may be generated according to the location attribute of the unstable behavior position. The location attribute may be distinguished as different attributes for exclusive roads and general roads, and may be distinguished as different attributes for narrow roads and general roads 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 according to individual vehicle attributes such as vehicle type. The individual vehicle attributes may be classified as different attributes for large-sized vehicles, medium-sized vehicles, and small-sized vehicles, or may be classified as different attributes for freight vehicles and passenger vehicles. Various attributes that can be determined from the vehicle specifications can be used as the 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 determination model is a vehicle behavior model generated from the target vehicle data of the target vehicle 2 that exhibits unstable behavior at the unstable behavior position. The stable behavior determination model is a vehicle behavior model generated from the target vehicle data of the target vehicle 2 that does not exhibit unstable behavior at the unstable behavior position.

The model generation unit 11e acquires the unstable running data for each set area from the target vehicle data of the target vehicle 2 that exhibits unstable behavior at the unstable behavior position, for example. The unstable running data is data related to the past running state of the target vehicle 2 in which unstable behavior occurred at the unstable behavior position. The unstable running data can be, for example, data on the running state of the target vehicle 2 from a position a certain distance before the unstable behavior position to the unstable behavior position. The unstable running data may be data on the running state of the target vehicle 2 during a certain period of time until unstable behavior occurs. The unstable travel data may include the travel locus of the target vehicle 2 . The model generation unit 11e may generate an unstable behavior determination model for the set area using the unstable travel data for each set area.

The model generation unit 11e may acquire the stable running data of the target vehicle 2 that has passed through the unstable behavior position without exhibiting unstable behavior for each set area. The stable running data is data related to the past running state of the target vehicle 2 that passed through an unstable behavior position in a discontinuous situation without causing an unstable behavior. The stable running data can be, for example, data on the running state of the target vehicle 2 from a position a certain distance before the unstable behavior position to the unstable behavior position. The stable running data may be data on the running state of the target vehicle 2 during a certain period of time until reaching the unstable behavior position. The stable travel data may include the travel locus of the target vehicle 2 . The model generation unit 11e may generate a stable behavior determination model for each set area using stable running data for each set area.

The support target vehicle determination unit 11f determines whether or not the 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. judge. The support target vehicle is the target vehicle 2 that is estimated to exhibit unstable behavior at the unstable behavior position within a certain period of time. The fixed time may be 5 minutes or 10 minutes. The certain period of time may be less than 5 minutes or 30 minutes or longer. The certain period of time is not particularly limited, and may be arbitrarily set by the server administrator.

Specifically, based on the unstable behavior position and the position information of the target vehicle 2 included in the target vehicle data, the support target vehicle determination unit 11f determines the candidate target vehicle 2 that will reach the unstable behavior position within a certain period of time. Determine whether a vehicle exists. A candidate vehicle is a vehicle that is a candidate for assistance. The traveling direction of the target vehicle 2 may be recognized based on the temporal change of the positional information, or may be recognized from the traveling direction of the lane included in the map information by matching the positional information and the map information.

When it is determined that there is a candidate vehicle that will reach the unstable behavior position within a certain period of time, the support target vehicle determination unit 11f estimates whether or not the candidate vehicle will exhibit unstable behavior at the unstable behavior position.

The support target vehicle determination unit 11f uses the 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 to determine whether the candidate vehicle behaves unstable at the unstable behavior position. Estimate whether or not to be The support target vehicle determination unit 11f performs determination using a behavior determination model (for example, an unstable behavior determination model and/or a stable behavior determination model) in a set area including an unstable behavior position. The support target vehicle determination unit 11f may estimate whether or not the candidate vehicle exhibits unstable behavior at the unstable behavior position without using the behavior determination model.

The support target vehicle determination unit 11f is based on the determination result of the situation determination unit 11c, the location attribute of the unstable behavior position (distinction between motorway and general road), and the individual vehicle attribute of the candidate vehicle (distinction between freight vehicle and passenger car). You may select a model to use for The support target vehicle determination unit 11f may further consider the surrounding environment such as outside temperature and weather information to select a model.

The support target vehicle determination unit 11f determines, for example, which of the unstable behavior determination model and the stable behavior determination model the running state of the candidate vehicle is closer to. The support target vehicle determination unit 11f determines the value of the feature amount parameter included in the unstable behavior determination model and the value of the feature amount parameter included in the stable behavior determination model for arbitrary parameters such as the vehicle speed, acceleration, and yaw rate of the candidate vehicle. Determine which one is closest. The support target vehicle determination unit 11f estimates that the candidate vehicle exhibits unstable behavior at the unstable behavior position when it determines that the running state of the candidate vehicle is close to the unstable behavior determination model. The support target vehicle determination unit 11f determines that the support target vehicle exists when it is estimated that the candidate vehicle exhibits unstable behavior at the unstable behavior position.

The support target vehicle determination unit 11f may estimate whether or not the candidate vehicle exhibits unstable behavior at the unstable behavior position using a neural network (deep learning network) learned by reflecting the behavior discrimination model. As an example of a neural network, CNN [Convolutional neural networks] can be used. A neural network learning method is not particularly limited, and a known method such as back propagation can be used. Note that the support target vehicle determination unit 11f may employ well-known machine learning (for example, support vector machine, boosting, etc.) other than the neural network. The behavior discrimination model may be used in some way to determine the support target vehicle.

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

As vehicle support, the vehicle support unit 11g provides notification of unstable behavior position information that is information about the unstable behavior position, travel route change support, and information regarding the running state of the target vehicle 2 that did not exhibit unstable behavior at the unstable behavior position. At least one of the notification of the stable running data and the stable running instruction to bring the running state of the support target vehicle closer to the running state of the stable running data is performed.

The vehicle support unit 11g, for example, notifies the support target vehicle of the unstable behavior position information. The support target vehicle can suppress unstable behavior by using the unstable behavior position information. Further, the vehicle support unit 11g may instruct the support target vehicle to change the travel route so as to avoid the unstable behavior position. The vehicle to be supported avoids the occurrence of unstable behavior at the unstable behavior position by changing the route (road) or the travel route within the road.

The vehicle support unit 11g may notify the support target vehicle of stable running data regarding the running state of the target vehicle 2 that did not exhibit unstable behavior at the unstable behavior position. The vehicle to be supported can suppress unstable behavior by referring to the stable running data.

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

Note that, when the support target vehicle is automatically driving, the vehicle support unit 11g may instruct to cancel the automatic driving at the unstable behavior position together with the notification of the unstable behavior position information. By canceling automatic driving and shifting to manual driving by the driver, the support target vehicle can avoid unstable behavior of the support target vehicle at an unstable behavior position while in automatic driving.

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

The statistics include vehicle traffic statistics or unstable behavior statistics. First, the case of vehicle traffic statistics will be described.

A vehicle traffic statistic value is a value related to the traffic of the target vehicle 2 within the set area. The vehicle traffic statistic value can be, for example, the number of times the target vehicle 2 has traveled within the set area in a certain period of time. The statistical value calculation unit 11h may calculate the vehicle traffic statistical value by distinguishing the number of times of traffic for each vehicle, or may calculate the statistical value of vehicle traffic by distinguishing the number of times of traffic by driver.

Alternatively, the statistic value calculation unit 11h may calculate the vehicle traffic statistic value based on the conditions of the divided areas obtained by dividing the set area into a plurality of areas. A divided area is an area obtained by dividing a setting area into a plurality of areas. FIG. 6 is a diagram showing an example of setting areas and divided areas. In FIG. 6, Tokyo is shown as a set area A, and each ward is shown as divided areas B1 to B4. Note that FIG. 6 shows only a part of the divided areas. The divided area does not have to be a municipality, and can be an area of any shape as long as it is within one set area. The divided areas do not overlap each other. A blank area that is not set as a divided area may exist within the set area.

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 traffic 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. The statistical value calculation unit 11h may calculate the vehicle traffic calculation value considering the number of times each vehicle passes, or may calculate the vehicle traffic calculation value considering the number of times each driver passes. Note that the fixed period may be a different period for each calculation target, or may be common regardless of the calculation target.

The statistical value calculator 11h counts the number of divided areas in which the vehicle traffic calculated value is greater than or equal to 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 value. For the first divided area threshold, a common value may be set within one set area, or a different value may be set 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 the divided areas in which the vehicle traffic calculated value is equal to or greater than the first divided area threshold value. The statistical value calculator 11h may directly calculate the number of divided areas in which the vehicle traffic calculated value is greater than or equal to the first divided area threshold value as the vehicle traffic statistical value.

The statistical value calculator 11h may calculate the vehicle traffic statistical value from the number of divided areas in which the vehicle traffic calculated value is greater than or equal to the first divided area threshold using a predetermined arithmetic expression. The statistic value calculation unit 11h sums up the evaluation values of each divided area in which the vehicle traffic calculation value is equal to or greater than the first divided area threshold within the set area, using the evaluation value set in advance for each divided area. Values may be calculated as vehicle traffic statistics. The evaluation value is set in advance according to, for example, the position or area of each divided area. The evaluation value may be set to a value that increases as the area of the divided area increases.

The statistical value calculator 11h may calculate the vehicle traffic statistical value based on the road conditions in the set area instead of the divided areas. FIG. 7 is a diagram for explaining an example of a road. FIG. 7 shows roads C1 to C5 in the divided area B1. The roads C1 to C5 are also roads within the set area A in FIG. Note that the divided areas do 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 distances, or may be recognized as being divided at arbitrary intersections. The vehicle traffic calculation value for each road is a value related to traffic of the target vehicle 2 for each road. The vehicle traffic calculation value for each road can be, for example, the number of times the target vehicle 2 travels for each road in a certain period of time.

The statistical value calculator 11h counts the number of roads with the vehicle traffic calculated value equal to or greater than the first road threshold. The first road threshold is a preset value threshold. Any value can be adopted as the first road threshold. The first road threshold value 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 calculator 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 calculated value is equal to or greater than the first threshold value for roads. The statistical value calculation unit 11h may directly calculate the number of roads whose vehicle traffic calculated value is equal to or greater than the first road threshold value as the vehicle traffic statistical value.

The statistical value calculation unit 11h may calculate the vehicle traffic statistical value from the number of roads whose vehicle traffic calculated value is equal to or greater than the first road threshold using a predetermined arithmetic expression. The statistic value calculation unit 11h uses the evaluation value preset for each road to determine the total evaluation value of the roads in the set area where the vehicle traffic calculation value is equal to or greater than the first threshold value for vehicle traffic. It may be calculated as a statistical value. The evaluation value is set in advance according to, for example, the position or length of each road. The evaluation value may be a value that increases as the length of the road increases.

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

Based on the target vehicle data, the statistical value calculation unit 11h calculates the number of passing target vehicles 2 for each road in a certain period (see FIG. 7). The statistical value calculation unit 11h calculates the vehicle traffic calculation value of the divided area from the number of passing target vehicles 2 for each road in the divided area for a certain period of time. For example, the statistical value calculation unit 11h can directly calculate the number of roads on which the number of passing target vehicles 2 is greater than or equal to the first threshold for calculation within the divided area as the vehicle traffic calculation value for the divided area. The first calculation threshold is a preset threshold value. Any value can be adopted as the first calculation threshold. The first calculation threshold value 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 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 greater than or equal to the first threshold for calculation, using a predetermined arithmetic expression. The statistic value calculation unit 11h uses an evaluation value preset for each road, and calculates a value obtained by summing the evaluation values of the roads having a vehicle traffic calculation value greater than or equal to the first threshold for calculation in the divided area. You may calculate as a vehicle traffic calculation value of an area.

The statistical value calculation unit 11h calculates the vehicle traffic statistical value of the set area from the vehicle traffic calculated values of a plurality of divided areas within the set area. For example, the statistical value calculator 11h can directly calculate the number of divided areas in the set area in which the 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 preset value threshold. Any value can be adopted as the second statistical threshold. The second statistical threshold value 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 may calculate the vehicle traffic statistical value of the set area from the number of divided areas in which the vehicle traffic calculated value is equal to or greater than the second statistical threshold using a predetermined arithmetic expression. The statistical value calculation unit 11h uses an evaluation value set in advance for each divided area to obtain a value obtained by summing the evaluation values of the divided areas in which the vehicle traffic calculation value is equal to or greater than the second statistical threshold within the set area. may be calculated as vehicle traffic statistics.

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

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

Alternatively, the statistic value calculation unit 11h may calculate the unstable behavior statistic value based on the conditions of the divided areas obtained by dividing the set area (see FIG. 6). The statistical value calculator 11h calculates an unstable behavior calculated 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 for a certain period of time. The unstable behavior calculation value for each divided area can be, for example, the number of times unstable behavior positions are recognized in the divided area or the number of occurrences of unstable behavior within a certain period. The statistical value calculation unit 11h may calculate the unstable behavior calculation value considering the number of times each vehicle passes, or may calculate the unstable behavior calculation value considering the number of times each driver passes.

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

The statistical value calculator 11h may calculate the unstable behavior statistical value from the number of divided areas in which the unstable behavior calculated value is greater than or equal to the second divided area threshold using a predetermined arithmetic expression. The statistical value calculation unit 11h uses an evaluation value set in advance for each divided area, and sums the evaluation values of the divided areas in which the unstable behavior calculation value is equal to or greater than the second divided area threshold within the set area. You may calculate the value which carried out as an unstable behavior statistic value. The evaluation value can be the same as in the case of vehicle traffic statistics.

The statistical value calculator 11h may calculate the unstable behavior statistical value based on the road conditions in the set area instead of the divided areas. The statistical value calculator 11h calculates an unstable behavior calculated value for each road based on the target vehicle data. Each road may be recognized as a different road at regular distances, or may be recognized as being divided at arbitrary intersections. The calculated unstable behavior value for each road is a value related to unstable behavior for each road during a certain period of time. The unstable behavior calculation value for each road can be, for example, the number of recognition times of unstable behavior positions or the number of occurrences of unstable behavior for each road in a certain period.

The statistical value calculator 11h counts the number of roads with the unstable behavior calculated value equal to or greater than the second road threshold. The second threshold for road is a preset value threshold. Any value can be adopted as the second road threshold. The second road threshold value 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 calculator 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 calculated value is equal to or greater than the second threshold value for roads. 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 threshold value for roads as the unstable behavior statistical value.

The statistical value calculation unit 11h may calculate the unstable behavior statistical value from the number of roads whose unstable behavior calculated value is equal to or greater than the second threshold value for roads using a predetermined arithmetic expression. The statistic value calculation unit 11h uses an evaluation value preset for each road, and calculates the value obtained by summing the evaluation values of the roads having the unstable behavior calculation value equal to or greater than the second threshold value for roads in the set area. It may be calculated as a stable behavior statistic. The evaluation value can be the same as in the case of vehicle traffic statistics.

Furthermore, the statistical value calculator 11h may calculate statistical values in two stages using divided areas and roads. The statistical value calculation unit 11h calculates an unstable behavior calculation value for the divided area from, for example, the number of unstable behaviors for each road in the divided area, and calculates an unstable behavior statistical value for the set area from the unstable behavior calculation value for the divided area. Calculate (see FIG. 7). The number of unstable behaviors for each road can be, for example, the number of recognition times of unstable behavior positions or the number of occurrences of unstable behaviors for each road in a certain period of time.

The statistical value calculator 11h calculates the number of unstable behaviors for each road in a certain period based on the target vehicle data. The statistical value calculator 11h calculates an unstable behavior calculation value for the divided area from the number of unstable behaviors for each road in the divided area. For example, the statistical value calculator 11h can directly calculate the number of roads with the number of unstable behaviors equal to or larger than the second threshold value for calculation in the divided area as the unstable behavior calculated value of the divided area. The second calculation threshold is a preset threshold value. Any value can be adopted as the second calculation threshold. The second calculation threshold value 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 calculate the unstable behavior calculation value of the divided area from the number of roads having the number of unstable behaviors equal to or larger than the second calculation threshold using a predetermined arithmetic expression. The statistic value calculation unit 11h uses an evaluation value preset for each road, and calculates a value obtained by summing the evaluation values of the roads in the divided area for which the unstable behavior calculation value is equal to or greater than the second threshold for calculation. You may calculate as an unstable behavior calculation value of a division|segmentation area.

The statistical value calculator 11h calculates the unstable behavior statistical value of the set area from the unstable behavior calculated values of a plurality of divided areas within the set area. For example, the statistical value calculator 11h can directly calculate the number of divided areas in 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 preset value threshold. Any value can be adopted as the second statistical threshold. The second statistical threshold value 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 calculator 11h may calculate the unstable behavior statistical value of the set area from the number of divided areas in which the unstable behavior calculated value is equal to or greater than the second statistical threshold using a predetermined arithmetic expression. The statistical value calculator 11h sums up the evaluation values of the divided areas in which the unstable behavior calculation value is equal to or greater than the second statistical threshold within the set area, using the evaluation value set in advance for each divided area. The value may be calculated as an unstable behavior statistic.

The model update necessity determination unit 11j determines whether the vehicle traffic statistics have changed since the behavior discrimination model was generated in the set area, or the unstable behavior statistics have changed since the behavior discrimination model was generated in the set area. Based on the change, it is determined whether or not the behavior discrimination model of the set area needs to be updated. The model update necessity determination unit 11j determines whether or not the model update is necessary on a ratio basis based on the change from when the behavior discrimination model was generated in the set area.

For example, the model update necessity determination unit 11j determines that the amount of change in the latest vehicle traffic statistic value in the set area with respect to the vehicle traffic statistic value when the behavior discrimination model is generated in the set area is greater than or equal to the first model update threshold. case, it is determined that the behavior discrimination model of the setting area needs to be updated. The first model update threshold is a preset value threshold. A different value may be set for each set area as the first model update threshold. The first model update threshold may be changed according to the elapsed time since the model update. The first model update threshold may be changed so that the longer the elapsed time from the model update, the smaller the value.

The vehicle traffic statistic value when the behavior discrimination model is generated in the set area is the vehicle traffic statistic value at the timing when the behavior discrimination model is generated by the model generation unit 11e among the vehicle traffic statistic values repeatedly calculated at regular intervals. It's a statistic. The amount of change is an absolute value. The model update necessity determination unit 11j determines whether the vehicle traffic statistic value when the behavior discrimination model was generated in the set area or the latest vehicle traffic statistic value is a larger value, and the model update is performed when the change amount is the first model update. If it is equal to or greater than the threshold, it is determined that the behavior discrimination model needs to be updated.

Alternatively, the model update necessity determination unit 11j determines that the amount of change in the latest unstable behavior statistic value in the set area with respect to the unstable behavior statistic value when the behavior discrimination model is generated in the set area is the second model update threshold. In the case above, it is determined that the behavior discrimination model of the setting area needs to be updated. The second model update threshold can be set similarly to the first model update threshold. The unstable behavior statistic when the behavior discrimination model is generated in the set area is the same as the vehicle traffic statistic. The model update necessity determination unit 11j may determine whether or not the model update is necessary using a neural network or the like, based on changes in statistical values.

The model update unit 11k updates the behavior determination model of the set area when the model update necessity determination unit 11j determines that the behavior determination model of the set area needs to be updated. 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 behavior discrimination model can be updated in the same manner as the model generation by the model generation unit 11e.

The model updating unit 11k may update the behavior discrimination model using only the target vehicle data after the behavior discrimination model is generated, and the target vehicle data before the behavior discrimination model is generated and the behavior discrimination model are The behavior discrimination model may be updated using both of the generated target vehicle data.

[program]
The program comprises the processor 11 of the information processing server 10, the target vehicle data recognition unit 11a, the unstable behavior position recognition unit 11b, the situation determination unit 11c, the storage processing unit 11d, the model generation unit 11e, the support target vehicle It functions (operates) as a determination unit 11f, a vehicle support unit 11g, a statistical value calculation unit 11h, a model update necessity determination unit 11j, and a model update unit 11k. A program is provided by non-temporary recording media, such as ROM or semiconductor memory, for example. Also, the program may 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 this embodiment will be described with reference to the drawings. FIG. 8 is a flowchart showing an example of unstable behavior position storage processing. The unstable behavior position storage processing is executed, for example, when the information processing server 10 manages the target vehicle 2 .

As shown in FIG. 8, the information processing server 10 recognizes the target vehicle data by the target vehicle data recognition unit 11a as S10 (target vehicle data recognition step). The target vehicle data recognition unit 11 a recognizes 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 running state of the target vehicle 2 . After that, the information processing server 10 shifts to S12.

In S12, the information processing server 10 determines whether or not the target vehicle 2 exhibits unstable behavior by the unstable behavior position recognition unit 11b (unstable behavior determination step). The unstable behavior position recognition unit 11b determines whether or not the target vehicle 2 exhibits unstable behavior based on the target vehicle data. The unstable behavior position recognition unit 11b determines whether or not the object vehicle 2A has slipped as an unstable behavior based on the wheel speed of each wheel of the object vehicle 2, for example, thereby determining the presence or absence of the unstable behavior. When the information processing server 10 does not determine that the target vehicle 2 has exhibited unstable behavior (S12: NO), the information processing server 10 ends the current unstable behavior position information storage processing. The information processing server 10 proceeds to S14 when it is determined that the target vehicle 2 has become unstable (S12: YES).

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 recognizing unit 11b recognizes the unstable behavior position based on the position information of the target vehicle 2 on the map when it is determined that the target vehicle 2 exhibits unstable behavior. 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. FIG. 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 when the determination is made by the situation determination unit 11c. The storage processing unit 11d may associate the unstable behavior position information with the target vehicle data of the target vehicle 2 that has traveled the unstable behavior position, and store them in the storage database 15 .

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

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

In S22, the information processing server 10 uses the model generation unit 11e to generate a behavior discrimination model in the set area (model generation step). The model generation unit 11e generates a behavior discrimination model from the target vehicle data of the target vehicle 2 that exhibits unstable behavior at the unstable behavior position for each of a plurality of set areas set on the map.

FIG. 9B is a flow chart showing an example of the vehicle assistance execution process. The vehicle assistance execution process is periodically executed, for example, for each target vehicle 2 that is running.

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

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

In S34, the information processing server 10 executes vehicle support for the vehicle to be supported by the vehicle support unit 11g (vehicle support step). As vehicle support, the vehicle support unit 11g provides notification of unstable behavior position information that is information about the unstable behavior position, travel route change support, and information regarding the running state of the target vehicle 2 that did not exhibit unstable behavior at the unstable behavior position. At least one of the notification of the stable running data and the stable running instruction to bring the running state of the support target vehicle closer to the running state of the stable running data is performed. After that, the information processing server 10 ends the current process.

FIG. 10 is a flowchart illustrating an example of model update processing. Model update processing is performed periodically. The model update process may be executed each time the statistical value of the set area is calculated.

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

In S42, the information processing server 10 uses the model update necessity determination unit 11j to determine whether or not the behavior discrimination model of the set area needs to be updated based on the change in the statistical value from the time of model refinement (model update required). negative judgment step). For example, the model update necessity determination unit 11j determines that the amount of change in the latest vehicle traffic statistic value in the set area with respect to the vehicle traffic statistic value when the behavior discrimination model is generated in the set area is greater than or equal to the first model update threshold. case, it is determined that the behavior discrimination model of the setting area needs to be updated. Alternatively, the model update necessity determination unit 11j determines that the amount of change in the latest unstable behavior statistic value in the set area with respect to the unstable behavior statistic value when the behavior discrimination model is generated in the set area is the second model update threshold. In the case above, it is determined that the behavior discrimination model of the setting area needs to be updated.

When the information processing server 10 determines that the behavior determination model of the setting area needs to be updated (S42: YES), the process proceeds to S44. When the information processing server 10 does not determine that updating of the behavior determination model of the setting area is necessary (S42: NO), the current process ends.

In S44, the information processing server 10 updates the behavior determination model of the set area by the model updating unit 11k (model updating 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 is generated in the set area. After that, the information processing server 10 ends the current process.

FIG. 11A is a flowchart showing an example of statistical value calculation processing using divided areas. As shown in FIG. 11A, in S50, the information processing server 10 calculates the vehicle traffic calculation value or the unstable behavior calculation value for each divided area in the 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 unstable behavior positions are recognized or the number of times unstable behavior occurs within a certain period of time.

In S52, the information processing server 10 calculates the statistical value (vehicle traffic statistical value or unstable behavior statistical value) of the set area by the statistical value calculation unit 11h (statistic value calculation step). The statistical value calculation unit 11h calculates the vehicle traffic statistical value of the set area including the divided area from the vehicle traffic calculated value of each divided area. Alternatively, the statistical value calculator 11h calculates the unstable behavior statistical value of the set area including the divided area from the unstable behavior calculated value for each divided area. After that, the information processing server 10 ends the current process.

FIG. 11B is a flowchart showing an example of statistical value calculation processing using roads. As shown in FIG. 11B, in S60, the information processing server 10 calculates the vehicle traffic calculation value or the 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 travels on the road in a certain period of time. The unstable behavior statistic value for each road may be the number of recognition times of unstable behavior positions or the number of occurrences of unstable behavior of the road 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) of the set area by the statistical value calculation unit 11h (statistic value calculation step). The statistical value calculation unit 11h calculates the vehicle traffic statistical value of the set area from the vehicle traffic calculated value for each road. Alternatively, the statistical value calculator 11h calculates the unstable behavior statistical value of the set area from the unstable behavior calculated value for each road. After that, the information processing server 10 ends the current process.

In the information processing server 10 (and the processing method and program of the information processing server 10) according to the present embodiment described above, the vehicle traffic statistics in the set area are calculated based on the target vehicle data, and the behavior discrimination model is generated in the set area. It is determined whether or not the behavior discrimination model of the set area needs to be updated based on the change in the vehicle traffic statistics value since it was generated. Alternatively, the information processing server 10 calculates the unstable behavior statistic value in the set area based on the target vehicle data, and based on the change in the unstable behavior statistic value from when the behavior discrimination model is generated in the set area, It is determined whether or not the behavior discrimination model of the set area needs to be updated.

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 setting area, compared to the case where the model is updated periodically without considering the situation of the setting area. can be done. As a result, it is possible to prevent the learned behavior discrimination model from diverging from the market environment. In addition, the load on the information processing server 10 and the operation cost can be reduced compared to the case where the model is updated periodically with high frequency.

Furthermore, in this information processing server 10, instead of making a determination based only on the latest vehicle traffic statistical value or unstable behavior statistical value, a rate-based determination is made by focusing on the change from when the behavior discrimination model was generated in the set area. , it is possible to update the behavior discrimination model in consideration of changes in the situation of the set area (market changes, etc.) since the behavior discrimination model was generated.

Further, according to the information processing server 10, the amount of change in the latest vehicle traffic statistic value in the set area with respect to the vehicle traffic statistic value when the behavior discrimination model was generated in the set area is greater than or equal to the first model update threshold, or , when the amount of change in the latest unstable behavior statistic value of the setting area with respect to the unstable behavior statistic value when the behavior discrimination model is generated in the setting area is greater than or equal to the second model update threshold, the setting area Since it is determined that the behavior discrimination model needs to be updated, the behavior discrimination model can be updated at an appropriate timing according to the situation of the set area by setting the threshold value.

Further, according to the information processing server 10, it is also possible to calculate the statistical value of the set area based on the situation of each divided area obtained by dividing the set area into a plurality of areas or the situation of each road in the set area. Even if extreme values occur in some areas, they can be gently 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-described embodiments. The present invention can be embodied in various forms with various modifications and improvements based on the knowledge of those skilled in the art, including the embodiment described above.

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 determination model based on data that has been determined by another server.

2, 2A to 2Z ... target vehicle 10 ... information processing server 11 ... processor 11b ... unstable behavior position recognition section 11e ... model generation section 11f ... support target vehicle determination section 11g ... vehicle support section 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 to B4 division Areas, C1 to C5: Road, D: Unstable behavior position.

Claims (6)

An information processing server for obtaining target vehicle data including the driving state of the target vehicle and the position information of the target vehicle on a map from a plurality of target vehicles,
an unstable behavior position recognition unit that recognizes, based on the target vehicle data, an unstable behavior position, which is a position on a map where the target vehicle exhibits unstable behavior;
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 set on a map;
The target estimated to exhibit unstable behavior 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 support target vehicle determination unit that determines whether or not a support target vehicle that is a vehicle exists;
a vehicle support unit that provides vehicle support to the support target vehicle when the support target vehicle determination unit determines that the support target vehicle exists;
a statistic value calculation unit that calculates a vehicle traffic statistic value in the set area or an unstable behavior statistic value in the set area based on the target vehicle data;
Based on changes in the vehicle traffic statistics since the behavior discrimination model was generated in the set area, or changes in the unstable behavior statistics since the behavior discrimination model was generated in the set area a model update necessity determination unit that determines whether or not the behavior determination model in the setting area needs to be updated;
a model update unit that updates the behavior discrimination model in the setting area when the model update necessity determination unit determines that the behavior discrimination model in the setting area needs to be updated;
Information processing server. The model update necessity determination unit determines that a change amount of the latest vehicle traffic statistic value in the set area with respect to the vehicle traffic statistic value when the behavior determination model is generated in the set area is a first model update threshold. or more, or the amount of change in the latest unstable behavior statistic value in the set area with respect to the unstable behavior statistic value when the behavior discrimination model is generated in the set area is a second model update threshold 2. The information processing server according to claim 1, which determines that said behavior discrimination model of said setting area needs to be updated in the above cases. The statistic value calculation unit calculates a vehicle traffic calculation value for each divided area obtained by dividing the set area based on the target vehicle data, and calculates the vehicle traffic calculation value for each of the divided areas. calculating the vehicle traffic statistic value of the set area from the number of the divided areas in which the difference between the traffic calculated value and the latest vehicle traffic calculated value is equal to or greater than a first divided area threshold;
Alternatively, an unstable behavior calculation value is calculated for each divided area obtained by dividing the set area based on the target vehicle data, and the unstable behavior calculation value is obtained when the behavior discrimination model is generated in the set area. 3. The unstable behavior statistic value of the set area is calculated from the number of the divided areas in which the difference between the calculated value and the latest unstable behavior calculation value is equal to or greater than a second divided area threshold value. information processing server. 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 the vehicle traffic calculation value when the behavior discrimination model is generated in the set area. calculating the vehicle traffic statistics value of the set area from the number of the roads whose difference from the latest vehicle traffic calculation value is equal to or greater than the first threshold for roads;
Alternatively, an unstable behavior calculated value for each road in the set area is calculated based on the target vehicle data, and the unstable behavior calculated value when the behavior discrimination model is generated in the set area and the latest the 3. The information processing server according to claim 1, wherein the unstable behavior statistic value of the set area is calculated from the number of the roads whose difference from the unstable behavior calculated value is equal to or greater than the second road threshold. A processing method for an information processing server for obtaining target vehicle data including the driving state of the target vehicle and the position information of the target vehicle on a map from a plurality of target vehicles,
an unstable behavior position recognition step of recognizing an unstable behavior position, which is a position on a map where the target vehicle exhibits unstable behavior, 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;
The target estimated to exhibit unstable behavior 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 support target vehicle determination step of determining whether or not a support target vehicle that is a vehicle exists;
a vehicle support step of providing vehicle support 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;
Based on changes in the vehicle traffic statistics since the behavior discrimination model was generated in the set area, or changes in the unstable behavior statistics since the behavior discrimination model was generated in the set area a model update necessity determination step of determining whether or not the behavior discrimination model in the setting area needs to be updated;
a model update step of updating the behavior discrimination model in the setting area when it is determined in the model update necessity determination step that the behavior discrimination model in the setting area needs to be updated;
A processing method of an information processing server including A program for operating an information processing server for obtaining target vehicle data including the 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 exhibits unstable behavior, 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 set on the map;
The target estimated to exhibit unstable behavior 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 support target vehicle determination unit that determines whether or not a support target vehicle that is a vehicle exists;
a vehicle support unit that provides vehicle support to the support target vehicle when the support target vehicle determination unit determines that the support target vehicle exists;
A statistic value calculation unit that calculates a vehicle traffic statistic value in the set area or an unstable behavior statistic value in the set area based on the target vehicle data;
Based on changes in the vehicle traffic statistics since the behavior discrimination model was generated in the set area, or changes in the unstable behavior statistics since the behavior discrimination model was generated in the set area a model update necessity determination unit that determines whether or not the behavior discrimination model in the setting area needs to be updated;
and an information processing server as a model update unit that updates the behavior discrimination model in the setting area when the model update necessity determination unit determines that the behavior discrimination model in the setting area needs to be updated. program that runs the

Images