JP7070827B2 - Driving evaluation device, in-vehicle device, driving evaluation system equipped with these, driving evaluation method, and driving evaluation program - Google Patents

Description

The present invention relates to a driving evaluation device, an in-vehicle device, a driving evaluation system including these, a driving evaluation method, and a driving evaluation program.

Patent Document 1 discloses an automatic driving behavior evaluation system that evaluates the driving behavior of a driver of an automobile with respect to a predetermined location.
The driving behavior automatic evaluation system described in Patent Document 1 is configured to include a measurement control computer and an analysis computer, and the analysis computer stores information on dangerous points in advance. The information on the dangerous place includes a name, coordinates, features, and the like, and also stores definition data of an accident prevention operation set for each dangerous place.

The driving behavior automatic evaluation system described in Patent Document 1 is based on the angular velocity data detected by the angular velocity sensor mounted on the driver's head, the position data of the automobile, and the definition data, and the accident prevention is performed at the dangerous place. It is configured to evaluate whether or not the operation has been performed correctly.

[Problems to be solved by the invention]
In the driving behavior automatic evaluation system described in Patent Document 1, it is necessary to store the information of the dangerous place for each dangerous place. As described in Patent Document 1, when the course to be driven by the driver is set to a predetermined course such as a driving school or a facility, the number of dangerous points to be set is limited. The amount of information is also small.

However, when the course to be driven by the driver is not determined or when the driver travels in a wide area, the number of dangerous points increases. Therefore, it is not efficient to set the definition data individually for all the dangerous points in advance, and if the information on the dangerous points is leaked, the evaluation of the accident prevention operation at the dangerous points is accurate. I can't do it. Further, when the accident prevention operation is evaluated under uniform conditions at each dangerous place, there is a problem that the accuracy of the evaluation is lowered.

Japanese Patent No. 5408572

Means for solving problems and their effects

The present invention has been made in view of the above problems, and is a driving evaluation device, an in-vehicle device, and a driving evaluation system provided with these, which can efficiently and accurately evaluate a driver's safety confirmation operation at an intersection. The purpose is to provide a driving evaluation method and a driving evaluation program.

In order to achieve the above object, the driving evaluation device (1) according to the present disclosure uses the detection data including the driver's condition, the position of the vehicle, and the behavior of the vehicle detected by the on-board unit, and the driver. It is an operation evaluation device that evaluates the operation of
A discrimination data acquisition unit that acquires discrimination data using at least one of the detected data detected at the intersection where the driving evaluation is performed, and a discrimination data acquisition unit.
Using the discrimination data, an intersection discrimination unit that discriminates the shape pattern of the intersection for which the operation evaluation is performed, and an intersection discrimination unit.
An evaluation table storage unit that stores an evaluation table in which evaluation conditions for safety confirmation operations to be performed by the driver are set for each intersection shape pattern, and
From the evaluation table, an evaluation condition selection unit that selects the evaluation condition corresponding to the shape pattern of the intersection determined by the intersection determination unit, and an evaluation condition selection unit.
Based on the evaluation conditions selected by the evaluation condition selection unit and the detection data detected at the intersection where the driving evaluation is performed, the safety confirmation evaluation unit that evaluates the safety confirmation operation of the driver at the intersection is provided. It is characterized by having.

According to the operation evaluation device (1), the discrimination data is acquired using at least one of the detection data detected at the intersection where the operation evaluation is performed, and the discrimination data is used to obtain the discrimination data. The shape pattern of the intersection where the driving evaluation is performed is determined. Then, the evaluation conditions corresponding to the determined shape pattern of the intersection are selected from the evaluation table, and based on the selected evaluation conditions and the detection data detected at the intersection where the operation evaluation is performed, the evaluation conditions are selected. The driver's safety confirmation operation at the intersection is evaluated.
Therefore, it is possible to efficiently and accurately evaluate the driver's safety confirmation operation at the intersection under the evaluation conditions corresponding to the shape pattern of the intersection where the driving evaluation is performed.

Further, the operation evaluation device (2) according to the present disclosure uses the detection data detected at the intersection where the operation evaluation is performed in the operation evaluation device (1) to determine the bending direction at the intersection where the operation evaluation is performed. Equipped with a direction estimation unit for estimation
In the evaluation table,
For each combination of the shape pattern of the intersection and the direction of the turn at the intersection, the evaluation conditions for the safety confirmation operation that the driver should perform are set.
The evaluation condition selection unit
The evaluation table is characterized in that the evaluation conditions corresponding to the shape pattern of the intersection determined by the intersection determination unit and the bending direction at the intersection estimated by the direction estimation unit are selected. It is supposed to be.

According to the operation evaluation device (2), the evaluation conditions selected from the evaluation table correspond to the shape pattern of the intersection where the operation evaluation is performed and the direction of the bend at the intersection. The driver's safety confirmation operation can be evaluated more accurately.

Further, the driving evaluation device (3) according to the present disclosure accumulates the detection data detected by the in-vehicle device when the vehicle is passing through the intersection in the driving evaluation device (1) or (2). Detection data storage unit and
A discrimination data storage unit for storing the discrimination data is provided.
The discrimination data acquisition unit
The discrimination data is acquired from the discrimination data storage unit by using at least one of the detection data stored in the detection data storage unit and detected at the intersection where the operation evaluation is performed. It is characterized by that.

According to the operation evaluation device (3), since the detection data storage unit is provided, the detection data can be stored for each on-board unit. Further, the discrimination data can be stored in the discrimination data storage unit. Therefore, even when the number of the in-vehicle devices increases and the amount of processing to be evaluated increases, it is possible to efficiently acquire the appropriate discrimination data for each in-vehicle device, and for each driver at the intersection. The evaluation process of the safety confirmation operation can be efficiently executed.

Further, in the driving evaluation device (4) according to the present disclosure, in the driving evaluation device (3), the discrimination data stored in the discrimination data storage unit includes an intersection map including the intersection through which the vehicle has passed. Contains extractable map data,
The discrimination data acquired by the discrimination data acquisition unit is intersection map data including the intersection through which the vehicle has passed, which is extracted from the map data.
The intersection discrimination unit is characterized in that it discriminates the shape pattern of the intersection for which the operation evaluation is performed based on the intersection map data.

According to the operation evaluation device (4), the intersection map data is acquired from the discrimination data storage unit using the detection data detected at the intersection where the operation evaluation is performed, and the intersection map data is used. , The shape pattern of the intersection where the operation evaluation is performed is determined. Therefore, it is possible to accurately discriminate the shape pattern of the intersection for which the operation evaluation is performed.

Further, the operation evaluation device (5) according to the present disclosure is the above-mentioned operation evaluation device (4).
The intersection discrimination unit
It has a neural network that includes an input layer, an intermediate layer, and an output layer.
The intersection map data is input to the input layer as the discrimination data, and the intersection map data is input.
In the intermediate layer, an operation based on a weighting coefficient is executed on the intersection map data, and the operation is performed.
It is characterized in that the discrimination result of the shape pattern of the intersection is output from the output layer.

According to the operation evaluation device (5), by using the neural network for the intersection discrimination unit, it is possible to efficiently and accurately discriminate the shape pattern of the intersection for which the operation evaluation is performed.

Further, in the driving evaluation device (6) according to the present disclosure, in the driving evaluation device (3), the detection data further includes the outside world image data obtained by capturing or detecting the outside world situation of the vehicle.
The discrimination data stored in the discrimination data storage unit includes the outside world image data transmitted from the vehicle-mounted device.
The discrimination data acquired by the discrimination data acquisition unit is the outside world image data imaged or detected at an intersection through which the vehicle has passed.
The intersection discrimination unit is characterized in that it discriminates the shape pattern of the intersection for which the operation evaluation is performed based on the outside world image data.

According to the operation evaluation device (6), the outside world image data is accumulated as the discrimination data in the discrimination data storage unit, and the shape of the intersection where the operation evaluation is performed is discriminated by using the outside world image data. Will be done. Therefore, the shape pattern of the intersection for which the operation evaluation is performed can be accurately discriminated based on the actual external world situation.

Further, the operation evaluation device (7) according to the present disclosure is the above-mentioned operation evaluation device (6).
The intersection discrimination unit
It has a neural network that includes an input layer, an intermediate layer, and an output layer.
The external image data is input to the input layer as the discrimination data, and the external image data is input to the input layer.
In the intermediate layer, an operation based on a weighting coefficient is executed on the external image data, and the operation is performed.
It is characterized in that the discrimination result of the shape pattern of the intersection is output from the output layer.

According to the operation evaluation device (7), by using the neural network for the intersection discrimination unit, it is possible to efficiently and accurately discriminate the shape pattern of the intersection for which the operation evaluation is performed.

Further, in the operation evaluation device (8) according to the present disclosure, in any of the operation evaluation devices (1) to (7), the timing of the safety confirmation operation to be performed by the driver is set in the evaluation condition of the evaluation table. , Confirmation angle, and confirmation time are included,
The safety confirmation evaluation unit evaluates the timing, the confirmation angle, and the confirmation time at each intersection where the operation evaluation is performed, and calculates the evaluation result scored or ranked based on each evaluation. It is a feature.

According to the operation evaluation device (8), it is possible to comprehensively evaluate the safety confirmation operation at each driver's intersection, and it is possible to output the evaluation result in an easy-to-understand form.

Further, the in-vehicle device (1) according to the present disclosure is an in-vehicle device mounted on a vehicle.
A driver monitoring unit that detects the state of the driver of the vehicle, and
A position detection unit that detects the position of the vehicle and
A behavior detection unit that detects the behavior of the warfare vehicle,
It is characterized by including a driver monitoring unit, a position detection unit, and an output unit that outputs data detected by the behavior detection unit to an operation evaluation device that evaluates the safety confirmation operation of the driver. There is.

According to the in-vehicle device (1), detection data including the state of the driver, the position of the vehicle, and the behavior of the vehicle is output to the driving evaluation device, so that the driver's safety in the driving evaluation device. The evaluation of the confirmation operation can be executed, and the processing load of the in-vehicle device can be reduced.

Further, the driving evaluation system according to the present disclosure is characterized in that it includes any of the driving evaluation devices (1) to (8) and the vehicle-mounted device (1).

According to the operation evaluation system, since any one of the operation evaluation devices (1) to (8) and the in-vehicle device (1) are included, the operation evaluation devices (1) to (8) are included. It is possible to obtain the effect of the above, reduce the processing load of the in-vehicle device, and construct a system that can be introduced at low cost.

Further, the driving evaluation method according to the present disclosure is a driving evaluation method for evaluating the driving of a driver by using detection data including the state of the driver, the position of the vehicle, and the behavior of the vehicle detected by the in-vehicle device. There,
The first step of acquiring discrimination data using at least one of the detected data detected at the intersection where the driving evaluation is performed, and
The second step of discriminating the shape pattern of the intersection for which the operation evaluation is performed using the discriminating data, and
From the evaluation table in which the evaluation conditions for the safety confirmation operation to be performed by the driver are set for each shape pattern of the intersection, the evaluation condition corresponding to the shape pattern of the intersection determined in the second step is selected. 3 steps and
Based on the evaluation conditions selected by the third step and the detection data detected at the intersection where the driving evaluation is performed, the fourth step of evaluating the driver's safety confirmation operation at the intersection. It is characterized by executing the steps including.

According to the operation evaluation method, the discrimination data is acquired using at least one of the detection data detected at the intersection where the operation evaluation is performed, and the operation evaluation is performed using the discrimination data. Determine the shape pattern of the intersection to be performed. Then, the evaluation conditions corresponding to the determined shape pattern of the intersection are selected from the evaluation table, and based on the selected evaluation conditions and the detection data detected at the intersection where the operation evaluation is performed, the evaluation conditions are selected. The driver's safety confirmation operation at the intersection is evaluated.
Therefore, it is possible to efficiently and accurately evaluate the driver's safety confirmation operation at the intersection under the evaluation conditions corresponding to the shape pattern of the intersection where the driving evaluation is performed.

Further, the driving evaluation program according to the present disclosure performs at least one process of evaluating the driving of the driver by using the detection data including the state of the driver, the position of the vehicle, and the behavior of the vehicle detected by the in-vehicle device. It is a driving evaluation program to be executed by one computer.
To the at least one computer
The first step of acquiring discrimination data using at least one of the detected data detected at the intersection where the driving evaluation is performed, and
The second step of discriminating the shape pattern of the intersection for which the operation evaluation is performed using the discriminating data, and
From the evaluation table in which the evaluation conditions for the safety confirmation operation to be performed by the driver are set for each shape pattern of the intersection, the evaluation condition corresponding to the shape pattern of the intersection determined in the second step is selected. 3 steps and
Based on the evaluation conditions selected by the third step and the detection data detected at the intersection where the driving evaluation is performed, the fourth step of evaluating the driver's safety confirmation operation at the intersection. It is characterized by being executed.

According to the operation evaluation program, the discrimination data is acquired using at least one of the detection data detected at the intersection where the operation evaluation is performed, and the operation evaluation is performed using the discrimination data. Determine the shape pattern of the intersection to be performed. Then, the evaluation conditions corresponding to the determined shape pattern of the intersection are selected from the evaluation table, and based on the selected evaluation conditions and the detection data detected at the intersection where the operation evaluation is performed, the evaluation conditions are selected. The driver's safety confirmation operation at the intersection is evaluated.
Therefore, it is possible to realize a driving evaluation device capable of efficiently and accurately evaluating the safety confirmation operation of the driver at the intersection under the evaluation conditions corresponding to the shape pattern of the intersection where the driving evaluation is performed.

It is a schematic block diagram which shows the application example of the operation evaluation system which concerns on embodiment. It is a block diagram which shows the main part of the in-vehicle device used in the driving evaluation system which concerns on embodiment. It is a figure which shows an example of the structure of the detection data transmitted from an in-vehicle device to a server device. It is a block diagram which shows the main part of the server apparatus used in the operation evaluation system which concerns on embodiment. It is a figure which shows an example of the structure of the detection data file stored in the detection data storage part. It is a figure which shows an example of the structure of the evaluation table stored in the evaluation table storage part. It is a figure which shows an example of the structure of the evaluation data file stored in the evaluation data storage part. It is a functional block diagram of the server apparatus which concerns on embodiment. It is a conceptual diagram which shows the example which configured the intersection discriminating part by a neural network. It is a flowchart which shows the processing operation performed by the driver monitoring unit of the vehicle-mounted device which concerns on embodiment. It is a flowchart which shows the processing operation performed by the control unit of the vehicle-mounted device which concerns on embodiment. It is a flowchart which shows the processing operation performed by the central processing unit of the server apparatus which concerns on embodiment. It is a flowchart which shows the processing operation of the safety confirmation evaluation performed by the central processing unit of the server device which concerns on embodiment.

Hereinafter, embodiments of a driving evaluation device, an in-vehicle device, a driving evaluation system including these, a driving evaluation method, and an embodiment of a driving evaluation program according to the present invention will be described with reference to the drawings.

[Application example]
FIG. 1 is a schematic configuration diagram showing an application example of the operation evaluation system according to the embodiment.
The driving evaluation system 1 is a system for evaluating the safety confirmation operation of the driver 3 who is in the vehicle 2, and is an in-vehicle device 10 mounted on at least one or more vehicles 2 and each in-vehicle device 10. It is configured to include at least one or more server devices 40 that process the data acquired in. The server device 40 is an example of the "operation evaluation device" according to the present invention.

The vehicle 2 on which the vehicle-mounted device 10 is mounted is not particularly limited. In this application example, a vehicle managed by a business operator engaged in various businesses may be targeted. For example, a truck managed by a transportation company, a bus managed by a bus company, a taxi managed by a taxi company, a car sharing vehicle managed by a car sharing company, a rental car managed by a car rental company, or owned by a company. It may be a company-owned vehicle that is owned by a company or a company-owned vehicle that is leased and used by a car leasing company.

The vehicle-mounted device 10 and the server device 40 are configured to be communicable via the communication network 4. The communication network 4 may include a wireless communication network such as a mobile phone network (3G / 4G) including a base station or a wireless LAN (Local Area Network), a wired communication network such as a public telephone network, the Internet, or a dedicated network. It may include a telecommunication line such as.

Further, the terminal device 80 of the business operator that manages the vehicle 2 (hereinafter referred to as the business operator terminal) is configured to be able to communicate with the server device 40 via the communication network 4. The business terminal 80 may be a personal computer having a communication function, or may be a mobile information terminal such as a mobile phone, a smartphone, or a tablet device. Further, the operator terminal 80 may be configured to be able to communicate with the vehicle-mounted device 10 via the communication network 4.

The server device 40 includes a communication unit 41, a server computer 50, and a storage unit 60.

In the driving evaluation system 1, the server device 40 accumulates the detection data including the state of the driver 3, the position of the vehicle 2, and the behavior of the vehicle 2 transmitted from each vehicle-mounted device 10, and the accumulated vehicle-mounted vehicle. The operation evaluation of each driver 3 is performed using the detection data of the machine 10. The driving evaluation item includes an item for evaluating the driver's safety confirmation operation at an intersection. The server device 40 uses, for example, the detection data acquired from each on-board unit 10 to execute an evaluation process of a safety confirmation operation at each intersection passed on that day after the day's operation of each vehicle 2 is completed. Alternatively, the evaluation process of the safety confirmation operation at each intersection passed within the fixed period may be executed at regular intervals, and the execution timing of the driving evaluation process is not particularly limited.

By the way, road intersections include not only general crossroads (four-way intersections) but also intersections of various shapes (types) such as three-way intersections, junctions, Y-shaped roads, and circular intersections. When the vehicle 2 turns at an intersection, the behavior that the driver 3 should confirm safety, for example, the direction, range, number of times, or time to confirm, generally differs depending on the shape of the intersection.

When the evaluation process of the safety confirmation operation at the intersection is executed under uniform evaluation conditions regardless of the shape of the intersection, it is difficult to perform accurate evaluation if the evaluation conditions do not match the shape of the intersection. ..

Further, when the evaluation conditions are individually set in advance for each intersection, the number of intersections increases, especially when the setting is performed in a wide area. Therefore, it is not efficient to set evaluation conditions individually in advance for all intersections. Further, when the evaluation conditions are individually set in advance for all the intersections, the memory capacity for storing the data of the evaluation conditions increases, and the evaluation conditions corresponding to the corresponding intersections are executed when the evaluation process is executed. The burden of processing to select is increased. Further, when the evaluation processes of a large number of drivers 3 are executed in parallel, there is a problem that the processing load is further increased.

Therefore, the server device 40 according to the present embodiment discriminates the shape pattern of the intersection for which the operation is evaluated, and the evaluation conditions corresponding to the shape pattern of the discriminated intersection and the detection acquired at the intersection for which the operation is evaluated. Based on the data, a process of evaluating the safety confirmation operation of the driver 3 at each intersection is executed.
For example, the server device 40 uses at least one of the detected data (for example, position data) detected at the intersection where the operation evaluation is performed, and is used for discriminating the shape pattern of the intersection where the operation evaluation is performed. Acquire data (for example, map data of an intersection portion).

After that, the server device 40 discriminates the shape pattern of the intersection where the operation evaluation is performed by using the acquired discriminating data. For example, the shape of an intersection included in the map data is recognized, and the shape pattern of the intersection is determined. Even at the same intersection, for example, at an intersection such as a junction, the shape pattern of the intersection as seen by the driver 3 differs depending on the direction in which the vehicle 2 enters the intersection. The shape pattern of the intersection in the present embodiment means the shape pattern of the intersection seen from the driver 3 of the vehicle 2.

Then, the server device 40 performs the evaluation corresponding to the determined shape pattern of the intersection from the evaluation table in which the evaluation conditions of the safety confirmation operation to be performed by the driver are set for each shape pattern of the intersection such as the crossroads. Select a condition.
Next, the server device 40 executes a process of evaluating the driver's safety confirmation operation at the intersection based on the selected evaluation conditions and the detection data detected at the intersection where the operation evaluation is performed.

Further, the server device 40 uses at least one of the detected data (for example, angular velocity, position data, etc.) detected at the intersection where the operation evaluation is performed, and uses at least one of the detected data (for example, angular velocity, position data, etc.) to determine the direction of the bend at the intersection where the operation evaluation is performed. You may estimate.

In this case, the server device 40 is equipped with an evaluation table in which the evaluation conditions for the safety confirmation operation to be performed by the driver are stored for each combination of the shape pattern of the intersection and the direction of turning at the intersection. The server device 40 selects evaluation conditions corresponding to the shape pattern of the intersection determined from the evaluation table and the direction of bending at the estimated estimated intersection, and the selected evaluation conditions and the operation. Based on the detection data detected at the intersection to be evaluated, a process of evaluating the driver's safety confirmation operation at the intersection may be executed.

With this configuration, the server device 40 does not use uniform evaluation conditions and does not set individual evaluation conditions for each intersection in advance, and the driver 3 at each intersection according to the shape of each intersection. It is possible to efficiently and accurately evaluate the safety confirmation operation.

[Configuration example]
FIG. 2 is a block diagram schematically showing a main part of the vehicle-mounted device 10 used in the operation evaluation system 1 according to the embodiment.
The on-board unit 10 includes a platform unit 11 and a driver monitoring unit 20. Further, the drive recorder unit 30 is connected to the vehicle-mounted device 10.

The platform unit 11 is equipped with an acceleration sensor 12 that measures the acceleration of the vehicle 2, an angular velocity sensor 13 that detects the rotational angular velocity of the vehicle 2, and a GPS (Global Positioning System) receiving unit 14 that detects the position of the vehicle 2. .. Further, the platform unit 11 includes a communication unit 15 that performs communication processing with an external device via the communication network 4, a notification unit 16 that outputs a warning sound such as a beep sound, or a voice such as a warning message, and a storage unit. 17 and an external interface (external I / F) 18 are equipped. Further, the platform unit 11 is equipped with a control unit 19 that controls the processing operation of each unit. The acceleration sensor 12 and the angular velocity sensor 13 are examples of the "behavior detection unit" of the vehicle-mounted device according to the present invention. The GPS receiving unit 14 is an example of the “position detecting unit” of the vehicle-mounted device according to the present invention.

The acceleration sensor 12 is composed of, for example, a three-axis acceleration sensor that measures acceleration in three directions of the XYZ axes. As the 3-axis accelerometer, a semiconductor type accelerometer such as a piezo resistance type can be used in addition to the capacitance type. A 2-axis and 1-axis accelerometer may be used as the accelerometer 12. The acceleration data measured by the acceleration sensor 12 is stored in the RAM 19b of the control unit 19 in association with the detection time.

The angular velocity sensor 13 is a sensor capable of detecting at least an angular velocity corresponding to rotation around the vertical axis (yaw direction), that is, angular velocity data corresponding to rotation (turning) of the vehicle 2 in the left-right direction, for example, a gyro sensor (yaw rate). It is also called a sensor).
The angular velocity sensor 13 includes a uniaxial gyro sensor around the vertical axis, a biaxial gyro sensor that detects an angular velocity around the horizontal axis in the left-right direction (pitch direction), and a horizontal axis around the front-back direction (roll direction). A 3-axis gyro sensor that also detects the angular velocity of the gyro may be used. As these gyro sensors, in addition to the vibration type gyro sensor, an optical type or a mechanical type gyro sensor can be used.

Further, regarding the detection direction of the angular velocity around the vertical axis of the angular velocity sensor 13, for example, clockwise may be set in the positive direction and counterclockwise may be set in the negative direction. In this case, if the vehicle 2 turns to the right, positive angular velocity data is detected, and if the vehicle 2 turns to the left, negative angular velocity data is detected. The angular velocity sensor 13 detects the angular velocity in a predetermined cycle (for example, a cycle of several tens of ms), and the detected angular velocity data is stored in the RAM 19b of the control unit 19 in association with the detection time, for example. For the acceleration sensor 12 and the angular velocity sensor 13, an inertial sensor in which these are mounted in one package may be used.

The GPS receiving unit 14 receives GPS signals from artificial satellites via the antenna 14a at a predetermined cycle, and detects position data (latitude and longitude) of the current location. The position data detected by the GPS receiving unit 14 is stored in the RAM 19b of the control unit 19 in association with the detection time. The device for detecting the position of the vehicle 2 is not limited to the GPS receiving unit 14. For example, positioning devices compatible with other satellite positioning systems such as the Quasi-Zenith Satellite in Japan, GLONASS in Russia, Galileo in Europe, or Compass in China may be used.

The communication unit 15 is an example of the “output unit” of the vehicle-mounted device according to the present invention, and includes a communication module that performs data output processing or the like on the server device 40 via the communication network 4.

The storage unit 17 is composed of, for example, a detachable storage device such as a memory card, or one or more storage devices such as a hard disk drive (HDD) and a solid state drive (SSD). The storage unit 17 stores, for example, data acquired from the acceleration sensor 12, the angular velocity sensor 13, the GPS receiving unit 14, the driver monitoring unit 20, or the drive recorder unit 30.

The external I / F 18 includes an interface circuit, a connector, and the like for exchanging data and signals with an in-vehicle device such as a drive recorder unit 30.

The control unit 19 is composed of a microcomputer including a CPU (Central Processing Unit) 19a, a RAM (Random Access Memory) 19b, and a ROM (Read Only Memory) 19c. The control unit 19 performs a process of storing various acquired data in the RAM 19b or the storage unit 17. Further, the control unit 19 reads out various data stored in the RAM 19b or the storage unit 17 as needed in addition to the program stored in the ROM 19c, and executes the program.

The driver monitoring unit 20 includes a driver camera 21, an image analysis unit 22, and an interface (I / F) 23.
The driver camera 21 includes, for example, a lens unit (not shown), an image sensor unit, a light irradiation unit, and a camera control unit that controls each of these units.

The image pickup device unit includes, for example, an image pickup device such as a CCD (Charge Coupled Device) or a CMOS (Complementary Metal Oxide Semiconductor), a filter, a microlens, and the like. The image pickup element unit includes an infrared sensor such as a CCD, CMOS, or a photodiode capable of receiving infrared rays such as near infrared rays or ultraviolet rays to form an image pickup image, in addition to those capable of forming an image captured image by receiving light in the visible region. It may be included. The light irradiation unit includes a light emitting element such as an LED (Light Emitting Diode), and an infrared LED or the like may be used so that the state of the driver can be imaged day or night. The driver camera 21 may be a monocular camera or a stereo camera.

The camera control unit includes, for example, a processor and the like. The camera control unit controls the image pickup element unit and the light irradiation unit to irradiate light (for example, near infrared rays) from the light irradiation unit, and the image pickup element unit captures the reflected light. I do. The driver camera 21 captures an image at a predetermined frame rate (for example, 30 to 60 frames per second), and the data of the image captured by the driver camera 21 is output to the image analysis unit 22.

The image analysis unit 22 is configured to include, for example, an image processing processor, and information on at least one of the driver's face direction, line-of-sight direction, and eye opening from the image captured by the driver camera 21. Performs processing to detect (driver's condition). Data indicating the driver's state, image data, and imaging date / time data detected by the image analysis unit 22 are transmitted to the platform unit 11 via the interface (I / F) 23, and the RAM 19b or the storage unit of the platform unit 11 It is stored in 17.

The orientation of the driver's face detected by the image analysis unit 22 is, for example, a pitch angle which is an angle (vertical orientation) around the X axis (left and right axis) of the driver's face, and a Y axis of the face ( It may be indicated by the Yaw angle, which is the angle around the vertical axis (horizontal axis), and the Roll angle, which is the angle around the Z axis (front-back axis) of the face (left-right tilt), and is at least left and right. Includes yaw angle to indicate orientation. Further, these angles can be indicated by an angle with respect to a predetermined reference direction, and for example, the reference direction may be set to the front direction of the driver.

Further, the direction of the driver's line of sight detected by the image analysis unit 22 is, for example, the direction of the driver's face detected from the image and information on the eye region (position of the inner corner of the eye, the outer corner of the eye, or the pupil). It is estimated from the relationship with and can be indicated by a line-of-sight vector V (three-dimensional vector) on three-dimensional coordinates. The line-of-sight vector V is, for example, a pitch angle that is an angle (up and down direction) around the X axis (left and right axis) of the driver's face, and a yaw that is an angle (left and right direction) around the Y axis (up and down axis) of the face. It may be estimated from at least one of the angle and the roll angle (tilt left and right) around the Z axis (front-back axis) of the face and the information in the eye area. Further, the line-of-sight vector V shows a part of the value of the three-dimensional vector in common with the value of the vector of the face orientation (for example, the origin of the three-dimensional coordinates is common), or the vector of the face orientation is used as a reference. It may be indicated by the relative angle (relative value of the vector of the orientation of the face).

Further, the driver's eye opening detected by the image analysis unit 22 is, for example, information on the driver's eye region (the inner corner of the eye, the outer corner of the eye, the position of the upper and lower eyelids, the position of the pupil, etc.) detected from the image. ) Is estimated.

The drive recorder unit 30 is connected to the external I / F 18 of the platform unit 11. The drive recorder unit 30 includes an outside camera 31 and an inside camera 32.
The out-of-vehicle camera 31 is a camera that captures an image in front of the vehicle 2, and the in-vehicle camera 32 is a camera that captures an image of the interior of the vehicle 2. The outside camera 31 and the inside camera 32 may be composed of, for example, a visible light camera, but may be composed of a near infrared camera or the like.
The out-of-vehicle camera 31 and the in-vehicle camera 32 each capture an image at a predetermined frame rate (for example, 30 to 60 frames per second), and the images captured by the out-of-vehicle camera 31 and the in-vehicle camera 32 and data such as the imaging date and time are stored in the platform unit. It is sent to 11 and stored in the RAM 19b or the storage unit 17 of the platform unit 11. The drive recorder unit 30 may be configured to include only the external camera 31.

The on-board unit 10 can have a compact configuration in which the platform unit 11 and the driver monitoring unit 20 are housed in one housing. In that case, the location where the vehicle-mounted device 10 is installed in the vehicle is not particularly limited as long as the driver camera 21 can capture at least the field of view including the driver's face. For example, in addition to the vicinity of the center of the dashboard of the vehicle 2, it may be installed in the steering wheel column portion, the vicinity of the meter panel, the position near the rearview mirror, the A pillar portion, or the like. In addition, information including the specifications of the driver camera 21 (for example, the angle of view, the number of pixels (vertical x horizontal), etc.) and the position / posture (for example, the mounting angle, the distance from a predetermined origin (the center position of the handle, etc.), etc.) is the driver. It may be stored in the monitoring unit 20 or the platform unit 11. Further, the driver monitoring unit 20 may be configured separately from the platform unit 11 in addition to the form integrated with the platform unit 11.

FIG. 3 is a diagram showing an example of the structure of detection data transmitted from the vehicle-mounted device 10 to the server device 40.
The detection data includes the identification information (serial number, etc.) of the in-vehicle device 10, the transmission date and time, the direction of the driver's face (pitch, yaw, and roll), the direction of the line of sight (pitch, yaw), and the eye opening (right eye). , And left eye), vehicle acceleration (front-back, left-right, and up-down), angular velocity (yaw), driver image, out-of-vehicle image, vehicle position data (latitude and longitude), and running speed. The structure of the detection data transmitted to the server device 40 is not limited to the structure shown in FIG.
The detection data is transmitted, for example, when the vehicle-mounted device 10 detects that the vehicle 2 has passed the intersection. According to the vehicle-mounted device 10, for example, when the predetermined time elapses after the value of the rotational angular velocity detected by the angular velocity sensor 13 exceeds a predetermined threshold value (threshold value for estimating the approach to the intersection), the vehicle 2 passes through the intersection. It is possible to judge.

FIG. 4 is a block diagram schematically showing a main part of the server device 40 used in the operation evaluation system 1 according to the embodiment.
The server device 40 includes a communication unit 41, a server computer 50, and a storage unit 60, and these are connected via a communication bus 42.

The communication unit 41 is composed of a communication device for realizing transmission / reception of various data and signals to and from the vehicle-mounted device 10 and the business terminal 80 via the communication network 4.

The server computer 50 includes a central processing unit 51 including one or more CPUs, and a main memory 53 in which the control program 52 is stored. The central processing unit 51 executes various processes according to the control program 52 in the main memory 53.

The storage unit 60 is composed of one or more large-capacity storage devices such as a hard disk drive and a solid state drive, and includes a detection data storage unit 61, a discrimination data storage unit 62, an evaluation table storage unit 63, and an evaluation data storage unit. It includes 64 and a driver information storage unit 65.

The detection data storage unit 61 stores the detection data detected by the vehicle-mounted device 10 while the vehicle 2 is passing through the intersection in association with the identification information of each vehicle-mounted device 10.
FIG. 5 is a diagram showing an example of the structure of the detection data file stored in the detection data storage unit 61.
In the detection data file 61a, for example, the date and time when the detection data is transmitted from the in-vehicle device 10 in association with the identification information (serial number, etc.) of the in-vehicle device 10, the direction of the driver's face (pitch, yaw, and). Roll), line-of-sight direction (pitch and yaw), eye opening (right eye and left eye), vehicle acceleration (front and back, left and right, and up and down), angular speed (yaw), driver image (capture date and time, frame number, And image data), outside vehicle image (capture date and time, frame number, and image data), vehicle position data (latitude and longitude), traveling speed, and other information are accumulated.

The discrimination data storage unit 62 stores discrimination data used for discriminating the shape pattern of the intersection through which the vehicle 2 has passed.
The discrimination data is, for example, map data. The map data is, for example, map data capable of extracting a portion of an intersection map including an intersection through which a vehicle has passed. The type of map data is not particularly limited as long as the map data can recognize the shape of the road including the intersection. For example, it may be a plane map image data or a map image data by an aerial photograph. Further, the color map image data may be used, or the map image data in which the road portion and the other portion are binarized may be used. The map image data may be at least data in the range in which the vehicle 2 travels, and of course, map image data over a wide area may be used. Further, the data may be data other than the map image.

The evaluation table storage unit 63 stores an evaluation table in which evaluation conditions for safety confirmation operations to be performed by the driver are set for each combination of the shape pattern of the intersection and the direction of turning at the intersection.

FIG. 6 is a diagram showing an example of the structure of the evaluation table stored in the evaluation table storage unit 63. The items of the evaluation table 63a include a pattern number, an intersection shape pattern, a turning direction, and a confirmation action to be performed by the driver. The evaluation table shown in FIG. 6 is an example of a case where a car is traveling on the left side, such as in Japan. In a country where a car is on the left side and a country where a car is on the right side, the right turn and the left turn are reversed. In addition, the contents of the evaluation table are appropriately set according to the traffic regulations of the country.

In the item of the intersection shape pattern, the shape pattern of the intersection categorized in advance is registered. For example, a crossroads of a living road, a left junction of a living road, a left junction of a living road, a crossroads of a main road, a left junction of a main road, and a right junction of a main road are stored. The left junction is an intersection where the vehicle can turn left and go straight in the direction of travel. The right-hand junction is an intersection where you can turn right and go straight in the direction of travel of the vehicle. Further, it may be further categorized by items such as the presence / absence of a signal at an intersection or the presence / absence of a pedestrian crossing.

In the item of the turning direction, the direction of turning at the intersection of each intersection shape pattern, for example, a left turn or a right turn is stored.
The items of confirmation action include items such as the direction (right or left), the angle, and time that the driver should confirm safety when passing through the intersection of each intersection shape pattern (before or during approach). More than one is remembered. The angle indicates, for example, the direction of the driver's face or the direction of the line of sight with respect to the front direction of the vehicle.

For example, No. 1 in the evaluation table 63a. In the case of 1 (the pattern of turning left at the crossroads of the living road), an item for evaluating whether the left confirmation before approaching is performed at a degree or more for t1 second or more is set as the confirmation action 1. This evaluation is an evaluation of whether or not pedestrians, motorcycles, etc. have been confirmed to be involved before entering the intersection.
In the confirmation action 2, an item for evaluating whether the right confirmation before approaching is performed at a degree or more for t1 second or more is set. This evaluation is an evaluation of whether or not it is confirmed whether or not there is an oncoming vehicle turning right at the intersection, or whether or not there is a pedestrian or the like who is trying to cross the intersection from the front of the vehicle.
In the confirmation action 3, an item for evaluating whether the left confirmation during the approach is performed at b degrees or more for t2 seconds or more is set. This evaluation is an evaluation of whether or not there are pedestrians or the like trying to cross the intersection, and whether or not the safety of the front (course switching direction) after turning left has been confirmed.

The evaluation data storage unit 64 stores the evaluation result of the driver's safety confirmation operation executed by the server computer 50, that is, information on the evaluation result of the driver 3's safety confirmation operation at least at the intersection for each in-vehicle device 10. Will be done.
FIG. 7 is a diagram showing an example of the structure of the evaluation data file stored in the evaluation data storage unit 64. The evaluation data file 64a includes identification information of the in-vehicle device 10, the date and time of passing the intersection, the position of the intersection, the evaluation conditions (the number of the intersection shape pattern selected from the evaluation table), and the evaluation result. In the evaluation result, for example, evaluation data scored or ranked based on the evaluation under the evaluation conditions selected from the evaluation table 63a is stored.
The driver information storage unit 65 stores various information about the driver 3 managed by the business operator.

FIG. 8 is a functional configuration block diagram of the server device according to the embodiment.
The central processing unit 51 of the server computer 50 includes a discrimination data acquisition unit 54, a direction estimation unit 55, an intersection discrimination unit 56, an evaluation condition selection unit 57, and a safety confirmation evaluation unit 58.

The discrimination data acquisition unit 54 reads out the detection data detected at the intersection where the operation is evaluated from the detection data storage unit 61, and uses the detection data to perform the operation evaluation from the discrimination data storage unit 62. The shape pattern of the intersection. Performs the process of acquiring the discrimination data used for the discrimination of.
For example, the discrimination data acquisition unit 54 reads out the position data (longitudinal and latitude) detected at the intersection where the operation is evaluated from the detection data storage unit 61, and the discrimination data storage unit 62 is based on the position data. Performs the process of acquiring the map data (intersection map data) of the intersection portion for which driving evaluation is performed from the accumulated map data (wide area map data).

The direction estimation unit 55 reads the detection data detected at the intersection where the driving evaluation is performed from the detection data storage unit 61, and executes a process of estimating the bending direction at the intersection where the driving evaluation is performed using the detection data. do.
For example, the direction estimation unit 55 reads out the angular velocity data detected when passing through the intersection for driving evaluation from the detection data storage unit 61, and estimates the right / left turn direction of the vehicle 2 based on the positive / negative of the value of the angular velocity data. You may. For example, if the value of the angular velocity data is positive, it can be estimated to be a right turn, and if the integrated value is negative, it can be estimated to be a left turn. Further, the integrated value of the angular velocity data may be calculated, and the right / left turn direction of the vehicle 2 may be estimated based on the positive / negative of the calculated integrated value.
Alternatively, the direction estimation unit 55 reads out the position data (latitude and longitude) detected when passing through the intersection for driving evaluation from the detection data storage unit 61, and based on the time-series change of the position data, the vehicle 2 You may estimate the right / left turn direction of.

The intersection discriminating unit 56 executes a process of discriminating the shape pattern of the intersection to be evaluated for operation by using the discriminating data acquired by the discriminating data acquisition unit 54.
For example, the intersection discrimination unit 56 performs image recognition processing of the intersection shape of the road using the intersection map data acquired by the discrimination data acquisition unit 54, and performs processing to discriminate the shape pattern of the intersection for which driving evaluation is performed. .. For the image recognition process, for example, the intersection shape recognition process by template matching may be performed. In this case, a plurality of types of intersection shape templates may be stored in advance, and the shape pattern of the intersection to be evaluated for driving may be determined by pattern matching processing between the intersection map data and the template.

Alternatively, a neural network may be used for the intersection discrimination unit 56 to perform image recognition processing of the shape pattern of the intersection.
FIG. 9 is a conceptual diagram showing an example in which the intersection discrimination unit 56 is configured by a neural network.
The neural network 56a is divided into a plurality of layers including an input layer 56b, an intermediate layer (also referred to as a hidden layer) 56c, and an output layer 56d, and signals are processed by a plurality of neurons (also referred to as units) and classified from the output layer 56d. The result is output.
The neural network 56a functions as a trained classifier created by performing learning processing in advance using map images of various intersection shapes as teacher data, for example.

For example, the pixel information of the intersection map data 62a acquired from the discrimination data storage unit 62 by the discrimination data acquisition unit 54 is input to the input layer 56b, and the intermediate layer 56c has already learned the intersection map data 62a. The calculation based on the weighting coefficient is executed, and the output layer 56d is configured to output the discrimination result of the shape pattern of the intersection.

The input layer 56b is a layer that captures information given to the neural network 56a. For example, the number of units corresponding to the number of pixels of the map image is provided, and each pixel information of the map image is input to each neuron.

Each neuron constituting the intermediate layer 56c is a value obtained by adding a weight coefficient to a plurality of input values, adding the weight coefficient, and subtracting a threshold value, and processing the value by a transfer function (for example, a step function or a sigmoid function). Is performed, and the characteristics of the intersection shape of the map image input to the input layer 56b are extracted. For example, in the shallow layer of the intermediate layer 56c, small features of the intersection shape in the map image (characteristics of the boundary portion of the road, etc.) are recognized, and as the layer becomes deeper (toward the output side), the smaller features are combined. Major features of intersections (such as features of road intersections) are recognized.

Each neuron constituting the output layer 56d outputs the result calculated by the neural network 56a. For example, it is composed of neurons having the number of intersection shape patterns set in the evaluation table 63a, and outputs the result of classifying (identifying) which intersection shape pattern belongs to.

The evaluation condition selection unit 57 reads the evaluation table 63a from the evaluation table storage unit 63, and bends at the intersection shape pattern determined by the intersection determination unit 56 and the intersection estimated by the direction estimation unit 55 from the evaluation table 63a. The process of selecting the evaluation condition corresponding to the direction is executed.

The safety confirmation evaluation unit 58 executes a process of evaluating the driver's safety confirmation operation at the intersection based on the evaluation conditions selected by the evaluation condition selection unit 57 and the detection data detected at the intersection where the driving evaluation is performed. Then, the evaluation result is stored in the evaluation data storage unit 64.

The safety confirmation evaluation unit 58 first performs a process of estimating the approach time of the vehicle 2 to the intersection. For example, the safety confirmation evaluation unit 58 calculates the integrated value of the angular velocity data (angular velocity data transmitted from the in-vehicle device 10 when passing through the intersection) stored in the detection data storage unit 61, and the calculated integrated value is predetermined. The time when the integration ratio is reached is detected, and this detection time is estimated as the approach time to the intersection.

Then, the safety confirmation evaluation unit 58 performs a process of evaluating the driver's safety confirmation operation within a predetermined time before and after the approach time to the intersection based on the evaluation conditions selected from the evaluation table 63a.
First, based on the detection data stored in the detection data storage unit 61, the safety confirmation evaluation unit 58 has at least one swing angle and a swing angle of at least one of the driver's face direction and the line-of-sight direction within a predetermined time before and after the approach time. Detect the swing time.

Next, the safety confirmation evaluation unit 58 determines the safety before entering the intersection based on the evaluation conditions selected from the evaluation table 63a and the swing angle and swing time at the predetermined time (before the approach) before the approach time. Performs the process of evaluating the confirmation operation. When the selected evaluation conditions include confirmation actions in each of the left and right directions, the detection order in each of the left and right directions does not matter.

For example, regarding safety confirmation in the right direction, (1) whether or not the confirmation timing in the right direction is appropriate, (2) whether or not the swing angle in the right direction is equal to or greater than a predetermined angle, and (3) the predetermined value. It is determined whether or not the state of the angle or more continues for a predetermined time or more. (1) Whether or not the confirmation timing in the right direction is appropriate can be determined by, for example, whether or not the swing angle in the right direction is equal to or greater than the predetermined angle by a predetermined time before the approach time. ..

Similarly, regarding safety confirmation in the left direction, (4) whether or not the confirmation timing in the left direction is appropriate, (5) whether or not the swing angle in the left direction is equal to or greater than a predetermined angle, and (6) the above. It is determined whether or not the state of the predetermined angle or more continues for the predetermined time or more. (4) Whether or not the confirmation timing in the left direction is appropriate can be determined by, for example, whether or not the swing angle in the left direction is equal to or greater than the predetermined angle by a predetermined time before the approach time. ..

Further, the safety confirmation evaluation unit 58 determines the course change direction at the intersection based on the evaluation conditions selected from the evaluation table 63a and the swing angle and swing time at a predetermined time (during approach) after the approach time. For example, a process for evaluating the safety confirmation operation at the right turn destination or the left turn destination is performed.
For example, regarding the safety confirmation of the course change direction, (7) whether or not the confirmation timing of the course change direction is appropriate, and (8) whether or not the swing angle in the course change direction of turning right or left is equal to or more than a predetermined angle. Or (9) it is determined whether or not the state of the predetermined angle or more continues for a predetermined time or more. Whether or not the confirmation timing of the above (7) course change direction is appropriate is determined by, for example, whether or not the swing angle in the course change direction is equal to or greater than the predetermined angle by a predetermined time after the approach time. be able to.

Further, the safety confirmation evaluation unit 58 may perform a process of evaluating the driver's deceleration consciousness based on the traveling speed of the vehicle 2 at a predetermined time before and after the approach time. For example, (10) deceleration by determining whether the maximum value of the traveling speed of the vehicle 2 at a predetermined time before and after the approach time is equal to or less than a predetermined speed (for example, an upper limit speed at which a vehicle can safely turn and pass through an intersection). You may evaluate the presence or absence of consciousness.

The safety confirmation evaluation unit 58 evaluates each confirmation action set in the evaluation conditions selected from the evaluation table 63a based on the determination methods of the above items (1) to (10), and evaluates the evaluation points for each intersection. Perform the calculation process. For example, for each confirmation action set in the evaluation conditions, the evaluation results of confirmation timing, confirmation angle, and confirmation time are totaled, and necessary statistical processing (averaging, normalization processing, etc.) is performed for each intersection. Calculate the score by scoring or ranking.
Then, the safety confirmation evaluation unit 58 performs a process of storing the calculated evaluation points for each intersection in the evaluation data storage unit 64 of the storage unit 60.

Further, the server computer 50 may process various requests requested from the browser of the business terminal 80 and present the processing results corresponding to the various requests to the business terminal 80 through the browser.
For example, when the server computer 50 receives a transmission request for a driver's driving evaluation result from the operator terminal 80, a predetermined type of driving evaluation is performed based on the driver's evaluation data stored in the evaluation data storage unit 64. You may create a report and execute a process of presenting a driving evaluation report through the browser of the business operator terminal 80. These processes may be executed on a server computer other than the server computer 50.

[Operation example]
FIG. 10 is a flowchart showing a processing operation performed by the driver monitoring unit 20 in the vehicle-mounted device 10 according to the embodiment. This processing operation is executed, for example, at the timing when an image is captured by the driver camera 21 (for example, every frame or every frame at a predetermined interval).

First, the image analysis unit 22 acquires an image captured by the driver camera 21 (step S1).

Next, the image analysis unit 22 performs a process of detecting the driver's face or a face area from the acquired image (step S2). The method for detecting a face from an image is not particularly limited, but it is preferable to adopt a method for detecting a face at high speed and with high accuracy.

The image analysis unit 22 performs a process of detecting the position and shape of facial organs such as eyes, nose, mouth, and eyebrows from the facial region detected in step S2 (step S3). The method for detecting the facial organs from the facial region in the image is not particularly limited, but it is preferable to adopt a method that can detect the facial organs at high speed and with high accuracy. For example, a method in which the image analysis unit 22 creates a three-dimensional face shape model, fits the three-dimensional face shape model to the face region on the two-dimensional image, and detects the position and shape of each organ of the face can be adopted. According to this method, it is possible to accurately detect the position and shape of each organ of the face regardless of the installation position of the driver camera 21 or the orientation of the face in the image. As a technique for fitting a three-dimensional face shape model to a human face in an image, for example, the technique described in JP-A-2007-249280 can be applied, but the technique is not limited thereto.

Next, the image analysis unit 22 detects the orientation of the driver's face based on the position and shape data of each organ of the face obtained in step S3 (step S4). For example, the pitch angle of vertical rotation (around the X axis), the yaw angle of the horizontal rotation (around the Y axis), and the roll angle of the total rotation (around the Z axis) included in the parameters of the above three-dimensional face shape model are set. It may be detected as information regarding the orientation of the driver's face.

Next, the image analysis unit 22 determines the orientation of the driver's face determined in step S4, and the position and shape of the driver's facial organs determined in step S3, particularly the characteristic points of the eyes (outer corners of the eyes, inner corners of the eyes, and pupils). The direction of the line of sight is detected based on the position and shape (step S5).
The direction of the line of sight is, for example, the feature amount of the image of the eye in various face directions and the direction of the line of sight (relative position of the outer corner of the eye, the inner corner of the eye, the pupil, or the relative position of the sclera (so-called white eye) part and the iris (so-called black eye) part. , Shading, texture, etc.) may be detected by learning in advance using a learning device and evaluating the degree of similarity with the learned feature amount data. Alternatively, using the fitting result of the three-dimensional face shape model, the size and center position of the eyeball are estimated from the size and orientation of the face and the position of the eyes, and the position of the pupil is detected to determine the center of the eyeball. The vector connecting the center of the pupil may be detected as the line-of-sight direction.

Next, the image analysis unit 22 determines the eye opening degree based on the position and shape of the driver's facial organs obtained in step S3, particularly the positions and shapes of the characteristic points of the eyes (outer corners of the eyes, inner corners of the eyes, pupils, and eyelids). Detect (step S6).

Next, the image analysis unit 22 determines the direction of the driver's face detected in step S4, the direction of the driver's line of sight detected in step S5, the opening of the driver's eyes detected in step S6, and the driver's eyes. The image and the imaging time are associated with each other and transmitted as image information to the platform unit 11 (step S7), after which the image analysis unit 22 returns to step S1 and repeats the process.

FIG. 11 is a flowchart showing a processing operation performed by the control unit 19 in the vehicle-mounted device 10 according to the embodiment. This processing operation is executed, for example, in a predetermined cycle of several tens of ms to several seconds.

The control unit 19 acquires the acceleration data measured by the acceleration sensor 12 and stores it in the RAM 19b (step S11).

Further, the control unit 19 acquires the angular velocity data detected by the angular velocity sensor 13 and stores it in the RAM 19b (step S12).

Further, the control unit 19 acquires the image information transmitted from the driver monitoring unit 20 and stores it in the RAM 19b (step S13).

Further, the control unit 19 acquires the data (data of the image outside the vehicle and the image inside the vehicle) transmitted from the drive recorder unit 30 and stores it in the RAM 19b (step S14).

Further, the control unit 19 acquires the position data detected by the GPS receiving unit 14 (step S15).

Next, the control unit 19 calculates the traveling speed based on the unit-time change of the position data, and stores the position data and the traveling speed in the RAM 19b (step S16).

Next, the control unit 19 determines whether or not the vehicle 2 has entered the intersection (step S17). For example, the control unit 19 determines whether or not the integrated value of the angular velocity is equal to or higher than the threshold value indicating the approach of the vehicle to the intersection and the traveling speed is equal to or lower than the predetermined upper limit speed indicating the approach to the intersection.

In step S17, if the control unit 19 determines that the vehicle 2 has entered the intersection, then the control unit 19 determines whether or not the vehicle 2 has passed the intersection (step S18). For example, it is determined whether or not the integrated value of the angular velocity is equal to or less than the threshold value indicating that the vehicle has passed the intersection and the traveling speed is equal to or more than the predetermined lower limit speed indicating the passage of the intersection.

If the control unit 19 determines that the vehicle 2 has passed the intersection, the control unit 19 reads the detection data acquired at a predetermined time (about several tens of seconds) before and after the vehicle 2 enters the intersection from the RAM 19b, and reads the detection data from the RAM 19b. 15 is controlled to execute a process of transmitting detection data to the server device 40 (step S19). After that, the control unit 19 returns to step S11 and repeats the process. An example of the structure of the detection data transmitted from the vehicle-mounted device 10 to the server device 40 is shown in FIG.

The control unit 19 may store the detection data read from the RAM 19b in the storage unit 17 (for example, a removable storage medium) instead of the processing in step S19. Then, after the end of the day's driving, the driver 3 removes the storage unit 17 from the vehicle-mounted device 10, causes the business operator terminal 80 to read the detection data stored in the storage unit 17, and the business operator terminal 80 uses the server device 40. You may send it to.

FIG. 12 is a flowchart showing a processing operation performed by the central processing unit 51 in the server device 40 according to the embodiment.
The central processing unit 51 determines whether or not the detection data transmitted from the vehicle-mounted device 10 has been received (step S21).

If the central processing unit 51 determines that the detection data has been received from the vehicle-mounted device 10, then the detection data received from the vehicle-mounted device 10 is stored in the detection data storage unit 61 in association with the identification information of the vehicle-mounted device 10. (Step S22).

After that, the central processing unit 51 determines whether or not it is the timing to execute the evaluation of the driver's safe driving behavior (step S23). Whether or not it is the timing to execute the evaluation may be determined, for example, by whether or not a signal indicating the end of operation of the vehicle 2 is received from the vehicle-mounted device 10.

If the central processing unit 51 determines that it is not the timing to execute the evaluation, the process returns to step S21.

On the other hand, if the central processing unit 51 determines that it is time to execute the evaluation, the central processing unit 51 performs an evaluation process of the safety confirmation operation at the intersection of the driver 3 of the vehicle 2 on which the vehicle-mounted device 10 is mounted (step S24). The evaluation process executed in step S24 will be described with reference to the flowchart of FIG.

After that, the central processing unit 51 performs a process of storing the result of the evaluation process in the evaluation data storage unit 64 in association with the information of the vehicle-mounted device 10 or the driver 3 (step S25), and ends the process.

FIG. 13 is an example of a flowchart showing an evaluation processing operation of the safety confirmation operation performed by the central processing unit 51 in the server device 40 according to the embodiment. It should be noted that this flowchart shows a processing operation for evaluating the safety confirmation operation of the driver 3 at each intersection that has passed by the end of the day's driving of the vehicle 2 on which the on-board unit 10 is mounted.

The central processing unit 51 reads out the detection data of the vehicle-mounted device 10 from the detection data storage unit 61 (step S31).

Next, the central processing unit 51 counts the number of read detection data, that is, the number N of the passed intersections in chronological order (step S32).

Next, the central processing unit 51 determines whether or not the number N of the passed intersections is 1 or more (step S33).

When the central processing unit 51 determines in step S33 that the number N of intersections is not 1 or more, the central processing unit 51 stores the number N = 0 of intersections in the evaluation data storage unit 64 (step S34), and ends the process.

On the other hand, if the central processing unit 51 determines in step S33 that the number N of intersections is 1 or more, 1 is added to the counter C that counts the number of evaluated intersections in chronological order (step S35).

Next, the central processing unit 51 acquires the discrimination data of the shape pattern of the intersection to be evaluated from the discrimination data storage unit 62 by using the detection data at the intersection that has passed the Cth position (step S36).
For example, the central processing device 51 has a map data (for example, a wide area map) stored in the discrimination data storage unit 62 based on the position data (longitude and latitude) at the time of passing the intersection included in the C-th detection data. Performs the process of acquiring the map data (intersection map data) of the intersection portion for which driving evaluation is performed from the data).

Next, the central processing unit 51 executes a process of estimating the bending direction at the intersection using the read detection data (step S37).
For example, the central processing unit 51 estimates the right / left turn direction of the vehicle 2 based on whether the angular velocity data included in the detection data is positive or negative. For example, if the angular velocity data is positive, it is estimated to be a right turn, and if it is negative, it is estimated to be a left turn. Further, the integrated value of the angular velocity may be calculated, and the right / left turn direction of the vehicle 2 may be estimated based on the positive / negative of the calculated integrated value, or the time series of the position data (latitude and longitude) included in the detection data. The right / left turn direction of the vehicle 2 may be estimated based on the change.

Next, the central processing unit 51 inputs the discrimination data acquired in step S37 to the intersection discrimination unit 56, and executes a process of discriminating the shape pattern of the intersection (step S38).
For example, using the above-mentioned neural network 56a, image recognition processing of the shape pattern of an intersection is performed. The intersection map data 62a acquired as discrimination data is input to the input layer 56b, the intermediate layer 56c executes an operation based on the weighting coefficient for the intersection map data 62a, and the output layer 56d determines the shape pattern of the intersection. Execute the process to output the discrimination result. Alternatively, the central processing unit 51 uses the intersection map data acquired as discrimination data to discriminate the intersection shape of the road by image recognition processing such as template matching, and outputs the determination result of the shape pattern of the intersection. May be executed.

Next, the central processing unit 51 acquires the determination result of the shape pattern of the intersection determined by the intersection determination unit 56 (step S39).

After that, the central processing unit 51 reads the evaluation table 63a from the evaluation table storage unit 63 (step S40).

Next, the central processing unit 51 performs a process of selecting from the evaluation table 63a the evaluation conditions corresponding to the shape pattern of the intersection acquired in step S39 and the direction of bending at the intersection estimated in step S37. (Step S41).

Next, the central processing unit 51 performs a process of evaluating the driver's safety confirmation operation at the intersection passed through the Cth position based on the evaluation conditions selected in step S41 and the Cth detection data (step S42). ).

For example, the central processing unit 51 confirms the confirmation timing, the confirmation angle, and the confirmation for each confirmation action set in the evaluation conditions selected from the evaluation table, based on the determination methods of the above items (1) to (10). Evaluation such as time is performed, the evaluation result is statistically processed, and the evaluation score based on the score or ranking at the intersection is calculated. Next, the central processing unit 51 performs a process of storing the evaluation result in the evaluation data storage unit 64 (step S43).

Next, the central processing unit 51 determines whether or not the counter C has reached the number N of intersections (step S44), and if it is determined that the counter C has not reached the number N of intersections, the central processing unit 51 determines. The process returns to step S35 and the process is repeated. On the other hand, if the central processing unit 51 determines that the counter C has reached the number N of intersections, the central processing unit 51 ends the processing thereafter.

According to the server device 40 according to the embodiment, discrimination data is acquired from the discrimination data storage unit 62 using the position data of the detection data detected at the intersection where the operation is evaluated, and the discrimination data is used. Therefore, the shape pattern of the intersection where the operation evaluation is performed is determined. Then, the evaluation conditions corresponding to the determined shape pattern of the intersection are selected from the evaluation table 63a, and the driver at the intersection is based on the selected evaluation conditions and the detection data detected at the intersection where the driving evaluation is performed. The safety confirmation operation is evaluated.
Therefore, it is possible to efficiently and accurately evaluate the driver's safety confirmation operation at the intersection under the evaluation conditions corresponding to the shape pattern of the intersection where the driving evaluation is performed.

Further, in the evaluation table 63a of the evaluation table storage unit 63, evaluation conditions are set for each combination of the intersection shape pattern and the direction of bending at the intersection, so that it is not necessary to set the evaluation conditions for each intersection. In the server device 40, the memory capacity for storing the evaluation conditions can be reduced, and the evaluation conditions can be efficiently selected.

Further, according to the server device 40, since the detection data storage unit 61 is provided, the detection data can be stored for each in-vehicle device 10. Further, the discrimination data can be stored in the discrimination data storage unit 62. Therefore, even when the number of on-board units 10 increases and the amount of processing to be evaluated increases, appropriate discrimination data can be efficiently acquired for each on-board unit 10, and safety for each driver at an intersection can be obtained. The evaluation process of the confirmation operation can be executed efficiently.

Further, according to the server device 40, the intersection map data is acquired from the discrimination data storage unit 62 using the position data detected at the intersection where the operation evaluation is performed, and the operation evaluation is performed using the intersection map data. The shape pattern of the intersection is determined. Therefore, it is possible to efficiently and accurately discriminate the shape pattern of the intersection for which the operation is evaluated.

Further, according to the server device 40, by using the neural network 56a for the intersection discrimination unit 56, it is possible to efficiently and accurately discriminate the shape pattern of the intersection for which the operation is evaluated.

Further, according to the server device 40, the safety confirmation evaluation unit 58 evaluates the confirmation timing, the confirmation angle, and the confirmation time at each intersection for which the operation evaluation is performed, and scores or ranks the evaluation based on each evaluation. Since the result is calculated, it is possible to comprehensively evaluate the safety confirmation operation at each driver's intersection, and it is possible to output the evaluation result to the operator terminal 80 in an easy-to-understand form.

Further, according to the server device 40, it is possible to accurately evaluate the left and right safety confirmation operation of the driver before entering the intersection according to the shape of the intersection. In addition, it is possible to accurately evaluate the safety confirmation operation for the driver's course change direction after entering the intersection. Further, since the driver's deceleration consciousness is evaluated based on the speed of the vehicle 2 at a predetermined time before and after the approach time, the driver's safety confirmation consciousness at the intersection can be accurately evaluated. Further, by calculating the evaluation points of the safety confirmation operation for each intersection based on these evaluation results, the safety confirmation operation at the entire intersection passed by the driver can be evaluated to a higher degree.

Further, according to the on-board unit 10, detection data including the state of the driver 3, the position of the vehicle 2, and the behavior of the vehicle 2 is output to the server device 40, so that the server device 40 evaluates the safety confirmation operation of the driver. Can be executed, and the processing load of the in-vehicle device 10 can be reduced. In addition, the configuration can be simplified and retrofitting to the vehicle 2 can be easily performed.

Further, according to the operation evaluation system 1, since the server device 40 and the in-vehicle device 10 are included, the effect of the server device 40 can be obtained, and the processing load of the in-vehicle device 10 is reduced. It is possible to build a system that can be introduced at low cost.

Although the embodiments of the present invention have been described in detail above, the above description is merely an example of the present invention in all respects. Needless to say, various improvements and changes can be made without departing from the scope of the present invention.

[Modification example]
In the server device 40 according to the above embodiment, the case where the map data is stored as the discrimination data in the discrimination data storage unit 62 has been described. In another embodiment, the vehicle exterior image data (outside world image data) captured by the vehicle exterior camera 31 of the vehicle-mounted device 10 may be configured to be stored in the discrimination data storage unit 62 as discrimination data.
In this case, the discrimination data acquired by the discrimination data acquisition unit 54 is the vehicle exterior image data captured or detected at the intersection through which the vehicle has passed, and the intersection discrimination unit 56 performs driving evaluation based on the vehicle exterior image data. The process of determining the shape pattern of the intersection to be performed is executed.

Further, in this case, the intersection discrimination unit 56 may be configured by the neural network 56a. For example, the vehicle exterior image data is input as the discrimination data in the input layer 56b, and the intermediate layer 56c weights the vehicle exterior image data. The calculation based on the coefficient may be executed, and the determination result of the shape pattern of the intersection may be output from the output layer 56d.

[Additional Notes]
Embodiments of the present invention may also be described as, but are not limited to, the following appendices.
(Appendix 1)
A driving evaluation device (40) that evaluates a driver's driving using detection data including the driver's condition, the position of the vehicle, and the behavior of the vehicle detected by the on-board unit (10).
A discrimination data acquisition unit (54) that acquires discrimination data using at least one of the detected data detected at the intersection where the driving evaluation is performed, and a discrimination data acquisition unit (54).
Using the discrimination data, an intersection discrimination unit (56) for discriminating the shape pattern of the intersection for which the operation evaluation is performed, and an intersection discrimination unit (56).
An evaluation table storage unit (63) that stores an evaluation table (63a) in which evaluation conditions for safety confirmation operations to be performed by the driver are set for each shape pattern of an intersection, and
An evaluation condition selection unit (57) that selects the evaluation condition corresponding to the shape pattern of the intersection determined by the intersection determination unit (56) from the evaluation table (63a).
A safety confirmation evaluation that evaluates the safety confirmation operation of the driver at the intersection based on the evaluation conditions selected by the evaluation condition selection unit (57) and the detection data detected at the intersection where the operation evaluation is performed. An operation evaluation device including a unit (58).

(Appendix 2)
An on-board unit (10) mounted on a vehicle.
The driver monitoring unit (20) that detects the state of the driver of the vehicle, and
The position detection unit (14) that detects the position of the vehicle and
The behavior detection unit (12, 13) that detects the behavior of the warfare vehicle, and
The operation evaluation device (40) that evaluates the safety confirmation operation of the driver by using the data detected by the driver monitoring unit (20), the position detection unit (14), and the behavior detection unit (12, 13). An in-vehicle device characterized by having an output unit (15) for outputting to.

(Appendix 3)
A driving evaluation system (1) characterized by including a driving evaluation device (40) and an in-vehicle device (10).

(Appendix 4)
A driving evaluation method for evaluating a driver's driving using detection data including the driver's condition, the position of the vehicle, and the behavior of the vehicle detected by the on-board unit (10).
The first step (S36) of acquiring discrimination data using at least one of the detected data detected at the intersection where the driving evaluation is performed, and
The second step (S39) of discriminating the shape pattern of the intersection for which the operation evaluation is performed using the discriminating data, and
From the evaluation table in which the evaluation conditions for the safety confirmation operation to be performed by the driver are set for each shape pattern of the intersection, the evaluation condition corresponding to the shape pattern of the intersection determined in the second step is selected. Step 3 (S41) and
A fourth evaluation of the driver's safety confirmation operation at the intersection based on the evaluation conditions selected in the third step (S41) and the detection data detected at the intersection where the driving evaluation is performed. An operation evaluation method comprising executing a step including a step (S42).

(Appendix 5)
To have at least one computer perform a process of evaluating the driver's driving using the detection data including the driver's condition, the position of the vehicle, and the behavior of the vehicle detected by the on-board unit (10). There is (53) in the driving evaluation program,
To the at least one computer (51)
The first step (S36) of acquiring discrimination data using at least one of the detected data detected at the intersection where the driving evaluation is performed, and
The second step (S39) of discriminating the shape pattern of the intersection for which the operation evaluation is performed using the discriminating data, and
From the evaluation table in which the evaluation conditions for the safety confirmation operation to be performed by the driver are set for each shape pattern of the intersection, the evaluation condition corresponding to the shape pattern of the intersection determined in the second step is selected. Step 3 (S41) and
A fourth evaluation of the driver's safety confirmation operation at the intersection based on the evaluation conditions selected in the third step (S41) and the detection data detected at the intersection where the driving evaluation is performed. A driving evaluation program characterized by executing a step (S42).

1 Driving evaluation system 2 Vehicle 3 Driver 4 Communication network 10 In-vehicle device 11 Platform unit 12 Accelerometer 13 Angular velocity sensor 14 GPS receiver 15 Communication unit 16 Notification unit 17 Storage unit 18 External interface (external I / F)
19 Control unit 19a CPU
19b RAM
19c ROM
20 Driver monitoring unit 21 Driver camera 22 Image analysis unit 23 Interface (I / F)
30 Drive recorder section 31 Outside camera 32 Inside camera 40 Server device (driving evaluation device)
41 Communication unit 50 Server computer 51 Central processing device 52 Main memory 53 Control program 54 Discrimination data acquisition unit 55 Direction estimation unit 56 Crossing point discrimination unit 57 Evaluation condition selection unit 58 Safety confirmation evaluation unit 60 Storage unit 61 Detection data storage unit 61a Detection Data file 62 Discrimination data storage unit 63 Evaluation table storage unit 63a Evaluation table 64 Evaluation data storage unit 64a Evaluation data file 65 Driver information storage unit 80 Business terminal

Claims (9)

A driving evaluation device that evaluates a driver's driving using detection data including the driver's condition, the position of the vehicle, and the behavior of the vehicle detected by the on-board unit.
A discrimination data acquisition unit that acquires discrimination data using at least one of the detected data detected at the intersection where the driving evaluation is performed, and a discrimination data acquisition unit.
Using the discrimination data, an intersection discrimination unit that discriminates the shape pattern of the intersection for which the operation evaluation is performed, and an intersection discrimination unit.
An evaluation table storage unit that stores an evaluation table in which evaluation conditions for safety confirmation operations to be performed by the driver are set for each intersection shape pattern, and
From the evaluation table, an evaluation condition selection unit that selects the evaluation condition corresponding to the shape pattern of the intersection determined by the intersection determination unit, and an evaluation condition selection unit.
A safety confirmation evaluation unit that evaluates the driver's safety confirmation operation at the intersection based on the evaluation conditions selected by the evaluation condition selection unit and the detection data detected at the intersection where the driving evaluation is performed .
A detection data storage unit that stores the detection data detected by the on-board unit when the vehicle is passing through an intersection, and a detection data storage unit.
A discrimination data storage unit for storing the discrimination data is provided.
The discrimination data acquisition unit
The discrimination data is acquired from the discrimination data storage unit by using at least one of the detection data stored in the detection data storage unit and detected at the intersection where the operation evaluation is performed. ,
The discrimination data stored in the discrimination data storage unit includes map data capable of extracting an intersection map including an intersection through which the vehicle has passed.
The discrimination data acquired by the discrimination data acquisition unit is intersection map data including the intersection through which the vehicle has passed, which is extracted from the map data.
The intersection discriminating unit discriminates the shape pattern of the intersection for which the driving evaluation is performed based on the intersection map data.
It has a neural network that includes an input layer, an intermediate layer, and an output layer.
The intersection map data is input to the input layer as the discrimination data, and the intersection map data is input.
In the intermediate layer, an operation based on a weighting coefficient is executed on the intersection map data, and the operation is performed.
An operation evaluation device characterized in that the output layer is configured to output the determination result of the shape pattern of the intersection . A driving evaluation device that evaluates a driver's driving using detection data including the driver's condition, the position of the vehicle, and the behavior of the vehicle detected by the on-board unit.
A discrimination data acquisition unit that acquires discrimination data using at least one of the detected data detected at the intersection where the driving evaluation is performed, and a discrimination data acquisition unit.
Using the discrimination data, an intersection discrimination unit that discriminates the shape pattern of the intersection for which the operation evaluation is performed, and an intersection discrimination unit.
An evaluation table storage unit that stores an evaluation table in which evaluation conditions for safety confirmation operations to be performed by the driver are set for each intersection shape pattern, and
From the evaluation table, an evaluation condition selection unit that selects the evaluation condition corresponding to the shape pattern of the intersection determined by the intersection determination unit, and an evaluation condition selection unit.
A safety confirmation evaluation unit that evaluates the driver's safety confirmation operation at the intersection based on the evaluation conditions selected by the evaluation condition selection unit and the detection data detected at the intersection where the driving evaluation is performed.
A detection data storage unit that stores the detection data detected by the on-board unit when the vehicle is passing through an intersection, and a detection data storage unit.
A discrimination data storage unit for storing the discrimination data is provided.
The discrimination data acquisition unit
The discrimination data is acquired from the discrimination data storage unit by using at least one of the detection data stored in the detection data storage unit and detected at the intersection where the operation evaluation is performed. ,
The detection data further includes external image data obtained by capturing or detecting the external situation of the vehicle.
The discrimination data stored in the discrimination data storage unit includes the outside world image data transmitted from the vehicle-mounted device.
The discrimination data acquired by the discrimination data acquisition unit is the outside world image data imaged or detected at an intersection through which the vehicle has passed.
The intersection discriminating unit discriminates the shape pattern of the intersection for which the operation evaluation is performed based on the outside world image data.
It has a neural network that includes an input layer, an intermediate layer, and an output layer.
The external image data is input to the input layer as the discrimination data, and the external image data is input to the input layer.
In the intermediate layer, an operation based on a weighting coefficient is executed on the external image data, and the operation is performed.
An operation evaluation device characterized in that the output layer is configured to output the determination result of the shape pattern of the intersection . Using the detection data detected at the intersection where the driving evaluation is performed, a direction estimation unit for estimating the bending direction at the intersection where the driving evaluation is performed is provided.
In the evaluation table,
For each combination of the shape pattern of the intersection and the direction of the turn at the intersection, the evaluation conditions for the safety confirmation operation that the driver should perform are set.
The evaluation condition selection unit
The evaluation table is characterized in that the evaluation conditions corresponding to the shape pattern of the intersection determined by the intersection determination unit and the bending direction at the intersection estimated by the direction estimation unit are selected. The operation evaluation device according to claim 1 or 2 . The evaluation conditions of the evaluation table include the timing, confirmation angle, and confirmation time of the safety confirmation operation to be performed by the driver.
The safety confirmation evaluation unit evaluates the timing, the confirmation angle, and the confirmation time at each intersection where the operation evaluation is performed, and outputs the evaluation result scored or ranked based on each evaluation. The operation evaluation device according to any one of claims 1 to 3 , wherein the operation evaluation device is characterized. The operation evaluation device according to any one of claims 1 to 4 , and the operation evaluation device.
Consists of including on-board equipment mounted on the vehicle ,
The in-vehicle device
A driver monitoring unit that detects the state of the driver of the vehicle, and
A position detection unit that detects the position of the vehicle and
A behavior detection unit that detects the behavior of the vehicle,
It is characterized by including a driver monitoring unit, a position detection unit, and an output unit that outputs data detected by the behavior detection unit to an operation evaluation device that evaluates the safety confirmation operation of the driver. Driving evaluation system. A driving evaluation method for evaluating a driver's driving using detection data including the driver's condition, the position of the vehicle, and the behavior of the vehicle detected by the on-board unit.
A first step of acquiring discrimination data used for discriminating the shape pattern of the intersection for which the driving evaluation is performed by using at least one of the detected data detected at the intersection for which the driving evaluation is performed.
Using the discrimination data acquired in the first step, the second step of discriminating the shape pattern of the intersection for which the operation evaluation is performed, and the second step.
From the evaluation table in which the evaluation conditions for the safety confirmation operation to be performed by the driver are set for each shape pattern of the intersection, the evaluation condition corresponding to the shape pattern of the intersection determined in the second step is selected. 3 steps and
Based on the evaluation conditions selected by the third step and the detection data detected at the intersection where the driving evaluation is performed, the fourth step of evaluating the driver's safety confirmation operation at the intersection. Perform the steps that include
The first step is
At least of the detection data detected at the intersection where the driving evaluation is performed, which is stored in the detection data storage unit that stores the detection data detected by the on-board unit when the vehicle is passing through the intersection. It is a step of acquiring the discrimination data from the discrimination data storage unit by using any of them.
The discrimination data stored in the discrimination data storage unit includes map data capable of extracting an intersection map including an intersection through which the vehicle has passed.
The discrimination data acquired in the first step is intersection map data including the intersection through which the vehicle has passed, which is extracted from the map data.
The second step is a step of determining the shape pattern of the intersection to be evaluated for driving based on the intersection map data.
The intersection map data is input to the input layer as the discrimination data, and the intersection map data is input.
In the middle layer, an operation based on the weighting coefficient is executed on the intersection map data, and the operation is performed.
An operation evaluation method comprising a neural network including the input layer, the intermediate layer, and the output layer, which is configured to output the discrimination result of the shape pattern of the intersection from the output layer . A driving evaluation method for evaluating a driver's driving using detection data including the driver's condition, the position of the vehicle, and the behavior of the vehicle detected by the on-board unit.
A first step of acquiring discrimination data used for discriminating the shape pattern of the intersection for which the driving evaluation is performed by using at least one of the detected data detected at the intersection for which the driving evaluation is performed.
Using the discrimination data acquired in the first step, the second step of discriminating the shape pattern of the intersection for which the operation evaluation is performed, and the second step.
From the evaluation table in which the evaluation conditions for the safety confirmation operation to be performed by the driver are set for each shape pattern of the intersection, the evaluation condition corresponding to the shape pattern of the intersection determined in the second step is selected. 3 steps and
Based on the evaluation conditions selected by the third step and the detection data detected at the intersection where the driving evaluation is performed, the fourth step of evaluating the driver's safety confirmation operation at the intersection. Perform the steps that include
The first step is
At least of the detection data detected at the intersection where the driving evaluation is performed, which is stored in the detection data storage unit that stores the detection data detected by the on-board unit when the vehicle is passing through the intersection. It is a step of acquiring the discrimination data from the discrimination data storage unit by using any of them.
The detection data further includes external image data obtained by capturing or detecting the external situation of the vehicle.
The discrimination data stored in the discrimination data storage unit includes the outside world image data transmitted from the vehicle-mounted device.
The discrimination data acquired in the first step is the outside world image data imaged or detected at an intersection through which the vehicle has passed.
The second step is a step of determining the shape pattern of the intersection to be evaluated for operation based on the external image data.
The external image data is input to the input layer as the discrimination data, and the external image data is input to the input layer.
In the intermediate layer, an operation based on the weighting coefficient is executed on the external image data, and the operation is performed.
An operation evaluation method comprising a neural network including the input layer, the intermediate layer, and the output layer, which is configured to output the discrimination result of the shape pattern of the intersection from the output layer. A driving evaluation program for causing at least one computer to perform a process of evaluating a driver's driving using detection data including the state of the driver, the position of the vehicle, and the behavior of the vehicle detected by the in-vehicle device. And,
To the at least one computer
A first step of acquiring discrimination data used for discriminating the shape pattern of the intersection for which the driving evaluation is performed by using at least one of the detected data detected at the intersection for which the driving evaluation is performed.
Using the discrimination data acquired in the first step, the second step of discriminating the shape pattern of the intersection for which the operation evaluation is performed, and the second step.
From the evaluation table in which the evaluation conditions for the safety confirmation operation to be performed by the driver are set for each shape pattern of the intersection, the evaluation condition corresponding to the shape pattern of the intersection determined in the second step is selected. 3 steps and
Based on the evaluation conditions selected by the third step and the detection data detected at the intersection where the driving evaluation is performed, the fourth step of evaluating the driver's safety confirmation operation at the intersection. Let it run
The first step is
At least of the detection data detected at the intersection where the driving evaluation is performed, which is stored in the detection data storage unit that stores the detection data detected by the on-board unit when the vehicle is passing through the intersection. It is a step of acquiring the discrimination data from the discrimination data storage unit by using any of them.
The discrimination data stored in the discrimination data storage unit includes map data capable of extracting an intersection map including an intersection through which the vehicle has passed.
The discrimination data acquired in the first step is intersection map data including the intersection through which the vehicle has passed, which is extracted from the map data.
The second step is a step of determining the shape pattern of the intersection to be evaluated for driving based on the intersection map data.
The intersection map data is input to the input layer as the discrimination data, and the intersection map data is input.
In the middle layer, an operation based on the weighting coefficient is executed on the intersection map data, and the operation is performed.
An operation evaluation program characterized by using a neural network including the input layer, the intermediate layer, and the output layer, which is configured to output the discrimination result of the shape pattern of the intersection from the output layer . A driving evaluation program for causing at least one computer to perform a process of evaluating a driver's driving using detection data including the state of the driver, the position of the vehicle, and the behavior of the vehicle detected by the in-vehicle device. And,
To the at least one computer
A first step of acquiring discrimination data used for discriminating the shape pattern of the intersection for which the driving evaluation is performed by using at least one of the detected data detected at the intersection for which the driving evaluation is performed.
Using the discrimination data acquired in the first step, the second step of discriminating the shape pattern of the intersection for which the operation evaluation is performed, and the second step.
From the evaluation table in which the evaluation conditions for the safety confirmation operation to be performed by the driver are set for each shape pattern of the intersection, the evaluation condition corresponding to the shape pattern of the intersection determined in the second step is selected. 3 steps and
Based on the evaluation conditions selected by the third step and the detection data detected at the intersection where the driving evaluation is performed, the fourth step of evaluating the driver's safety confirmation operation at the intersection. Let it run
The first step is
At least of the detection data detected at the intersection where the driving evaluation is performed, which is stored in the detection data storage unit that stores the detection data detected by the on-board unit when the vehicle is passing through the intersection. It is a step of acquiring the discrimination data from the discrimination data storage unit by using any of them.
The detection data further includes external image data obtained by capturing or detecting the external situation of the vehicle.
The discrimination data stored in the discrimination data storage unit includes the outside world image data transmitted from the vehicle-mounted device.
The discrimination data acquired in the first step is the outside world image data imaged or detected at an intersection through which the vehicle has passed.
The second step is a step of determining the shape pattern of the intersection to be evaluated for operation based on the external image data.
The external image data is input to the input layer as the discrimination data, and the external image data is input to the input layer.
In the intermediate layer, an operation based on the weighting coefficient is executed on the external image data, and the operation is performed.
An operation evaluation program characterized by using a neural network including the input layer, the intermediate layer, and the output layer, which is configured to output the discrimination result of the shape pattern of the intersection from the output layer.

Images