JP2012247871A - Driving skill discrimination device, and driving skill discrimination program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To discriminate whether safety confirmation by the visual sense of a driver becomes emasculated or not.SOLUTION: A driving skill discrimination device includes: a visual line data obtaining portion for obtaining visual line information of a driver of a vehicle; a driving data obtaining portion for obtaining driving operation information and behavior information of the vehicle; a peripheral vehicle data obtaining portion for obtaining distance information from the vehicle to other vehicles; a history storage portion for accumulating each of the obtained visual line information, the driving operation information, the behavior information, and the distance information; a driving action storage portion for storing driving action according to the behavior information and the driving operation information; a safety information storage portion for storing information on the vehicle to the driving action, a safe distance relationship from other vehicles, and visual line movement; a driving action history producing portion for referring to the driving action storage portion and producing a driving action history from the driving operation information and the behavior information accumulated in the history storage portion; and a driver classifying portion for extracting from the driving action history the driving action which is dangerous and during which the visual line movement to the driving action was correct, on the basis of the distance from the peripheral vehicles.

Description

  The present invention relates to an apparatus for discriminating a driving skill of a driver.

  There is a technique for evaluating driving skill by acquiring data when a vehicle is driven by a driver. For example, there is a technique used as an evaluation index for evaluating the degree of risk and annoyance of driving by a driver when changing the lane of a vehicle. This technology measures a physical quantity required for a driving operation by a driver who is changing lanes, and calculates a risk level and an annoyance level from a measurement result and an evaluation function.

  In addition, there is a technique for diagnosing a driving technique based on an operation behavior when a vehicle is driven by a driver. For example, this technique pays attention to the fact that the fluctuation is small when the car travels smoothly, acquires the fluctuation amount of the driver's operation, and determines that the driving is good when the operation is small.

  There is a technology that compares the driver's head check when changing lanes with existing good habits.

  There is also a technique for evaluating the driver's safety confirmation behavior based on gaze measurement data. This technique selects an object to be watched for safety confirmation in a driving scene to be diagnosed, and also measures the driver's line of sight to determine whether the driver is gazing at the object. It is measured whether or not is greater than or equal to a threshold value, and the visual behavior of the driver is evaluated.

Japanese Patent No. 3593502 Japanese Patent No. 4307833 JP 2007-293495 A

Automotive Engineering Society Academic Lecture Preprints No. 38-10 (May 2010), literature number: 201005255.

  As described above, there is a conventional technique for individually evaluating a driver's confirmation behavior and driving operation. The conventional technology can determine whether or not the user is looking in the direction to be viewed according to the evaluation criterion.

  However, even if the driver looks at an appropriate direction at an appropriate timing, the driver may be in danger driving as a result. In this case, since the prior art determines that the driver's safety check is appropriate, it cannot be determined that the driver's driving is dangerous.

  It is an object of the disclosed driving skill determination device to be able to determine whether or not a driver's visual safety confirmation is in a form.

  The disclosed driving technology determination device includes a line-of-sight data acquisition unit that acquires information on the line of sight of a driver who drives the vehicle, a driving data acquisition unit that acquires information related to the driving operation of the driver and information about the behavior of the vehicle, and the vehicle Of the surrounding vehicle data acquisition unit that acquires information on the distance from the vehicle to the other vehicle, information on the driver's line of sight, information on the driving operation, information on the behavior of the vehicle, and information on the distance to the other vehicle A history storage unit that accumulates each, a driving operation storage unit that stores information related to vehicle behavior and driving operation information, a vehicle for driving operation, a safe distance relationship with other vehicles, And the safety information storage unit that stores the information on the movement of the line of sight and the driving operation storage unit, the information related to the driving operation and the information about the behavior of the vehicle that is accumulated in the history storage unit. Referring to the driving operation history generation unit that generates the driving operation history and the safety information storage unit, based on the distance from the surrounding vehicle in the generated driving operation history, the movement of the line of sight with respect to the driving operation is dangerous And a driver classifying unit that extracts a driving action that is correct.

  The disclosed driving skill determination device can determine whether or not the driver's visual safety confirmation is broken up.

FIG. 1 is a functional block diagram of the driving skill determination device 1 of this embodiment. FIG. 2 is a hardware configuration diagram of the driving technique determination device 1 according to the present embodiment. FIG. 3 is a first flowchart of a process when calculating the driving risk of the driving skill determination device 1. FIG. 4 is a second flowchart of the process when calculating the driving risk of the driving skill determination device 1. FIG. 5 is a third flowchart of the process when calculating the driving risk of the driving skill determination device 1. FIG. 6 is a fourth flowchart of the processing when calculating the driving risk of the driving skill determination device 1. FIG. 7 shows the configuration of the operation data table 23. FIG. 8 shows an example of the driving operation storage unit 25. FIG. 9 is a flowchart of the calculation process of the driving risk when changing lanes. FIG. 10 is a continuation of the flowchart of the calculation process of the driving risk when changing lanes. FIG. 11 is an explanatory diagram of an index of driving risk when changing lanes. FIG. 12 is a flowchart of line-of-sight data acquisition processing. FIG. 13 is an explanatory diagram for obtaining the direction of the driver's line of sight. FIG. 14 shows the data structure of the line-of-sight data table 26. FIG. 15 is a diagram illustrating the relationship between the driver's line of sight and the mirror. FIG. 16 is a diagram illustrating the relationship between the driver's line of sight, the mirror, and the rear vehicle. FIG. 17 is a flowchart of the mirror confirmation process flag storage process. FIG. 18 is an example of a desirable form of the visual confirmation action required in the preparation stage when changing lanes. FIG. 19 is a flowchart of the safety confirmation degree calculation process. FIG. 20 is a second flowchart of the safety confirmation degree calculation process. FIG. 21 shows an example of driver classification results.

[Device functions]
FIG. 1 is a functional block diagram of the driving skill determination device 1 of this embodiment.

  The driving technology determination device 1 includes a driving data acquisition unit 11, a surrounding vehicle data acquisition unit 12, a surrounding vehicle speed, an acceleration calculation unit 13, a driving risk calculation unit 14, a line-of-sight data acquisition unit 15, a mirror confirmation detection unit 16, and a safety confirmation. A degree calculation unit 17, a driver classification unit 18, a driving guidance presentation unit 19, a history storage unit 20, a driving operation storage unit 21, and a safety information storage unit 22 are included.

  The driving data acquisition unit 11 acquires information on operations such as steering angle of the steering wheel, turn-on / off of the turn signal switch, depression angle of the accelerator pedal, depression angle of the brake pedal, and the speed of the host vehicle by the driver.

  The surrounding vehicle data acquisition unit 12 acquires a distance from a vehicle (hereinafter referred to as a surrounding vehicle) existing around a vehicle driven by the driver (hereinafter referred to as the own vehicle). The surrounding vehicle data acquisition unit 12 periodically acquires the distance between the host vehicle and the surrounding vehicle. In addition, the surrounding vehicle data acquisition part 12 acquires the positional information on a surrounding vehicle, and the information of the area | region where a surrounding vehicle exists, for example. The position information of the surrounding vehicle and the information on the area where the surrounding vehicle exists are used when determining whether or not the driver has visually recognized the surrounding vehicle through the mirror.

  The surrounding vehicle speed / acceleration calculation unit 13 calculates the speed and acceleration of the surrounding vehicle from the difference between the distance between the own vehicle and the surrounding vehicle that is periodically acquired and the speed of the own vehicle.

  The driving risk calculation unit 14 calculates the driving risk of the driver from the speed difference and distance between the host vehicle and the surrounding vehicles. The risk level of the present embodiment is high because the accident is more likely when the speed of the host vehicle is slower than the rear vehicle or when the distance between the rear vehicle and the host vehicle is closer.

The line-of-sight data acquisition unit 15 acquires a driver's line-of-sight history. For example, a driver's face image is acquired by a camera or the like, and the mirror confirmation detection unit 16 determines whether the driver's line of sight is a mirror from the face and eye positions. Further, when the driver looks at the mirror, it detects whether the surrounding vehicle is reflected on the mirror.

  The safety confirmation degree calculation unit 17 generates a history of driving operations and determines whether or not the direction of the driver's line of sight is looking at an appropriate direction at an appropriate timing.

  The driver classification unit 18 classifies the type of the driver from the past driving risk history and safe driving history.

  The driving guidance presentation unit 19 registers a guidance method in advance according to the type of the driver, and outputs a guidance method according to the type of the driver classified by the driver classification unit 18.

  The history storage unit 20 includes each driving data acquisition unit 11, surrounding vehicle data acquisition unit 12, surrounding vehicle speed, acceleration calculation unit 13, driving risk calculation unit 14, line-of-sight data acquisition unit 15, mirror confirmation detection unit 16, safety The results output by the confirmation degree calculation unit 17, the driver classification unit 18, and the driving guidance presentation unit 19 are stored as history information.

  The driving operation storage unit 21 stores information for specifying the driving operation from the relationship between the information related to the behavior of the host vehicle and the information related to the driving operation.

  The safety information storage unit 22 stores information on a safe distance between the host vehicle and surrounding vehicles and a desired timing of eye movement with respect to the driving operation of the driver.

[Hardware configuration]
FIG. 2 is a hardware configuration diagram of the driving technique determination device 1 according to the present embodiment. The driving skill determination device 1 includes, for example, a CPU 31, a RAM 32, a ROM 33, a monitor 34, a driver camera 35, a radar 36, an input unit 37, a storage unit 38, and a vehicle camera 39.

  The CPU 31 executes the program so that the RAM 32, the ROM 33, the monitor 34, the driver camera 35, the radar 36, the input unit 37, the storage unit 38, and the vehicle camera 39 are interlocked with the driving data acquisition unit 11 and the surrounding vehicle data acquisition. Unit 12, surrounding vehicle speed, acceleration calculation unit 13, driving risk calculation unit 14, line-of-sight data acquisition unit 15, mirror confirmation detection unit 16, safety confirmation degree calculation unit 17, driver classification unit 18, and driver guidance presentation unit 19 Function, driving data acquisition processing of the driving technology discrimination device 1, vehicle surrounding data acquisition processing, surrounding vehicle speed and acceleration calculation processing, driving risk calculation processing, line-of-sight data acquisition processing, mirror confirmation detection processing, safety confirmation degree calculation processing, A driver classification process and a driving guidance presentation process are executed.

  The RAM 32 is a readable / writable memory and functions as a work area when the CPU 31 executes various processes. Data, output results, programs, and the like during the various processes by the CPU 31 are temporarily stored.

  The ROM 33 is a read-only memory, and stores a program used for the CPU 31 to operate in advance. The CPU 31 reads an execution program from the ROM 33 into the RAM 32 and executes various processes.

  The monitor 34 displays the processing result by the CPU 31 and the like.

  The driver camera 35 is a camera that captures a driver's face image in order to acquire the direction of the driver's line of sight. The CPU 31 specifies the direction of the driver's line of sight from the image of the driver camera 35.

  The radar 36 obtains the distance to the object around the host vehicle. The CPU 31 acquires distance information between the host vehicle and surrounding vehicles from the radar 36. In the present embodiment, the host vehicle is equipped with a plurality of radars 36 in order to acquire a wide range behind the host vehicle. Further, the CPU 31 uses the data of the radar 36 to acquire information on the position and size of surrounding vehicles with respect to the host vehicle.

  The input unit 37 receives state information of the host vehicle. The state information of the host vehicle is information such as the speed of the host vehicle, the steering angle of the steering of the host vehicle, the turn signal switch, and the like. The vehicle of the present embodiment has a controller area network (CAN), and receives state information of the host vehicle via the CAN.

  The storage unit 38 is a non-volatile data storage area, and stores various processing results by the CPU 31, execution programs, table information registered in advance, and the like.

  The vehicle camera 39 captures an image around the host vehicle. The vehicle camera 39 may image a road, for example. CPU31 specifies that the own vehicle straddled the lane from the image which the vehicle camera 39 imaged. A plurality of vehicle cameras 39 may be attached to the own vehicle.

[Driving risk calculation process]
3, 4, 5, and 6 are flowcharts of processing procedures for various sensors mounted on the host vehicle. FIG. 3 is a first flowchart of a process when calculating the driving risk of the driving skill determination device 1. FIG. 3 shows detection by the radar 36 mounted on the host vehicle and processing of the surrounding vehicle data acquisition unit 12. FIG. 4 is a second flowchart of the process when calculating the driving risk of the driving skill determination device 1. FIG. 4 shows the detection of the speed of the host vehicle by acquiring the value of the speed sensor mounted on the host vehicle via the input unit 37 and the processing of the surrounding vehicle data acquisition unit 12. FIG. 5 is a third flowchart of the process when calculating the driving risk of the driving skill determination device 1. FIG. 5 shows detection of a situation around the host vehicle by the vehicle camera 39 mounted on the host vehicle, and processing of the surrounding vehicle data acquisition unit 12. FIG. 6 is a fourth flowchart of the processing when calculating the driving risk of the driving skill determination device 1. FIG. 6 shows the detection of the steering angle of the host vehicle by acquiring the value of the steering sensor mounted on the host vehicle via the input unit 37 and the processing of the surrounding vehicle data acquisition unit 12. Note that the radar 36, the speed sensor, the steering sensor, and the vehicle camera 39 operate independently, and each acquire data at a predetermined timing.

  The driving data acquisition unit 11 and the surrounding vehicle data acquisition unit 12 store various information related to the own vehicle and the surrounding vehicles acquired by executing the processes of the flowcharts of FIGS. 3 to 6 in the operation data table 23 of the history storage unit 20. Store.

  FIG. 7 shows the configuration of the operation data table 23. The operation data table 23 includes a frame number 231, a speed 232, a rear vehicle distance 233, a rear vehicle relative speed 234, a rear vehicle distance fluctuation amount 235, a TTC 236, a THW 237, a winker lighting flag 238, a steering start flag 239, a white line arrival flag 250, And lane change time 241.

  The frame number 231 is information for identifying data to be stored every predetermined time. The speed 232 is the speed of the host vehicle. The rear vehicle distance 233 is a distance between the rear vehicle and the host vehicle. The rear vehicle relative speed 234 is a speed difference between the rear vehicle and the host vehicle. The rear vehicle distance fluctuation amount 235 is a change amount of the distance between the rear vehicle and the host vehicle.

  TTC 236 is an index of the degree of risk that the rear vehicle will collide with the host vehicle, and is a collision margin time (Time To Collation) representing the time until the rear vehicle collides with the host vehicle. The collision margin time TTC indicates a time until the vehicle collides with the host vehicle when the vehicle behind the vehicle travels while maintaining the traveling state at the time of calculating the collision margin time TTC. This is an index representing the “probability of collision” that a rear vehicle always collides with the host vehicle (with a probability of 100%) when the vehicle travels while maintaining the traveling state at the time of calculation.

  THW237 is the head time between the rear vehicle and the host vehicle.

  The turn signal lighting flag 238 is information indicating whether or not the turn signal is turned on. The turn signal lighting flag 238 is managed separately for the right turn signal and the left turn signal. In the present embodiment, since the right lane change is targeted, the turn signal lighting flag 238 stores only the state of the right turn signal.

  The steering start flag 239 stores “1” when the steering angle of the steering of the host vehicle becomes a predetermined angle or more with respect to the traveling direction. The steering start flag 239 stores “0” except when the steering angle of the steering of the host vehicle is not less than a predetermined angle with respect to the traveling direction. Note that the steering start flag 239 can store not only information indicating whether or not the steering angle is equal to or greater than a predetermined steering angle (information indicating whether or not the steering is started) but also a history of steering angles every predetermined time. is there.

  The white line arrival flag 240 stores “1” when the host vehicle passes the white line portion of the road information. The white line arrival flag 240 stores “1” except when the vehicle passes the white line portion of the road information.

  The lane change time 241 stores “1” when it is from the start to the end of the lane change. The lane change time 241 stores “1” except when it is from the start to the end of the lane change.

  The flowchart will be described below. In each flowchart described below, the data recording itself needs to be performed in real time during the operation. However, the processing performed using the recorded data may be performed during the operation or after the fact. It may be executed.

  The surrounding vehicle data acquisition unit 12 acquires the output result of the radar 36 mounted on the vehicle (S01). The surrounding vehicle data acquisition unit 12 calculates a distance between the own vehicle and a surrounding vehicle located behind the own vehicle (hereinafter referred to as a rear vehicle) from the output result of the radar 36 acquired in S01 (S02). ). In the present embodiment, an example will be described in which the radar 36 calculates the degree of risk from the distance to surrounding vehicles existing behind the host vehicle. It is assumed that the radar 36 can also acquire the distance for the surrounding vehicle in front of the host vehicle. The driving skill determination device 1 can also calculate the degree of risk from the distance to the surrounding vehicle ahead. In addition, the surrounding vehicle data acquisition part 12 calculates the distance with a surrounding vehicle for every predetermined time, such as 0.1 second, for example. The surrounding vehicle data acquisition unit 12 stores information on the distance (rear distance) with the surrounding vehicle existing behind the host vehicle calculated in S02 in the rear vehicle distance 233 of the operation data table 23 (S03).

  The surrounding vehicle data acquisition unit 12 accumulates information on the rear distance for each predetermined time. The peripheral vehicle speed / acceleration calculation unit 13 calculates the amount of change in the distance between the rear vehicle and the host vehicle (S04). The peripheral vehicle speed / acceleration calculation unit 13 calculates the amount of change in the distance per predetermined time from the information on the rear distance accumulated every predetermined time in S03. The surrounding vehicle data acquisition unit 12 stores the amount of change in the distance between the rear vehicle and the host vehicle every predetermined time calculated in S04 in the rear vehicle distance fluctuation amount 235 (S05). It is assumed that the surrounding vehicle data acquisition unit 12 accumulates information on the amount of change in distance every predetermined time.

  The surrounding vehicle speed / acceleration calculation unit 13 calculates the relative speed of the vehicle behind the host vehicle (S06). The peripheral vehicle speed / acceleration calculation unit 13 calculates the speed difference between the host vehicle and the rear vehicle based on the rear distance accumulated in S03. For example, the surrounding vehicle speed / acceleration calculation unit 13 calculates the amount of change in the rear distance every predetermined time, and calculates the relative speed of the rear vehicle with respect to the host vehicle from the amount of change in the rear distance and the predetermined time.

  The surrounding vehicle data acquisition unit 12 stores the information on the relative speed of the rear vehicle with respect to the host vehicle calculated in S04 in the rear vehicle relative speed 234 (S07). It is assumed that the surrounding vehicle data acquisition unit 12 accumulates information on the relative speed for each predetermined time.

  The surrounding vehicle speed / acceleration calculation unit 13 calculates TTC (S08). The surrounding vehicle data acquisition unit 12 stores the TTC for each predetermined time calculated in S08 in the TTC 236 of the operation data table 23 (S09). The surrounding vehicle data acquisition part 12 shall accumulate | store TTC for every predetermined time.

  For example, the driving data acquisition unit 11 acquires the speed of the host vehicle every predetermined time via a CAN mounted on the vehicle (S10). The driving data acquisition unit 11 stores the speed of the host vehicle every predetermined time acquired in S10 in the storage unit 38 (S11). The driving data acquisition part 11 shall accumulate | store the speed of the own vehicle for every predetermined time.

  The driving data acquisition unit 11 determines the speed of the host vehicle for each predetermined time accumulated in S11, the relative speed of the rear vehicle for each predetermined time accumulated in S07, and the rear vehicle for each predetermined time accumulated in S03. From the distance between the vehicles, THW is calculated every predetermined time (S12). The vehicle head time indicates the time from when the host vehicle passes a certain point until the rear vehicle passes the point. The driving data acquisition unit 11 stores the calculated vehicle head time for each predetermined time in the THW 237 of the operation data table 23 (S13). The driving data acquisition unit 11 accumulates the vehicle head time for each predetermined time.

  The driving data acquisition unit 11 acquires the turn-on / off state of the blinker (S14). The driving data acquisition unit 11 stores the blinker lighting / extinguishing state acquired in S14 in the blinker lighting flag 238 (S15). It is assumed that the operation data acquisition unit 11 accumulates a history of the turn-on / off state of the blinker every predetermined time.

  Next, the driving data acquisition unit 11 detects the start of the steering of the host vehicle (S16). The driving data acquisition unit 11 stores the steering start timing detected in S16 in the steering start flag 239 (S17). The driving data acquisition unit 11 accumulates a history of the timing of the start of steering of the steering of the host vehicle.

  The driving data acquisition unit 11 detects the arrival of the host vehicle on the white line from the road image captured by the vehicle camera 39 mounted on the host vehicle (S18). The reason why the arrival of the own vehicle to the white line is detected is to detect that the own vehicle straddles the lane when the lane is changed. Note that the white line detected when the lane is changed is a lane boundary line that divides the first lane and the second lane in one direction when the road has four lanes. The driving data acquisition unit 11 stores the timing at which the host vehicle reaches the white line detected in S18 in the white line arrival flag 250 (S19). The driving data acquisition unit 11 accumulates a history of the timing of arrival of the host vehicle to the white line.

  The driving data acquisition unit 11 determines the driver from the timing of the steering start state of the host vehicle stored in the storage unit 38 in S17 and the timing of the host vehicle reaching the white line stored in the storage unit 38 in S19. Calculates the time from the start of steering until the host vehicle reaches the white line (S20). The driving data acquisition unit 11 stores, in the lane change time 241, the time from when the driver calculated in S <b> 20 starts steering until the host vehicle reaches the white line (S <b> 21). For example, the driving data acquisition unit 11 associates with the steering start time so as to be able to determine at which time the lane change was made, and accumulate the lane change time.

  The driving risk calculation unit 14 changes the driver's lane using the rear vehicle distance fluctuation amount accumulated in S05, the rear collision time accumulated in S09, the rear head time accumulated in S13, and the lane change time accumulated in S21. Is calculated (S22).

[Driving operation identification]
The driving operation storage unit 25 is information in which types of driving operations such as left turn, right turn, curve, lane change, overtaking operation, and the like are classified according to the steering angle of the steering, the speed of the host vehicle, and the presence / absence of passage of the white line. It is assumed that the driving operation storage unit 25 is stored in advance. The driving operation is information of an operation such as a steering angle of the steering, an accelerator amount, a brake amount, and turn-on / off of the winker performed when the driver drives.

  FIG. 8 shows an example of the driving operation storage unit 25.

  The driving operation storage unit 25 includes, for example, a driving type 251, a speed 252, a steering angle 253, a blinker lighting 254, a white line passage 255, and the like. The driving type 251 indicates the type of driving operation. The speed 252 indicates the speed of the vehicle defined for each type of driving operation. The steering angle 253 indicates the steering angle of the driver's steering defined for each type of driving operation. The turn signal lighting 254 indicates whether or not the turn signal lighting is defined for each type of driving operation. A white line passage 255 indicates whether or not a white line defined for each type of driving operation has passed.

  In the case of a right or left curve, the blinker does not light and the steering angle is small. In the case of a right turn or a left turn, the turn signal lights, the traveling speed of the host vehicle is slow, and the steering angle is large.

  In this embodiment, a lane change will be described as an example. For example, the change in the steering angle when changing the lane from the left lane to the right lane is as follows. First, the rudder angle of the front wheel of the host vehicle is generated on the right, and the host vehicle starts moving to the right lane. Thereafter, the rudder angle of the front wheels of the host vehicle becomes 0 degrees, and the host vehicle travels straight during the period between the right lane and the left lane. Thereafter, the rudder angle of the front wheel of the host vehicle is generated on the left, and the host vehicle completes the movement to the right lane.

  The driving data acquisition unit 11 and the surrounding vehicle data acquisition unit 12 store all information at the time of driving in the operation data table 23, and the driving risk calculation unit 14 stores various history information stored in the operation data table 23 and The type of driving action is specified based on the degree of coincidence with the characteristics of the classified driving action types. The driving risk calculation unit 14 can also perform a driving risk determination process according to the specified driving type.

[Driving risk calculation process when changing lanes]
FIG. 9 is a flowchart of the calculation process of the driving risk when changing lanes.

  The driving data acquisition unit 11 detects the start of lane change (S31). For example, when the driving data acquisition unit 11 detects a steering angle equal to or greater than a predetermined angle, the driving data acquisition unit 11 determines that the lane change is started by the driver. The start of the lane change may be based on a time before a predetermined time of turning on the blinker.

  When the start of lane change is detected, the driving data acquisition unit 11 stores the start timing of lane change by setting the steering start flag 239 to “1” (S32). Further, the driving data acquisition unit 11 sets the value of the lane change time 241 to “1” until the lane change is completed.

  The driving data acquisition unit 11 acquires the speed of the host vehicle (S33). The driving data acquisition unit 11 stores the speed of the host vehicle acquired in S <b> 33 in the speed 232. In addition, the driving data acquisition part 11 accumulate | stores the speed of the own vehicle in the speed 232 for every predetermined time after S33 until lane change is completed.

  The surrounding vehicle data acquisition unit 12 detects the distance to the rear vehicle existing in the lane to be changed (S34). The surrounding vehicle data acquisition unit 12 stores the distance to the rear vehicle existing in the lane to be changed calculated in S34 in the rear vehicle distance 233. In addition, the surrounding vehicle data acquisition part 12 accumulate | stores the distance to the back vehicle which exists in a change destination lane in the back vehicle distance 233 for every predetermined time after S34 until lane change is completed.

  The surrounding vehicle speed / acceleration calculation unit 13 calculates the speed of the rear vehicle (S35). In this embodiment, the surrounding vehicle speed / acceleration calculation unit 13 calculates a speed difference (relative speed) between the host vehicle and the rear vehicle as the speed of the rear vehicle. The peripheral vehicle speed / acceleration calculation unit 13 stores the relative speed of the rear vehicle calculated in S35 in the rear vehicle relative speed 234. The peripheral vehicle speed / acceleration calculation unit 13 may accumulate the relative speed of the rear vehicle in the rear vehicle relative speed 234 every predetermined time until the lane change is completed after S35.

  The peripheral vehicle speed / acceleration calculation unit 13 calculates the TTC of the rear vehicle from the rear vehicle relative distance data 283 and the rear vehicle relative speed data 284 (S36). The surrounding vehicle speed / acceleration calculation unit 13 stores the TTC of the rear vehicle calculated in S36 in the TTC 236. The peripheral vehicle speed / acceleration calculation unit 13 accumulates the TTC of the rear vehicle in the TTC 236 every predetermined time until the lane change is completed after S36.

  The surrounding vehicle speed / acceleration calculation unit 13 calculates the THW of the rear vehicle from the own vehicle speed data 282, the rear vehicle relative distance data 283, and the rear vehicle relative speed data 284 (S37). The peripheral vehicle speed / acceleration calculation unit 13 stores the THW of the rear vehicle calculated in S37 in the THW 237. The peripheral vehicle speed / acceleration calculation unit 13 may accumulate the THW of the rear vehicle in the THW 237 every predetermined time until the lane change is completed after S37.

  The driving data acquisition unit 11 determines whether or not the host vehicle has exceeded the white line between the lanes (S38). When it is determined that the host vehicle does not exceed the white line between the lanes (S38: No), the driving data acquisition unit 11 continues to execute the processes after S33. When it is determined that the own vehicle has crossed the white line between the lanes (S38: Yes), the driving data acquisition unit 11 stores the timing at which the own vehicle has crossed the white line between the lanes in the white line arrival flag 250 (S39). In addition, since this flowchart exemplifies lane change, the host vehicle passes the white line between the lanes only once. Therefore, if it is determined that the host vehicle has exceeded the white line at S38 even once in the risk calculation process, the driving data acquisition unit 11 determines that the white line has been passed when executing S38 next time. To do.

  The driving data acquisition unit 11 determines whether or not the lane change process has been completed (S40). When it is determined that the lane change has not been completed (S40: No), the operation data acquisition unit 11 continuously executes the processes after S34. When it is determined that the lane change is completed (S40: Yes), the driving data acquisition unit 11 sets the value of the lane change time 241 to “0” after the timing when the lane change is completed (S41).

[Driving risk calculation process 2 when changing lanes]
FIG. 10 is a continuation of the flowchart of the calculation process of the driving risk when changing lanes.

  The driving risk calculation unit 14 calculates the risk based on the operation data table 23 calculated in the flowchart of the driving risk calculation process at the time of lane change in FIG.

  The driving risk calculation unit 14 calculates the lane change time 241 from when the lane change is started until the host vehicle exceeds the white line from the operation data table 23 (S51). The driving risk calculation unit 14 calculates the amount of change 235 in the inter-vehicle distance between the rear vehicle after the lane change and the host vehicle from the operation data table 23 (S52). The driving risk calculation unit 14 specifies the TTC of the rear vehicle at the time when the host vehicle is on the white line from the TTC data 236 (S53). Note that the time when the host vehicle is on the white line may be the time when the host vehicle is in the center of the white line or the time when the right end of the host vehicle is on the white line, and the reference is determined in advance. . The driving risk calculation unit 14 specifies the THW of the rear vehicle when the host vehicle is on the white line from the THW data 237 (S54).

  The driving risk calculation unit 14 calculates the driving risk of the driver from the data specified in S51 to S54 and a predetermined reference value (S55).

[Explanation of driving risk calculation method]
FIG. 11 is an explanatory diagram of an index of driving risk when changing lanes.

  Reference numerals 40 and 41 denote lanes. In both lanes 40 and 41, the traveling direction of the vehicle is from left to right. It is assumed that the speed of the vehicle traveling on the lane 41 is higher than the speed of the vehicle traveling on the lane 40. 42 is a white line located between lanes.

  43 is the own vehicle. 44, 45, and 46 are vehicles located in the periphery of the own vehicle. The own vehicle 43 is traveling in the lane 40 and is assumed to change the lane to the lane 41. At this time, the vehicle 44, which is one of the surrounding vehicles, is located particularly behind the host vehicle 43 and exists in the lane 41 to which the host vehicle 43 moves. In this case, the vehicle 44 is a vehicle 44 behind the host vehicle 43.

  t1, t2, and t3 indicate times. It is assumed that time elapses in the order of t1, t2, and t3. In addition, the own vehicle 43 shall start the operation | movement of the lane change from the lane 40 to the lane 41 at the time t1. T23 is the time during which the host vehicle 43 moves from time t1 to time t2. T23 indicates the time from the start of steering by the driver until the host vehicle 43 passes the white line.

  The own vehicle 43 exists on the white line 42 at time t2. THW is the THW of the rear vehicle 44 and the host vehicle 43 at time t2. TTC is the TTC of the rear vehicle 44 and the host vehicle 43 at the time t2.

  It is assumed that the own vehicle 43 has completed the lane change to the lane 41 at time t3. L35 is a change in the distance between the host vehicle 43 and the rear vehicle 44 that has changed from time t2 to time t3.

  Based on the above values, the driving risk level calculation unit 14 follows the method described in Non-Patent Document 1, for example, according to the data of each timing, as “risk level = −a * T23 + b * (1 / THW) + c * (1 / TTC) + d * L35 + e ”is input to a predetermined evaluation function.

In the above evaluation functions, a, b, c, d, and e are constants and are predefined values. The input value of the evaluation function is as follows. (1) Time from the start of steering to crossing the white line (T23). (2) THW of the rear vehicle. (3) TTC of the rear vehicle.
(4) Inter-vehicle distance change after lane change (L35). The above evaluation function has a high degree of danger when the degree of danger is large. The driving risk calculation unit 14 calculates the risk from the above evaluation function. Here, the evaluation function described in Non-Patent Document 1 was used, but another type of evaluation function using THW or TCC may be used, and the measurement amount that can be acquired by the surrounding vehicle data acquisition unit 12 is used. Another kind of evaluation function may be used.

[Obtaining Gaze Data]
FIG. 12 is a flowchart of line-of-sight data acquisition processing.

  The driver camera 35 acquires images such as the driver's face, eyeball, and iris (S61). FIG. 13 is an explanatory diagram for obtaining the direction of the driver's line of sight.

  The line-of-sight data acquisition unit 15 calculates the line-of-sight vector 150 indicating the driver's line-of-sight origin P and the direction of the line-of-sight based on the driver's face, eyeball, iris, and the like imaged by the driver camera 35, thereby driving the driver. The direction of the line of sight is acquired (S62). The line-of-sight data acquisition unit 15 calculates facial feature points based on images such as a face, eyeballs, and irises, and compares the feature points with driver feature values stored in advance. The line-of-sight data acquisition unit 15 extracts the orientation of the face based on the comparison result and the image of the face, eyeball, iris, and the like. The line-of-sight data acquisition unit 15 detects the center position of the driver's left eyeball 152L and right eyeball 152R as the line-of-sight origin P. The line-of-sight data acquisition unit 15 calculates the center position of the iris 153a, that is, the center position of the pupil 153b. The line-of-sight data acquisition unit 15 calculates the line-of-sight vector 150 based on the line-of-sight origin P and the center position of the pupil 153b.

  The line-of-sight vector 150 is defined by, for example, coordinates in a spatial coordinate system with an arbitrary center point O of the vehicle as the origin. Since the driver can change his / her head forward / backward / left / right and up / down, the line-of-sight data acquisition unit 15 determines the position of the line-of-sight origin P with respect to the center point O of the spatial coordinate system according to the position and orientation of the driver's head. change.

  The line-of-sight data acquisition unit 15 stores the calculated line-of-sight origin P and line-of-sight vector 150 in the line-of-sight data table 26.

  FIG. 14 shows the data structure of the line-of-sight data table 26. The line-of-sight data table 26 includes a frame number 261 indicating a predetermined time, a line-of-sight origin 262, and a line-of-sight vector 263. The line-of-sight origin 262 stores the driver's line-of-sight origin P calculated every predetermined time, and the line-of-sight vector 263 stores the driver's line-of-sight vector 263 calculated every predetermined time.

  Further, the line-of-sight data table 26 shows a room mirror confirmation flag 264, a right door mirror confirmation flag 265, a left door confirmation flag 267 indicating whether or not the driver has viewed the mirror, and indicates that the driver is looking at other than the mirror. The flag 268 includes a rear vehicle confirmation flag 268 indicating that the driver has visually recognized the rear vehicle via any mirror.

  The line-of-sight data acquisition unit 15 acquires the position of the door mirror and the position of the room mirror (S63). The line-of-sight data acquisition unit 15 acquires the angle of the mirror surface of the door mirror and the angle of the mirror surface of the room mirror (S64). It is assumed that the position of the door mirror and the angle of the mirror surface with respect to the center point O of the spatial coordinate system, the position of the room mirror and the angle of the mirror surface, the position of the driver camera and the imaging direction are set in advance.

  FIG. 15 is a diagram illustrating the relationship between the driver's line of sight and the mirror. The mirror confirmation detection unit 16 detects which direction the driver 156 was viewing based on the detected line-of-sight origin P and the line-of-sight vector 150 of the driver 156. If the direction of the line-of-sight vector 150 is either a room mirror, a left or right door mirror (hereinafter referred to as a door mirror 154), the mirror confirmation detection unit 16 is located behind or behind the host vehicle 43 via the door mirror 154. It is determined that he / she has been visually recognized. Since the door mirror 154 has a predetermined area, the mirror confirmation detection unit 16 determines that the driver has viewed the mirror when the extension line of the line-of-sight vector is included in the area.

  The mirror confirmation detection unit 16 determines the virtual line-of-sight origin P ′, which is the virtual line-of-sight origin, based on the angle between the surface of the door mirror 154 and the line-of-sight vector 150, the shape of the surface of the door mirror 154, and the distance between the door mirror 154 and the line-of-sight origin P. Calculate (S65). Further, the mirror confirmation detection unit 16 calculates a virtual line-of-sight direction 155 that is a virtual line-of-sight direction based on the angle between the surface of the door mirror 154 and the line-of-sight vector 150 (S66). The driver visually recognizes the tip of the direction of the virtual visual line direction 155 with reference to the virtual visual line origin P ′. The mirror line-of-sight origin P ′ and the virtual line-of-sight direction 155 are defined by a spatial coordinate system centered on an arbitrary center point O of the host vehicle 43.

  Next, the mirror confirmation detection unit 16 determines whether or not the driver can visually recognize the rear vehicle reflected in the mirror.

  FIG. 16 is a diagram illustrating the relationship between the driver's line of sight, the mirror, and the rear vehicle. The door mirror 154 is located on an extension line of the driver's 156 line-of-sight vector 150. The mirror confirmation detection unit 16 determines whether or not there is a rear vehicle or the like in the virtual direction and the range 157 that is ahead of the virtual visual line direction 155 with respect to the virtual visual line origin P ′. To do. The mirror confirmation detection unit 16 determines the virtual view through the mirror according to the distance between the virtual line-of-sight origin P ′ and the door mirror 154, the angle between the line-of-sight direction and the surface of the door mirror 154, the shape of the surface of the door mirror 154, and the area of the door mirror 154. The visual field range 157 is specified (S67).

  Since the surrounding vehicle data acquisition unit 12 acquires the distance of the surrounding vehicle and the angle with respect to the traveling direction of the host vehicle, the mirror confirmation detection unit 16 calculates position information of the surrounding vehicle with respect to the host vehicle (S68). The mirror confirmation detection unit 16 acquires information on a region where a surrounding vehicle is present. The position of the surrounding vehicle is, for example, a relative position when an arbitrary center point O of the vehicle is the origin. The area where the surrounding vehicle exists is information indicating the size of the surrounding vehicle.

  The mirror confirmation detection unit 16 executes a mirror confirmation processing flag storage process (S69).

  FIG. 17 is a flowchart of the mirror confirmation process flag storage process.

  Based on the driver's line-of-sight data acquired every predetermined time, the mirror confirmation detection unit 16 determines whether or not the driver has viewed the room mirror (RM) every predetermined time (S71). When the driver visually recognizes the room mirror (S71: Yes), the mirror confirmation detection unit 16 sets the room mirror confirmation flag 264 to “1” (S72). The mirror confirmation detection unit 16 determines whether or not a rear vehicle exists within the specified virtual visual field range (S73). When the rear vehicle exists (S73: Yes), the mirror confirmation detection unit 16 stores “1” in the rear vehicle confirmation flag 268 (S74).

  Based on the driver's line-of-sight data acquired every predetermined time, the mirror confirmation detection unit 16 determines whether or not the driver has viewed the right door mirror (RDM) every predetermined time (S75). When the driver visually recognizes the right door mirror (S75: Yes), the mirror confirmation detection unit 16 sets the right door mirror confirmation flag 265 to “1” (S76). The mirror confirmation detection unit 16 determines whether or not a rear vehicle exists within the specified virtual visual field range (S77). When the rear vehicle exists (S77: Yes), the mirror confirmation detection unit 16 stores “1” in the rear vehicle confirmation flag 268 (S78).

  Based on the driver's line-of-sight data acquired every predetermined time, the mirror confirmation detection unit 16 determines whether or not the driver has viewed the left door mirror (LDM) every predetermined time (S79). When the driver visually recognizes the left door mirror (S79: Yes), the mirror confirmation detection unit 16 sets the left door mirror confirmation flag 266 to “1” (S80). The mirror confirmation detection unit 16 determines whether or not a rear vehicle exists within the specified virtual visual field range (S81). When the rear vehicle exists (S81: Yes), the mirror confirmation detection unit 16 stores “1” in the rear vehicle confirmation flag 268 (S82).

  The mirror confirmation detection unit 16 determines and stores whether or not the driver has viewed the rear vehicle through the mirror at a predetermined time.

[Safety confirmation calculation process]
FIG. 18 is an example of a desirable form of the visual confirmation action required in the preparation stage when changing lanes. In order to change the lane safely, the driver drives the vehicle while checking the safety behind the vehicle through the mirror when changing the lane. A series of desirable mirror confirmation operations at the time of lane change is, for example, as follows. The driver decides to change lanes. Thereafter, the driver confirms the rear at least twice through the rearview mirror and once through the door mirror. Thereafter, the driver turns on the blinker 3 seconds after the driver decides to change the lane. Thereafter, the driver confirms the rear once through the rearview mirror and once through the door mirror. Thereafter, after 3 seconds or more have elapsed since the driver turned on the turn signal, the driver starts steering.

  In FIG. 18, the horizontal axis represents time. RM0, RM1, and RM2 are timings when the driver should check the room mirror. DM0 and DM2 are timings when the driver should check the door mirror.

  When the driver changes his / her vehicle from the left lane to the right lane, each DM is the timing when the driver visually recognizes the right door mirror, and the driver changes his / her vehicle from the right lane to the left lane. In this case, each DM is the timing when the driver visually recognizes the left door mirror.

  The diagnosis start is the timing when the driver decides to start the lane change. The blinker lighting ON is the timing when the driver lights the blinker. The steering start SS is a timing when the driver starts steering. It takes 3 seconds from the diagnosis start timing to the blinker lighting On timing, and from the blinker lighting ON timing to the steering start SS timing is 3 seconds or more.

  In this embodiment, instead of recognizing that the driver decides to change the lane, the safety confirmation degree calculation unit 17 detects the start of the lane change by turning on the blinker. The safety confirmation degree calculation unit 17 sets the timing at which the driver decides to change the lane three seconds before the timing when the driver turns on the turn signal. In this embodiment, it is 3 seconds ago, but the time can be appropriately changed by setting.

The safety confirmation degree calculation unit 17 determines the driver's safety confirmation degree using the following evaluation criteria.
(1) Before the blinker lights up, the safety confirmation degree calculation unit 17 evaluates whether the driver has checked the room mirror or the door mirror and prepared to change lanes 3 seconds before the blinker lights up. For example, the evaluation formula is as follows.

WON-RM1 <3
WON-RM0 <3
WON-DM0 <3
An evaluation formula for evaluating whether or not the door mirror is confirmed immediately before the blinker is turned on is, for example, as follows. For example, it is evaluated whether or not the driver has seen the Dora mirror after the rearview mirror. The evaluation formula is as follows.

DM0> RM0
(2) The safety confirmation degree calculation unit 17 evaluates whether or not the driver visually recognizes the rearview mirror or the door mirror after the turn signal is turned on and before the steering is steered, so that the driver can confirm the final confirmation before the steering. Evaluate whether or not The evaluation formula is as follows.

RM2-SS <3
DM2-SS <3
The safety confirmation degree calculation unit 17 evaluates whether or not the driver has viewed the door mirror immediately before starting the steering. The evaluation formula is as follows.

DM2> RM2
(3) The safety confirmation degree calculation unit 17 evaluates whether or not the driver turns on the turn signal lighting by evaluating whether or not the driver turns on the turn signal at least 3 seconds before the start of steering. evaluate. The evaluation formula is as follows.

SS-WON> 3
[Safety confirmation calculation process]
FIG. 19 is a first flowchart of the safety confirmation degree calculation process.

  The safety confirmation degree calculation unit 17 calculates the safety confirmation degree by using the line-of-sight data table 26 and the operation data table 23 that store the timing of the mirror confirmation of the driver who is changing the lane every predetermined time.

  The safety confirmation degree calculation unit 17 acquires the line-of-sight data table 26 (S91). The safety confirmation degree calculation unit 17 determines whether or not the driver has turned on the blinker (S92). When it is determined that the turn signal is turned on (S92: Yes), the safety confirmation degree calculation unit 17 starts analysis processing for starting the lane change by turning on the turn signal (S95). When it is not determined that the turn signal is turned on (S92: No), the safety confirmation degree calculation unit 17 determines whether or not the steering angle is larger than a certain angle (S93). When the steering angle is larger than a certain angle (S93: Yes), the safety confirmation degree calculation unit 17 starts analysis processing for starting lane change by steering (S94). When the turn signal is not turned on, the driving operation may be other than lane change. Therefore, the safety confirmation degree calculation unit 17 can also execute the safety confirmation degree of the driving operation without turning on other turn signals such as the right curve and the left curve. When the steering angle is smaller than a certain angle (S93: No), the safety confirmation degree calculation unit 17 executes the processes after S92.

  The safety confirmation degree calculation unit 17 acquires line-of-sight data from a predetermined time before the turn-on timing of the blinker (S96). The safety confirmation degree calculation unit 17 extracts the timing when the driver looks at the mirror based on the line-of-sight data read in S96 (S97). The safety confirmation degree calculation unit 17 stores the extracted timing when the driver sees the mirror in the mirror confirmation data from the predetermined time before the blinker lights up until the blinker lights up (S98). The safety confirmation degree calculation unit 17 acquires line-of-sight data after the timing when the turn signal is turned on (S99). The safety confirmation degree calculation unit 17 extracts the timing when the driver looks at the mirror based on the line-of-sight data acquired in S99 (S100). The safety confirmation degree calculation unit 17 stores the extracted timing when the driver sees the mirror in the mirror confirmation data after the blinker is turned on (S101).

  The safety confirmation degree calculation unit 17 determines whether or not the driver has started steering from the lane change history data (S102). When it is not determined whether or not the driver has started steering (S102: No), the safety confirmation degree calculation unit 17 executes the processes after S99. When it is determined whether or not the driver has started steering (S102: Yes), the safety confirmation degree calculation unit 17 stores the steering start timing in the mirror confirmation data (S103).

  The safety confirmation degree calculation unit 17 acquires line-of-sight data after the start of steering (S104). The safety confirmation degree calculation unit 17 extracts the timing when the driver looks at the mirror based on the line-of-sight data acquired in S104 (S105). The safety confirmation degree calculation unit 17 stores the extracted timing when the driver sees the mirror in the mirror confirmation data after the start of steering (S106). The safety confirmation degree calculation unit 17 determines whether or not the lane change is completed from the lane change history data (S107). When it is determined that the lane change has not ended (S107: No), the safety confirmation degree calculation unit 17 executes the processes after S104. When it is determined that the lane change has ended (S107: Yes), the safety confirmation degree calculation unit 17 ends the process.

  FIG. 20 is a second flowchart of the safety confirmation degree calculation process.

  In the present embodiment, the safety confirmation degree calculation unit 17 performs the visual confirmation behavior of the driver during the time from when the lane change is determined until the turn signal is turned on, until the steering is started after the turn signal is turned on. And a desired operation including the visual confirmation behavior of the driver at the time is defined and stored in advance. The safety confirmation degree calculation unit 17 determines the safety confirmation degree according to the degree to which the driver is driving along the desired action defined in FIG.

  The safety confirmation degree calculation unit 17 stores in advance the desired safety confirmation behavior defined in FIG. Specifically, the safety confirmation action is as follows.

・ Check the room mirror at least twice in 3 seconds until the turn signal is turned on after deciding the start of lane change ・ At least once in 3 seconds after turning on the turn signal after determining the start of lane change Confirmation of the door mirror ・ Confirmation of the door mirror immediately before turning on the blinker ・ Confirm at least one room mirror between turning on the blinker and starting steering ・ More than 3 seconds from turning on the blinker to starting steering Elapse of time-Confirmation of door mirror immediately before starting steering The safety confirmation degree calculation unit 17 reads the line-of-sight data table 26 and the operation data table 23 (S111).

  The safety confirmation degree calculation unit 17 determines whether or not the steering start timing has passed 3 seconds or more since the turn-on of the blinker (S112). For example, the safety confirmation degree calculation unit 17 determines whether or not 3 seconds or more have elapsed depending on the number of frames from when the steering start flag 239 of the operation data table 23 becomes “1” until the turn signal lighting flag 238 becomes “1”. Is determined. When the time has elapsed (S112: Yes), the safety confirmation degree calculation unit 17 adds “1” to the safety confirmation degree (S113).

  The safety confirmation degree calculation unit 17 determines whether the driver has viewed the room mirror twice or more from the time when the blinker is turned on until 3 seconds before the blinker is turned on (S114). For example, the safety confirmation degree calculation unit 17 extracts frame numbers from when the steering start flag 239 of the operation data table 23 becomes “1” until the winker lighting flag 238 becomes “1”. The safety confirmation degree calculation unit 17 extracts the frame number 261 of the corresponding line-of-sight data table 26. The safety confirmation degree calculation unit 17 extracts the room mirror confirmation flag 264 in the frame number record of the corresponding line-of-sight data table 26 extracted. The safety confirmation level calculation unit 17 extracts a set in which the value of the room mirror confirmation flag 264 is “1”. The safety confirmation degree calculation unit 17 determines whether or not the extracted set is two or more. When watching twice or more (S114: Yes), the safety confirmation degree calculation part 17 adds "2" point to a safety confirmation degree (S115). When not seeing twice or more (S114: No), the safety confirmation degree calculation part 17 discriminate | determines whether the driver | operator saw the door mirror once from the time of turning on the winker from 3 seconds before turning on the winker (S116). ). When viewing once (S116: Yes), the safety confirmation degree calculation unit 17 adds “1” point to the safety confirmation degree (S117).

  The safety confirmation degree calculation unit 17 determines whether the driver has seen the door mirror at least once from the time when the blinker is turned on to the time when the blinker is turned on (S118). If it is viewed once or more (S118: Yes), the safety confirmation degree calculation unit 17 adds “1” point to the safety confirmation degree (S119).

  In order to determine whether the driver has viewed the door mirror one or more times immediately before the blinker is turned on, the safety confirmation degree calculation unit 17 determines whether the last confirmed mirror before the blinker is turned on is a door mirror (S120). ). For example, the safety confirmation degree calculation unit 17 determines the frame number corresponding to the set of the room mirror confirmation flag 264 determined to have seen the room mirror and the right door mirror confirmation flag 265 determined to have seen the right door mirror. Is compared with the frame number corresponding to the set of “1”. In the present embodiment, the greater the frame number, the later the time, so the safety degree calculation unit 17 determines that this is an action that has been performed later on a set with a larger frame number. When it is a door mirror (S120: Yes), the safety confirmation degree calculation part 17 adds "1" point to a safety confirmation degree (S121).

  The safety confirmation degree calculation unit 17 determines whether the driver has seen the rearview mirror from the time when the turn signal is turned on to the time when the steering is started (S122). When the driver looks at the room mirror (S122: Yes), the safety confirmation degree calculation unit 17 adds “1” point to the safety confirmation degree (S123).

The safety confirmation degree calculation unit 17 determines whether the driver has seen the door mirror from the time when the turn signal is turned on to the time when the steering is started (S124).
When the driver looks at the door mirror (S124: Yes), the safety confirmation degree calculation unit 17 adds “1” point to the safety confirmation degree (S125).

  In order to determine whether the driver has seen the door mirror one or more times immediately before the start of steering, the safety confirmation degree calculation unit 17 determines whether or not the last mirror that the driver has checked before turning on the blinker is a door mirror (S126). ). When it is a door mirror (S126: Yes), the safety confirmation degree calculation part 17 adds "1" point to a safety confirmation degree (S127).

  The safety confirmation degree calculation unit 17 calculates the safety confirmation degree based on the above determination.

  Note that the safety confirmation degree calculation unit 17 can determine not only whether or not the mirror is seen, but also whether or not the rear vehicle is reflected on the mirror. Further, it is also possible to determine whether or not a rear vehicle is reflected on the mirror separately from whether or not the mirror is viewed.

[Classification of drivers]
FIG. 21 shows an example of driver classification results. The driver classification unit 18 classifies the driver into four types by using the obtained safety confirmation index and risk index, and analyzes the characteristics of the driver. FIG. 21 shows the result of having a plurality of subjects (driver A and driver B) change lanes on the drive simulator and measure the line of sight in synchronization with driving operation data. FIG. 21A shows the result of the simulation performed by the driver A, and FIG. 21B shows the result of the simulation performed by the driver B.

The simulation which obtained the classification result of FIG. 21 is the following conditions. The road is a two-lane highway including a passing lane (right lane) and a traveling lane (left lane). In the overtaking lane, there are trains of surrounding vehicles at arbitrary intervals between 30 m and 110 m. The speed difference between the right lane and the left lane is 10 to 20 km / h. The driver travels while changing lanes. The above-described simulation was performed, the safety confirmation degree calculation unit 17 calculated the safety confirmation degree, and the driving risk calculation unit 14 calculated the lane change risk degree.
The horizontal axis of the two-dimensional coordinate in FIG. The risk level is in the range of -3 to 5 points, and the safety confirmation level is in the range of 0 to 8 points. Each point drawn on the two-dimensional coordinates indicates the correlation between the safety confirmation degree calculated when the driver is driving and the risk degree.

  The driver classification unit 18 classifies the driving characteristics of the driver into four types according to the positions of the points on the two-dimensional coordinates. In the correlation diagram of FIG. 21, (1) the upper right region of the correlation diagram is a region with a high degree of safety confirmation and a low degree of dangerous driving. (2) The upper left area of the correlation diagram is an area with a high degree of safety confirmation and a high degree of dangerous driving. (3) The lower right area of the correlation diagram is an area with a low degree of safety confirmation and a low degree of dangerous driving. (4) The lower left area of the correlation diagram is an area with a low degree of safety confirmation and a high degree of dangerous driving.

  It is expected that there is a causal relationship between the prior confirmation action for safe driving and the driving risk resulting from the confirmation. Therefore, among the driver trends classified into the four types, the driver trends belonging to (1) and (4) can be estimated from the measurement results of either the safety confirmation level or the dangerous driving level. is there. However, the tendency of the drivers belonging to (2) and (3) cannot be obtained from individual measurements of the safety confirmation degree and the dangerous driving degree. The driving technique determination device 1 according to the present embodiment evaluates the correlation between the measurement results of two causal relations, so that it is possible to obtain the driving behavior of the driver that is out of the desired driving behavior.

  In particular, in the trend of (3), there are drivers who have always been able to confirm safety but have not been able to confirm safety accidentally (oversight due to carelessness), and drivers who have not intentionally performed safety confirmation actions. May be mixed. In order to classify these, the driver classifying unit 18 measures the statistics of the safe driving diagnosis result, thereby making it possible to classify drivers that have been inadvertently overlooked and drivers who have not thoroughly confirmed safety.

  For example, the driver classification unit 18 acquires a statistical result of the driver's diagnosis result for a driver whose safety driving diagnosis result is classified as having a low driver's safety confirmation level but driving operation is safe. The driver classification unit 18 classifies the driver as accidental incomplete safety confirmation when the diagnosis result is far from the statistical average obtained. In addition, the driver classification unit 18 classifies that the driver is inadequate in the safety check when the diagnosis result is not far from the acquired statistical average.

  The driver classifying unit 18 classifies the driver by combining the safety confirmation index and the risk index, thereby making it possible to classify the driving characteristics of the driver that could not be classified by individual measurement of the safety level and the risk level. To identify trends.

[Education for drivers]
The driving guidance presentation unit 19 presents an education policy to the driver according to the tendency of the driver classified by the driver classification unit 18.

  Drivers belonging to (1) have confirmed safety and have few dangerous driving. Driver characteristics are positioned as model drivers. The driving instruction presentation unit 19 instructs such a driver to continue the current safe driving practice.

  The driver belonging to (2) has confirmed safety, but has become unsafe. The characteristic of the driver is positioned as an inattentive driver who cannot recognize the danger only by the confirmation action. The driving guidance presentation unit 19 provides driving guidance and education for such a driver to enhance the recognition of dangerous goods. The driving guidance presentation unit 19 provides guidance and education on driving techniques that do not fall into a dangerous state thereafter.

  The driver belonging to (3) has not been able to confirm safety, but driving is not dangerous. The driver's characteristics are overconfident and part-time driving. In this case, it is only a coincidence that it is not in a dangerous state, and the current parting operation has been effective in the cases so far, but there is a high risk of causing an accident in the unknown cases. The driving guidance presentation unit 19 makes the danger of part-time driving known, recognizes the necessity of safety confirmation behavior, and provides education for implementing this.

  Drivers belonging to (4) are not driving safely and are driving dangerously. The characteristics of the driver are positioned for beginners who cannot grasp the danger. The driving guidance presentation unit 19 first provides education for implementing basic safety confirmation behavior. Thereafter, the driving guidance presentation unit 19 performs driving guidance and education that enhances the recognition of dangerous goods. After that, provide guidance and education on driving techniques that will not cause danger. These educational policies are stored in the storage unit 38 in advance.

  The driving guidance presentation unit 19 can improve the safety of the driving behavior of the driver.

  In particular, a driver having a tendency as in (2) or (3) has been instructing a driver using the teaching know-how accumulated by a safe driving instructor. It takes time and effort for a safe driving instructor to give guidance. Since the driving guidance presentation unit 19 can present an appropriate educational policy according to the driver's tendency, it is possible to greatly reduce the number of steps for safe driving guidance.

Further, the driving skill determination device 1 can be mounted on a normal vehicle.
The driving skill determination device 1 accumulates the daily driving behavior results of the driver, thereby acquiring a long-term tendency of the driver and presenting safe driving guidance suitable for the tendency of the driver. .

  It is also possible to calculate the insurance premium when renewing the car insurance by accumulating the driving behavior results of the driver. The driving skill determination device 1 accumulates the daily driving behavior results of the driver in the storage unit 38. At the time of insurance renewal, the tendency of the driver is specified based on the accumulated data. It is assumed that there is a table in which driver tendencies and insurance premiums are associated in advance. The insurance company sets the insurance premium according to the read driver tendency. Thus, for example, the insurance premium can be changed according to the degree of safe driving.

  Further, the driving skill determination device 1 can be attached to a truck of a carrier to acquire a tendency of the driver of the truck and reflect it in the driver's performance and the like.

  With the above processing, the driving skill determination device 1 can perform tests on a daily basis by combining the driving operation and the driver's line of sight. As a result, the driving skill discriminating apparatus 1 can perform safe driving education and guidance without requiring a special environment such as a regular class for vehicle workers, or an instructor with specialized knowledge. Become. In addition, the driving skill determination device 1 can appropriately provide training for correcting driving for the driver based on the index.

Driving technology discrimination device 1
Operation data acquisition unit 11
Peripheral vehicle data acquisition unit 12
Peripheral vehicle speed and acceleration calculation unit 13
Driving risk calculator 14
Line-of-sight data acquisition unit 15
Mirror confirmation detector 16
Safety confirmation degree calculation unit 17
Driver classification unit 18
Driving guidance presentation unit 19
Operation data table 23
Line-of-sight data table 26
CPU31
RAM32
ROM33
Monitor 34
Driver camera 35
Radar 36
Input unit 37
Storage unit 38
Vehicle camera 39

Claims (6)

A line-of-sight data acquisition unit for acquiring information on the line of sight of the driver driving the vehicle;
A driving data acquisition unit for acquiring information on the driving operation of the driver and information on the behavior of the vehicle;
A surrounding vehicle data acquisition unit for acquiring information on a distance from the vehicle to another vehicle;
A history storage unit for accumulating each of the acquired information on the driver's line of sight, information on the driving operation, information on the behavior of the vehicle, and information on the distance to the other vehicle;
A driving operation storage unit for storing driving operations according to information on vehicle behavior and information on driving operations;
A safety information storage unit that stores information on the vehicle, a safe distance relationship with other vehicles, and movement of the line of sight with respect to driving operation;
A driving operation history generation unit that generates a driving operation history from the information related to the driving operation and the information related to the behavior of the vehicle, which is stored in the history storage unit with reference to the driving operation storage unit;
A driver classification that refers to the safety information storage unit and extracts a driving operation that is dangerous and has a correct line-of-sight movement with respect to the driving operation based on the distance from the surrounding vehicle in the generated driving operation history And
A driving skill determination device characterized by comprising:   A history of the driving operation generated by the driving operation history generation unit is stored, and a driver who is determined to be safe driving or not confirming safety by analyzing a driver's tendency or confirming safety is confirmed. 2. The driving skill discriminating apparatus according to claim 1, wherein a driver who is judged to be a dangerous driving is extracted.   The driving skill determination device according to claim 1, further comprising a safety confirmation degree calculation unit that detects that the direction of the driver's line of sight is a room mirror or a door mirror.   4. The driving skill determination according to claim 3, wherein the safety confirmation degree calculation unit sets a direction to be seen for each operation and determines whether or not a direction to be seen for each action is seen. apparatus.   The driving operation is information storing a series of driving operations at the time of lane change, left turn, and right turn, and includes information on the steering angle amount of the steering of the vehicle, the presence / absence of turn signal lighting, and the presence / absence of passage of a white line. The driving skill discrimination device according to claim 1. Information on the line of sight of the driver driving the vehicle, information on the driving operation of the driver and information on the behavior of the vehicle, information on the distance from the vehicle to other vehicles, information on the behavior of the vehicle, and information on the driving operation A computer having a storage unit that stores driving operation information according to the vehicle, the vehicle with respect to the driving operation, a safe distance relationship with other vehicles, information on line-of-sight movement, and information on the distance to other vehicles,
From the information related to the driving operation instruction stored in the storage unit and the information related to the behavior of the vehicle, a history of driving operations is generated,
With reference to the vehicle, the safe distance relationship with the other vehicle, and the movement of the line of sight with respect to the driving operation stored in the storage unit, among the generated driving operation history, A driving technique determination program that executes a process of extracting a driving operation that is dangerous from the distance of the vehicle and that has a correct line-of-sight movement with respect to the driving operation.