JP7223275B2 - Learning method, driving assistance method, learning program, driving assistance program, learning device, driving assistance system and learning system - Google Patents

Description

The present invention relates to a learning method, a driving assistance method, a learning program, a driving assistance program, a learning device, a driving assistance system, and a learning system.

A device has been proposed that controls the provision of information to a driver according to the cognitive load of the driver who is driving a vehicle (see, for example, Patent Document 1).

JP 2010-33549 A

High cognitive load is known to increase the probability of a driver causing an accident. However, the prior art does not consider criteria for cognitive load that lead to dangerous driving or conditions that lead to accidents.

In one aspect, the present invention aims to obtain a criterion of cognitive load that can determine the possibility of dangerous driving.

In one aspect, in the learning method, the learning unit determines whether dangerous driving is detected in driving diagnosis based on driving data obtained by chronologically recording information about driving of the vehicle from time-series data of cognitive load obtained from the driver's physiological index. extracting a plurality of data between the determined time and a time a predetermined time before the time, learning the relationship between cognitive load and dangerous driving using the extracted plurality of data, Based on the learned relationship, a criterion for dangerous driving is defined with respect to the cognitive load, and the learning unit performs learning in which, among the plurality of data, data in which all the cognitive loads included in the data are equal to or less than a threshold value are not used for learning. The method.

It is possible to obtain a criterion of cognitive load that can determine the possibility of dangerous driving.

FIG. 1 is a block diagram showing the configuration of the driving support system according to the first embodiment. FIG. 2 is a diagram illustrating an example of a hardware configuration of an information processing apparatus; FIG. 3A is a diagram showing an example of the configuration of a cognitive load data DB, and FIG. 3B is a diagram showing an example of cognitive load data. 4A is a diagram showing an example of the configuration of the operation data DB, FIG. 4B is a diagram showing an example of diagnostic data, and FIG. 4C is a diagram showing the configuration of the diagnostic data DB. It is a figure which shows an example. FIG. 5 is a diagram illustrating an example of a hardware configuration of a driving support server; FIG. 6 is a diagram showing an example of the configuration of a teacher data DB. FIG. 7 is a flowchart showing an example of learning processing. 8A and 8B are diagrams (part 1) for explaining the learning process of FIG. 7. FIG. 9 is a diagram (part 2) for explaining the learning process of FIG. 7. FIG. FIG. 10A is a diagram showing an example of the configuration of the learning data DB, FIG. 10B is a diagram showing an example of the configuration of the reference data DB provided in the driving assistance server, and FIG. 4 is a diagram showing an example of a configuration of a reference data DB included in an information processing apparatus; FIG. FIG. 11 is a flowchart showing an example of dangerous driving prediction processing. FIG. 12 is a block diagram showing the configuration of the driving support system according to the second embodiment. FIG. 13 is a block diagram showing the configuration of the driving support system according to the third embodiment. FIG. 14 is a block diagram showing another example of the configuration of the driving support system according to the third embodiment. FIG. 15 is a flowchart illustrating an example of prediction accuracy determination processing.

A driving support system according to the first embodiment will be described in detail below with reference to FIGS. 1 to 11. FIG. A driving assistance system is a system that assists a driver's driving by predicting dangerous driving based on the driver's cognitive load.

FIG. 1 schematically shows the configuration of a driving assistance system 100 according to the first embodiment. As shown in FIG. 1 , the driving assistance system 100 includes a driving assistance device 10 and a driving assistance server 20 . In the first embodiment, the driving assistance device 10 is an example of a first information processing device, and the driving assistance server 20 is an example of a second information processing device. Also, in the first embodiment, the driving support server 20 is an example of a learning device. Also, the driving support system 100 is an example of a learning system.

The driving assistance device 10 is mounted on a vehicle (not shown) and includes a notification device 11 , a physiological index measurement device 12 , a travel information acquisition device 13 , a driver information acquisition device 14 and an information processing device 30 .

Examples of the notification device 11 include a display and a speaker, but are not limited to these. The display and the speaker may be a display and speaker pre-installed in the vehicle, respectively.

The physiological index measuring device 12 is a device that measures the physiological index of a driver who drives a vehicle. Examples of physiological indicators include, but are not limited to, heart rate, eye movement, pupil diameter, cerebral blood flow, perspiration, and facial skin temperature. From the physiological index, the cognitive load of the driver can be obtained. When the driver's cognitive load is high, heart rate increases, saccade-type intrusion increases in eye movements, pupil diameter increases, cerebral blood flow increases, perspiration increases, and facial skin temperature rises. . If the physiological index is heart rate, the physiological index measuring device 12 is a heart rate monitor, and if it is eye movement or pupil diameter, the physiological index measuring device 12 is a camera. Further, for example, if the physiological index is cerebral blood flow, the physiological index measuring device 12 is an optical topography device, and if it is the amount of perspiration, the physiological index measuring device 12 is a perspiration meter, and if it is the facial skin temperature, The physiological index measuring device 12 is a thermography. The physiological index measuring device 12 measures at least one physiological index and transmits the measurement result to the information processing device 30 .

The travel information acquisition device 13 is a device that acquires information about travel of the vehicle. Examples of information related to vehicle travel include, but are not limited to, engine speed, vehicle speed, acceleration, angular velocity, steering angle, accelerator pedal operation amount, brake pedal operation amount, and images around the vehicle. It is not something that can be done. When the information about the running of the vehicle is the engine speed, the running information acquisition device 13 is a tachometer, when it is the vehicle speed, it is a vehicle speed sensor, when it is the acceleration, it is an acceleration sensor, and when it is the angular velocity, it is a gyro sensor. and if it is the steering angle, it is the steering wheel angle sensor. Further, when the information regarding the travel of the vehicle is the operation amount of the accelerator pedal, the travel information acquisition device 13 is an accelerator pedal sensor, and when it is the operation amount of the brake pedal, it is a brake pedal sensor, and when it is an image around the vehicle. A drive recorder. The travel information acquisition device 13 acquires at least one piece of information about travel of the vehicle, and transmits the information to the information processing device 30 .

The driver information acquisition device 14 is a device that acquires driver identification information, and is, for example, a card reader. For example, when the driver holds the driver's license over the driver information acquisition device 14, the driver information acquisition device 14 acquires the driver's license number from the IC chip embedded in the license as identification information of the driver. The driver information acquisition device 14 transmits the acquired identification information of the driver to the information processing device 30 .

The information processing device 30 executes various processes described later based on the information acquired from the physiological index measuring device 12, the travel information acquiring device 13, and the driver information acquiring device 14. FIG. The information processing device 30 is connected to the driving support server 20 via a network NW such as the Internet.

The information processing device 30 has a hardware configuration as shown in FIG. Specifically, as shown in FIG. 2, the information processing device 30 includes a Central Processing Unit (CPU) 311, a Read Only Memory (ROM) 312, a Random Access Memory (RAM) 313, a storage device (Hard Disk Drive: HDD ) 314, a network interface 315, a portable storage medium drive 317 capable of reading data stored in a portable storage medium 316, and the like. Each component of the information processing device 30 is connected to the bus 320 . The CPU 311 executes a program stored in the ROM 312 or HDD 314 or a program read from the portable storage medium 316 by the portable storage medium drive 317, thereby causing the information processing apparatus 30 to function as each unit in FIG.

Specifically, as shown in FIG. 1, the CPU 311 executes a program, whereby the information processing device 30 includes an acquisition unit 31, a diagnosis unit 32, a communication unit 33, a first determination unit, a notification unit, and a first 2 functions as a control unit 34 as a judgment unit.

The acquisition unit 31 acquires the physiological index of the driver from the physiological index measuring device 12, and associates time-series data of cognitive load (hereinafter referred to as cognitive load data) obtained from the physiological index with identification information of the driver. Stored in the cognitive load data DB 351 . The cognitive load data may be, for example, the physiological index itself measured by the physiological index measuring device 12, or the result obtained by performing predetermined processing on the physiological index (for example, a moving average of the physiological index, etc.). There may be. The acquisition unit 31 also outputs the physiological index acquired from the physiological index measuring device 12 to the control unit 34 .

The cognitive load data DB 351 includes fields such as a driver ID, cognitive load data, and a transmitted flag, as shown in FIG. 3(A). The driver ID field stores the identification information of the driver (for example, the driver's license number) acquired by the driver information acquisition device 14 . The cognitive load data field stores time-series cognitive load data obtained during one trip, as shown in FIG. 3(B). In this embodiment, one trip is defined as one ON period of the ignition switch.

The transmitted flag is a flag indicating whether or not the cognitive load data has been transmitted to the driving assistance server 20 , and is checked when the cognitive load data has been transmitted to the driving assistance server 20 . In this way, the cognitive load data DB 351 stores cognitive load data for each driver. Note that the cognitive load data DB 351 may include, for example, fields for storing the acquisition start time and acquisition end time of the cognitive load data.

In addition, the acquisition unit 31 stores, in the operation data DB 352 , the driving data obtained by the driving information obtaining device 13 and recording the information about the driving of the vehicle in chronological order.

The driving data DB 352 includes fields such as a driver ID, driving data, and a diagnosed flag, as shown in FIG. 4(A). The driver ID field stores the identification information of the driver (for example, the driver's license number) acquired by the driver information acquisition device 14 . The driving data field stores driving data acquired during one trip. The diagnosed flag is a flag indicating whether or not a driving diagnosis, which will be described later, has been performed based on the driving data, and is checked when the driving diagnosis has been performed based on the driving data. Thus, the driving data DB 352 stores driving data for each driver. Note that the driving data DB 352 may include fields for storing the acquisition start time and the acquisition end time of the driving data, for example.

Returning to FIG. 1 , when the trip ends, the diagnosis unit 32 diagnoses the driving of the driver based on the driving data stored in the driving data DB 352 and records the driving diagnosis result in the diagnostic data DB 353 . After completing the driving diagnosis, the diagnosis unit 32 checks the diagnosis completed flag of the driving data for which the driving diagnosis has been performed in the driving data DB 352 . In this embodiment, the diagnosis unit 32 diagnoses a risky driving degree indicating the degree of risky driving based on the driving data. The diagnosis unit 32 diagnoses that the degree of dangerous driving is high, for example, when the rate of change in vehicle speed over time exceeds a predetermined threshold value (that is, when sudden acceleration or sudden deceleration occurs). Further, for example, the diagnosis unit 32 determines that the degree of dangerous driving is high when the rate of change in the steering angle over time exceeds a predetermined threshold value (that is, when the driver turns the steering wheel sharply). The diagnosis unit 32 may diagnose the degree of dangerous driving based on a plurality of pieces of information (eg, acceleration and images recorded by a drive recorder). As a result, in this embodiment, time-series data of the degree of dangerous driving as shown in FIG. 4B is acquired as diagnostic data.

The diagnostic data DB 353 includes fields such as a driver ID, diagnostic data, and a transmitted flag, as shown in FIG. 4(C). The driver ID field stores the driver ID included in the driving data used for the driving diagnosis. The diagnostic data field stores the time-series data of the degree of dangerous driving shown in FIG. 4(B). The transmitted flag is a flag indicating whether or not the diagnostic data has been transmitted to the driving support server 20, and is checked when the diagnostic data has been transmitted to the driving support server 20. FIG. Thus, the diagnostic data DB 353 stores diagnostic data for each driver.

Returning to FIG. 1, when the trip ends, the communication unit 33 transmits the cognitive load data stored in the cognitive load data DB 351 and the diagnostic data stored in the diagnostic data DB 353 together with the vehicle information 355 to the driving support server 20. do. After completing the transmission, the communication unit 33 checks the transmitted flags of the cognitive load data and diagnostic data transmitted to the driving support server 20 .

Vehicle information 355 is stored in HDD 314 or the like, and is information representing the type of vehicle. The information representing the vehicle type may be, for example, the vehicle name, model year, model, or a combination thereof, vehicle types such as ordinary passenger cars and small passenger cars, minivans, sedans, compacts, and SUVs (Sport Utility Vehicles). ) may be the body type of the vehicle. In this embodiment, it is assumed that the vehicle information 355 is the body type of the vehicle.

Further, at the start of the trip, the communication unit 33 transmits a reference data transmission request including the vehicle information 355 and the driver's identification information to the driving support server 20, and the driving support server 20 transmits the transmitted vehicle information 355 and driver's Obtaining (receiving) risky driving criteria for cognitive load defined for combination with identification information. The communication unit 33 stores the acquired criteria in the criteria data DB 354 .

The control unit 34 uses the driver's cognitive load obtained from the physiological index input from the acquisition unit 31 and the dangerous driving criteria for the cognitive load defined by the driving support server 20 described later. Determine the potential for dangerous driving. More specifically, the possibility of dangerous driving is determined based on whether or not the driver's cognitive load satisfies defined criteria for dangerous driving. When the controller 34 determines that there is a possibility of dangerous driving, it controls the notification device 11 to notify the driver of the possibility of dangerous driving. Note that the control unit 34 may acquire the physiological index directly from the physiological index measuring device 12 . Details of the processing executed by the control unit 34 will be described later.

The driving support server 20 is, for example, a cloud server, and defines the criteria for dangerous driving with respect to cognitive load based on the cognitive load data and diagnostic data received from the information processing device 30 .

The driving assistance server 20 has a hardware configuration as shown in FIG. Specifically, as shown in FIG. 5, the driving support server 20 includes a CPU 211, a ROM 212, a RAM 213, an HDD 214, a network interface 215, and a portable storage medium capable of reading data stored in a portable storage medium 216. A drive 217 and the like are provided. Each component of the driving support server 20 is connected to the bus 220 . The CPU 211 causes the driving support server 20 to function as each part in FIG.

Specifically, as shown in FIG. 1, the driving support server 20 functions as a communication unit 21 and a learning unit 22 as a learning unit and a definition unit by the CPU 211 executing a program.

The communication unit 21 receives the cognitive load data and diagnostic data together with the vehicle information 355 from the information processing device 30, and stores the received pairs of cognitive load data and diagnostic data as teacher data in the teacher data DB 251. FIG. Since the communication unit 33 of the information processing device 30 transmits cognitive load data and diagnostic data when the trip ends, one teacher data can be obtained for each trip.

For example, as shown in FIG. 6, the teacher data DB 251 includes fields such as vehicle type, driver ID, cognitive load data, diagnosis data, and learned flag. The vehicle type field stores the type of vehicle included in the vehicle information 355 received from the information processing device 30 . The driver ID field stores information on the driver ID included in the received cognitive load data or diagnostic data. The received cognitive load data and diagnostic data are stored in the cognitive load data and diagnostic data fields, respectively. The learned flag is a flag indicating whether learning of the relationship between cognitive load and dangerous driving using cognitive load data and diagnostic data has ended, and is checked when learning ends.

In addition, the communication unit 21 transmits to the information processing device 30 the dangerous driving criteria for the cognitive load defined by the learning unit 22 .

The learning unit 22 learns the relationship between cognitive load and dangerous driving based on teacher data. In addition, the learning unit 22 defines the criteria for dangerous driving with respect to cognitive load based on the learned relationships.

(learning process)
Next, learning processing executed by the learning unit 22 of the driving support server 20 will be described based on the flowchart of FIG. 7 . The process of FIG. 7 may be executed each time cognitive load data and diagnostic data are received from the information processing device 30, or may be executed when a predetermined number of teacher data for a combination of a certain vehicle type and driver ID is accumulated. You may In this embodiment, it is assumed that the driving support server 20 receives cognitive load data and diagnostic data from the information processing device 30 and executes the process of FIG. 7 .

In the process of FIG. 7, the learning unit 22 first reads teacher data stored in the teacher data DB 251 (step S11). More specifically, from the teacher data DB 251, the teacher for the same combination as the combination of the driver ID included in the cognitive load data and the diagnostic data and the vehicle type included in the vehicle information 355 received together with the cognitive load data and the diagnostic data. Among the data, read the data whose learned flag is not checked.

The teacher data are, for example, cognitive load data (shown in the lower part) and diagnostic data (shown in the upper part) collected during one trip as shown in FIG. 8(A).

Next, from the cognitive load data, the learning unit 22 extracts the data between the time when the dangerous driving was determined and the time a predetermined time (T) before that time, and the relationship between the cognitive load and the dangerous driving. is extracted as learning data for learning (step S13). The learning unit 22 determines dangerous driving when the dangerous driving degree of the diagnostic data exceeds a predetermined threshold. For example, as shown in FIG. 8B, when the dangerous driving degree exceeds the threshold at times t11 and t21, times t11 and t21 are determined as dangerous driving.

Then, as shown in FIG. 8(B), the learning unit 22 extracts the data between the time t11 and the time (t11-T) and the data between the time t21 and the time (t21-T) from the cognitive load data. Extract the data between and . Note that the learning unit 22 may extract data within a predetermined time (T) range before the time at which the dangerous driving is determined. For example, the learning unit 22 may extract data between time (t11-α) (α<T) before time t11 and time (t11-α-T). In other words, the learning data may or may not include cognitive load data at the time when dangerous driving is determined.

Returning to FIG. 7, next, the learning unit 22 excludes data whose cognitive load is equal to or less than a threshold value from among the plurality of extracted data (step S15). The threshold is set based on the normal cognitive load of the driver, which is measured in advance. For example, in the example of FIG. 9, since the cognitive load included in the data between time (t11-T) and time t11 is all below the threshold Vth, the data between time (t11-T) and time t11 are excluded from the training data.

Next, the learning unit 22 stores the remaining data in the learning data DB 253 as learning data for learning the relationship between cognitive load and dangerous driving (step S17). As a result, the data after the time when dangerous driving is determined (for example, in FIG. 9, the data from time t11 to time (t21-T), and the data after time t21) can be used to determine the relationship between cognitive load and dangerous driving. is not stored as training data for learning Therefore, the data when the cognitive load increases due to dangerous driving is not used as learning data. In addition, the learning unit 22 does not store, as learning data, data whose cognitive load is equal to or less than the threshold among the data between the time at which the dangerous driving is determined and the time before the time at which the driving is determined to be dangerous by a predetermined time. For this reason, data in which dangerous driving occurs due to an external factor (for example, an interruption by another vehicle, etc.) is excluded from the learning data. As a result, only the data when dangerous driving occurs due to cognitive load can be used as learning data.

The learning data DB 253 includes fields such as vehicle type, driver ID, and learning data, as shown in FIG. 10(A). The vehicle type and driver ID fields store the vehicle type and driver ID associated with the teacher data read in step S11, respectively. In the learning data field, among the cognitive load data, the data between the time when the dangerous driving was judged and the time before a predetermined time (T) from the time, and the cognitive load is not below the threshold. Stored.

Returning to FIG. 7, the learning unit 22 checks the learned flag of the teacher data from which the learning data is acquired in the teacher data DB 251 (step S18). Note that the learning unit 22 may delete teacher data from which learning data is acquired. In this case, the learned flag becomes unnecessary.

Next, the learning unit 22 defines a dangerous driving criterion for cognitive load based on the learning data stored in the learning data DB 253 (step S19). Specifically, the learning unit 22 reads the learning data for the combination of the vehicle type and the driver ID from the learning data DB 253, and based on the learning data, the cognitive load is calculated for the combination of the vehicle type and the driver ID. Define criteria for dangerous driving. For example, based on the learning data, the learning unit 22 defines, as a criterion for dangerous driving, the pattern of temporal change in cognitive load between the time when dangerous driving is determined and the time a predetermined time before that time. The learning unit 22 may define the standard deviation of the cognitive load between the time when dangerous driving is determined and the time a predetermined time before the time as the dangerous driving criterion.

The learning unit 22 registers the defined criteria in the criteria data DB 252 (step S21), and terminates the learning process.

As shown in FIG. 10B, the reference data DB 252 has fields for vehicle type, driver ID, and reference data. The vehicle type and driver ID fields store the vehicle type and driver ID used for selecting learning data in step S19, respectively. The reference data field stores the dangerous driving reference defined in step S19.

As described above, the communication unit 33 of the information processing device 30 transmits the cognitive load data and the diagnostic data to the driving support server 20 when the trip ends. increases. Therefore, by performing the processing of FIG. 7 each time the cognitive load data and diagnostic data are received, the learning accuracy (appropriateness of the criteria for dangerous driving) is improved.

For example, when the communication unit 21 receives a reference data transmission request including the vehicle information 355 and the driver identification information from the information processing device 30, the reference data DB 252 defined for the combination of the vehicle information 355 and the driver identification information. is transmitted to the information processing device 30 . As a result, reference data is stored in the reference data DB 354 of the information processing device 30 for each driver who drives a vehicle in which the information processing device 30 is mounted, as shown in FIG. 10(C).

(Dangerous driving prediction)
Next, the process of predicting dangerous driving of the driver executed by the information processing device 30 provided in the driving support device 10 mounted on the vehicle will be described. FIG. 11 is a flowchart showing dangerous driving prediction processing. The process of FIG. 11 is executed until the trip ends when the trip starts.

In FIG. 11, the control unit 34 of the information processing device 30 acquires the driver's cognitive load from the driver's physiological index input from the acquisition unit 31 (step S31).

Next, the control unit 34 acquires the reference data associated with the driver currently driving the vehicle from among the reference data stored in the reference data DB 354 (step S32).

Next, the control unit 34 determines the possibility of dangerous driving based on the cognitive load data and the criteria for dangerous driving included in the reference data acquired in step S32 (step S33). For example, when the degree of similarity between the change in cognitive load over time within a predetermined time (T) and the pattern of change in cognitive load over time defined as a criterion for dangerous driving is equal to or greater than a threshold, the control unit 34 Decide that you are likely to drive.

If there is no possibility of dangerous driving (step S35/NO), the control unit 34 returns to step S31. is notified to the driver (step S37). For example, the control unit 34 notifies the driver of the possibility of dangerous driving by using a buzzer sound. The control unit 34 may cause the notification device 11 to output a voice message such as "There is a possibility of dangerous driving." Further, the control unit 34 may display an image or the like indicating the possibility of dangerous driving on the display of the navigation device.

As described above in detail, according to the first embodiment, the driving support system 100 records information about the running of the vehicle in chronological order from the cognitive load chronological data obtained from the physiological index of the driver. Based on the driving diagnosis, multiple data are extracted between the time when the driving diagnosis was judged to be dangerous driving and the time a predetermined time before that time, and the relationship between cognitive load and dangerous driving is learned using the extracted multiple data. and a learning unit 22 that defines the criteria for dangerous driving with respect to cognitive load based on the learned relationships. The learning unit 22 does not store, in the learning data DB 253 , data in which all the cognitive loads included in the data are equal to or less than the threshold among the plurality of extracted data. That is, the learning unit 22 does not use data after the time at which dangerous driving is determined for learning, and does not use data whose cognitive load is below the threshold value among the plurality of extracted data for learning. . As a result, cases in which the cognitive load increases due to dangerous driving and cases in which dangerous driving occurs due to external factors (for example, cut-in by another vehicle) can be excluded from learning. It is possible to obtain a highly accurate reference that can be determined.

Further, in the first embodiment, the time series data of cognitive load (cognitive load data) and driving data are acquired for each combination of driver and vehicle type (see, for example, FIG. 6), and the learning unit 22 , for each combination of driver and vehicle type, define a dangerous driving criterion. As a result, it is possible to obtain a more accurate reference that reflects the influence of the type of vehicle on driving and individual differences.

In addition, in the first embodiment, the driving support system 100 determines the possibility of dangerous driving based on the driver's cognitive load obtained from the physiological index input from the physiological index measuring device 12 and the criteria for dangerous driving. and a control unit 34 that performs notification based on the determination result. Since the likelihood of dangerous driving can be determined using highly accurate criteria for cognitive load, the likelihood of dangerous driving can be determined with high accuracy. More specifically, the accuracy of the timing of alerting for dangerous driving is increased compared to the case where the alerting for dangerous driving is performed based on whether or not the physiological index exceeds the threshold value. In addition, by notifying the driver of the possibility of dangerous driving, the driver's safe driving can be supported.

Further, in the first embodiment, since the driving support server 20 includes the learning unit 22, it is easy to revise the learning algorithm appropriately, and highly accurate learning becomes possible.

In addition, in the first embodiment, the driving support system 100 includes a diagnostic unit 32 that performs driving diagnosis based on driving data. As a result, the amount of data to be transmitted to the driving support server 20 can be reduced compared to the case of transmitting driving data. Note that the driving data may be transmitted to the driving support server 20 instead of the driving data.

Note that in the process of FIG. 11 of the first embodiment, the control unit 34 controls the volume of the buzzer, the pattern of the sound, the sound The content of the message (mode of notification) may be changed. For example, when the possibility of dangerous driving is low, the possibility of dangerous driving may be notified by a voice message, and when the possibility of dangerous driving is high, the possibility of dangerous driving may be notified by a buzzer or the like. .

Further, in the first embodiment, the cognitive load data and driving data are acquired for each combination of driver and vehicle type (for example, see FIG. 6), but they may be acquired for each driver. , may be obtained for each type of vehicle. In this case, dangerous driving standards are defined for each driver or each type of vehicle.

Although the communication unit 33 of the information processing device 30 transmits diagnostic data to the driving support server 20 in the first embodiment, the present invention is not limited to this. For example, the communication unit 33 may transmit to the driving support server 20 the time at which the diagnostic data indicates dangerous driving (the time at which the degree of dangerous driving exceeds the threshold).

In the first embodiment, diagnostic data is transmitted to the driving support server 20, and the learning unit 22 of the driving support server 20 defines the criteria for dangerous driving. may

FIG. 12 is a block diagram showing the configuration of a driving assistance system 100A according to the second embodiment. As illustrated in FIG. 12 , an information processing apparatus 30A according to the second embodiment includes a learning unit 35. As illustrated in FIG. In the second embodiment, the information processing device 30A is an example of a learning device. Also, the information processing device 30A is an example of a first information processing device.

The learning unit 35 performs the same learning process as in FIG. Specifically, the learning unit 35 extracts learning data using the cognitive load data stored in the cognitive load data DB 351 and the diagnostic data stored in the diagnostic data DB 353 as teacher data, and stores the learning data in the learning data DB 356. do. Then, the learning unit 35 defines the dangerous driving criteria for the cognitive load based on the learning data stored in the learning data DB 356 .

Since other configurations are the same as those of the first embodiment, detailed description thereof will be omitted. In Example 2, the communication unit 33 can be omitted because it is not necessary to transmit the cognitive load data and diagnostic data to the driving support server 20 .

The driving assistance device 10 included in the driving assistance system may be realized by an electronic device such as a smart phone. FIG. 13 is a block diagram showing the configuration of a driving assistance system 100B according to the third embodiment. A driving assistance system 100B according to the third embodiment includes an electronic device 50 (driving assistance device 10). The electronic device 50 is, for example, a smartphone, and includes a notification device 11, a physiological index measurement device 12, a travel information acquisition device 13, and an information processing device 30A.

The notification device 11 is, for example, a speaker included in the electronic device 50 . The physiological index measurement device 12 is, for example, a camera provided in the electronic device 50. By attaching the electronic device 50 to the windshield, dashboard, or the like of the vehicle so as to photograph the eyeballs of the driver, the pupil diameter and/or eye movement can be measured. can function as the physiological index measuring device 12 that measures the

The travel information acquisition device 13 is, for example, an acceleration sensor and a gyro sensor built in the electronic device 50 .

Since other configurations are the same as those of the second embodiment, detailed description thereof will be omitted.

Note that, in the third embodiment, the communication unit 33 of the information processing device 30 of the electronic device 50A transmits cognitive load data and diagnostic data to the driving assistance server 20, as in the driving assistance system 100C shown in FIG. can be configured to At this time, the communication unit 33 transmits, for example, the phone number of the smartphone as the identification information of the driver. In addition, the communication unit 33 transmits, as the vehicle information 355, information on the type of vehicle input by the user in the application installed on the smartphone.

The driver information acquisition device 14 can be omitted by using electronic devices 50 and 50A owned by individuals such as smartphones.

In Examples 1 to 3, the control unit 34 may perform the dangerous driving prediction process described with reference to FIG. 11 after the prediction accuracy of dangerous driving reaches a predetermined accuracy. FIG. 15 is a flowchart illustrating an example of prediction accuracy determination processing. The process of FIG. 15 is executed when the vehicle finishes running.

First, the control unit 34 specifies the time (predicted dangerous driving time) at which it is determined that there is a possibility of dangerous driving when using the current reference data in the cognitive load data acquired on the current trip. (Step S51). In addition, the control unit 34 specifies the time (dangerous driving time) at which dangerous driving is determined in the diagnosis data of the current trip (step S53). The order of the processes in steps S51 and S53 may be reversed, or may be performed concurrently.

Next, the control unit 34 determines the validity of the reference data (step S55). For example, the control unit 34 determines the validity of the reference data, for example, based on the rate of coincidence between the predicted dangerous driving time and the dangerous driving time. The control unit 34 controls the rate of coincidence between the time when it is determined that there is a possibility of dangerous driving when using the current reference data and the time when the diagnostic data indicates that dangerous driving is determined to be a predetermined threshold value (for example, 70%). %) or more, the reference data is judged to be appropriate. If the reference data is not valid (step S55/NO), the process of FIG. 15 is terminated. If the reference data is valid (step S55/YES), the control unit 34 permits the start of the dangerous driving prediction process (step S57), and ends the process of FIG. For example, the control unit 34 turns ON a flag that permits the start of the dangerous driving prediction process, and if the flag is ON during the next run, the dangerous driving prediction process of FIG. 11 may be executed. As a result, it is possible to prevent the possibility of dangerous driving from being erroneously notified to the driver when the prediction accuracy of dangerous driving is low (when learning is not sufficient) using the current dangerous driving criteria.

Note that the processing functions described above can be realized by a computer. In that case, a program is provided that describes the processing contents of the functions that the processing device should have. By executing the program on a computer, the above processing functions are realized on the computer. A program describing the processing content can be recorded in a computer-readable recording medium (excluding carrier waves).

When a program is distributed, it is sold in the form of a portable storage medium such as a DVD (Digital Versatile Disc) or a CD-ROM (Compact Disc Read Only Memory) on which the program is recorded. It is also possible to store the program in the storage device of the server computer and transfer the program from the server computer to another computer via the network.

A computer that executes a program stores, for example, a program recorded on a portable storage medium or a program transferred from a server computer in its own storage device. The computer then reads the program from its own storage device and executes processing according to the program. The computer can also read the program directly from the portable storage medium and execute processing according to the program. In addition, the computer can also execute processing according to the received program every time the program is transferred from the server computer.

The embodiments described above are examples of preferred implementations of the present invention. However, the present invention is not limited to this, and various modifications can be made without departing from the spirit of the present invention.

10 driving assistance device 20 driving assistance server 22 learning unit 32 diagnosis unit 34 control unit 35 learning units 100, 100A, 100B, 100C driving assistance system

Claims (15)

The time at which the learning unit determines that dangerous driving is determined in driving diagnosis based on driving data in which information about driving of the vehicle is recorded in time series from time-series data of cognitive load obtained from the driver's physiological index, and from that time extracting a plurality of data between the time a predetermined time ago, using the extracted plurality of data, learning the relationship between cognitive load and dangerous driving,
The definition unit defines a criterion for dangerous driving with respect to cognitive load from the learned relationship,
wherein the learning unit does not use, among the plurality of data, data whose cognitive loads are all less than or equal to a threshold value for learning;
learning method. The time-series data of the cognitive load and the driving data are obtained for each driver, for each type of vehicle, or for each combination of types of the driver and the vehicle,
wherein the definition unit defines the criteria for each driver, each type of vehicle, or each combination of types of the driver and the vehicle;
A learning method according to claim 1 . the physiological index includes at least one of heart rate, eye movement, pupil diameter, cerebral blood flow, sweating state, and facial skin temperature;
3. The learning method according to claim 1 or 2. The information about the running of the vehicle includes at least one of engine speed, vehicle speed, acceleration, angular velocity, steering angle, accelerator pedal operation amount, brake pedal operation amount, and an image around the vehicle.
The learning method according to any one of claims 1 to 3. A first determination unit acquires a cognitive load from the physiological index measured by the physiological index measuring device, and the acquired cognitive load and the criterion defined by the learning method according to any one of claims 1 to 4. to determine the possibility of dangerous driving based on
The notification unit performs notification based on the judgment result,
Driving assistance method. The notification unit changes the mode of notification depending on whether the possibility of dangerous driving is high or low;
The driving support method according to claim 5. A result of a second judgment unit judging the possibility of dangerous driving using the criteria for the time-series data of the cognitive load acquired during one trip, and the driving data acquired during the one trip. Judging the validity of the criteria based on comparison with the results of driving diagnosis based on
The driving assistance method according to claim 5 or 6. The notification unit starts the notification based on the judgment result when it is determined that the criterion is appropriate.
The driving support method according to claim 7. A diagnosis unit performs driving diagnosis based on the driving data;
The driving support method according to any one of claims 5 to 8. to the computer,
Based on the time-series data of cognitive load obtained from the driver's physiological index, the time at which the driving diagnosis is based on the driving data recorded in time-series information about the driving of the vehicle, and the time before the predetermined time from that time. A process of extracting a plurality of data between times and learning the relationship between cognitive load and dangerous driving using the extracted plurality of data;
a process of defining dangerous driving criteria for cognitive load from the learned relationship;
and
In the learning process, among the plurality of data, data whose cognitive load is all less than a threshold value is not used for learning.
learning program. to the computer,
A cognitive load is obtained from the physiological index measured by the physiological index measuring device, and the possibility of dangerous driving is determined based on the obtained cognitive load and the criterion defined by the learning program according to claim 10. processing;
A process of notifying based on the judgment result;
A driving assistance program that runs Based on the time-series data of cognitive load obtained from the driver's physiological index, the time at which the driving diagnosis is based on the driving data recorded in time-series information about the driving of the vehicle, and the time before the predetermined time from that time. a learning unit that extracts a plurality of data between times and uses the extracted plurality of data to learn the relationship between cognitive load and dangerous driving;
a definition unit that defines a criterion for dangerous driving with respect to cognitive load from the learned relationship;
with
wherein, of the plurality of data, the learning unit does not use, for learning, data in which all the cognitive loads included in the data are equal to or less than a threshold;
learning device. A cognitive load is obtained from the physiological index measured by the physiological index measuring device, and the possibility of occurrence of dangerous driving is determined based on the obtained cognitive load and the criterion defined by the learning device according to claim 12. a first judgment unit that
A notification unit that performs notification based on the judgment result;
driving assistance system. a first information processing device mounted on a vehicle;
a second information processing device capable of communicating with the first information processing device,
The first information processing device includes a communication unit that transmits time-series data of cognitive load and results of driving diagnosis based on the driving data to the second information processing device,
The second information processing device comprises the learning device according to claim 12,
learning system. Equipped with a first information processing device mounted on a vehicle,
The first information processing device comprises the learning device according to claim 12,
learning system.

Images