JP7466725B2 - Traffic management device and computer program - Google Patents

Description

The present invention relates to a management device for commercial vehicle drivers, and to a technology for objectively grasping the driving suitability of drivers, for example, during a driver roll call.

According to the Trucking Business Transportation Safety Regulations, transportation companies are required to conduct driver roll calls of their drivers. The "driver roll call" referred to here is the task of verifying that a driver who is about to start driving duties or a driver who has finished driving duties is the "real" driver, but the main purpose is to check the health status of the drivers for the day (including whether they are drunk or not) before they start driving and to give each driver the necessary instructions to ensure safe operation, in other words, to reduce the occurrence of dangerous driving or vehicle accidents to as close to zero as possible.

Driver roll calls are generally conducted by the operations manager meeting face-to-face with the drivers who are coming in for a checkup. This type of driver roll call is called a "face-to-face roll call." However, face-to-face roll calls can have the problem that many drivers are concentrated in a certain time period, causing stress for drivers waiting their turn and potentially interfering with their driving thereafter. Operations managers may also not be able to pay attention and give appropriate instructions.

As an example of conventional technology that solves such problems, the driver roll call system disclosed in Patent Document 1 links a health management system unit that measures the blood pressure, alcohol concentration, and body temperature of the driver (i.e., the driver crew member) to a roll call system unit for personal authentication, and displays the measurement results by the health management system unit on a display device during the driver roll call. The health management system unit also includes an interview terminal with a video camera and microphone, which not only makes it easier for the operations manager to grasp the health condition of the driver crew member, but also makes it possible for the driver crew member to improve their own health awareness by communicating their current health condition to the interview terminal.

JP 2017-68656 A

The driver roll call system disclosed in Patent Document 1 has the advantage of being able to grasp the health condition of the driver at the time of the driver roll call in a short time through measurement data such as blood pressure, but it is difficult to predict the operating conditions after the driver starts driving based on only one measurement data.
It is said that vehicle accidents caused by experienced drivers are often caused by a moment of carelessness (lack of concentration, etc.). And while it is thought that a traffic manager who has been in contact with the driver for a long time could possibly detect the signs of such lack of concentration, such as lack of sleep, stress, and illness that the driver himself is not aware of, this is difficult for a traffic manager who has not been in contact with the driver for a long time.
Such problems also required improvement from the perspective of labor management for driver and crew members.
The main objective of the present invention is to provide technology that can solve the above problems, make it easier to detect signs of lack of concentration that even crew members themselves cannot recognize, and thereby enable proper labor management.

The operation management device of the present invention is characterized in that it comprises an analysis means for analyzing tendencies of variation in the driver's driving operation based on operation data representing the driver's vehicle driving conditions for a predetermined period before driving starts and quantifying the results of the analysis ; a detection means for detecting an involuntary reaction of the driver before driving starts by acquiring at least one of image data capturing an image of the driver's face immediately before driving starts and speech sound data representing the contents of the driver's speech, and comparing the acquired data with the data of the driver acquired in the past; a diagnosis means for assigning a score based on a predetermined standard to each of the quantified analysis results and the detection result of the involuntary reaction, storing the scores, and tallying up the results of the scores for each predetermined period starting from the date of detection by the detection means, and diagnosing the risk level of the driver on the date of detection before driving starts based on the tallying up results; and an output means for outputting instruction information corresponding to the risk level before driving starts.

The present invention makes it easy to detect signs of carelessness that even the driver himself cannot recognize, and thus provides a traffic management device that enables proper labor management.

FIG. 1 is a diagram showing an example of the overall configuration of a crew management system. FIG. 2 is a diagram illustrating the configuration of data stored in a data server. 13 is an example of a scatter diagram relating to braking operation; FIG. 13 is an example of a scatter diagram relating to steering wheel operation. 13 is an example of a scatter diagram relating to a right/left turn operation; 13(a) and 13(b) are diagrams illustrating the safety index and accident prediction level. 1A and 1B are diagrams illustrating face image analysis. 4A to 4C are diagrams illustrating vitality and activity values. 6A to 6C are diagrams illustrating examples of instruction information. FIG.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.
In this embodiment, a driver management system is described that verifies that the driver who is the subject of the roll call is the driver himself/herself and diagnoses the driver's driving risk on the day during the depot roll call (pre-drive roll call) in driving work. The driving risk is diagnosed based on the driver's vehicle driving situation and involuntary reactions. Based on the driving risk, an objective judgment is made on the driver's driving suitability on the day.

<Overall composition>
1 is a diagram showing an example of the overall configuration of a crew management system to which the present invention is applied, and mainly shows its characteristic parts. This crew management system 1 is configured to include a management server 10, which is a traffic management device connected to a digital network N such as the Internet.
The digital network N is appropriately connected to roll call terminals 20 and data servers 30 installed, for example, at business establishments that manage driving operations, as well as information terminals such as mobile phones (not shown), each of which can access the management server 10.

<Roll call terminal>
1, two roll call terminals 20 are connected to the digital network N, but the number may be any number. An imaging device 21, a sound collector 22, a card reader/writer (R/W) 23, a display device 24, an alcohol detector 25, and a driver's license reader 26 are connected to the roll call terminal 20. The imaging device 21 is a camera that captures a facial image of the driver who is the subject of the roll call. The sound collector 22 is a microphone that captures the voice of the driver who is the subject of the roll call. The display device 24 is a display that displays a guide screen during the roll call and the results of the roll call (driving risk level).

The R/W 23 is equipped with a card holding mechanism that detachably holds the memory card 231, which is an example of a portable memory medium that records the driving operation status of the driver, such as the brake, accelerator, and steering wheel, recorded by the drive recorder. The memory card 231 records individual information for uniquely identifying the drive recorder to which it is attached, the operation date of the vehicle, and the driver ID and driver name for identifying the driver. As a result, the memory card 231 records the driving operation status on the operation day as something unique to the driver. The various data recorded in the memory card 231 are restricted from being changed by the driver. The recording mechanism of the drive recorder outputs the driving operation status to the memory card 231 in a form that is concealed from the driver. "Concealed form" means that the data is automatically output without being based on the driver's instructions. The roll call terminal 20 acquires the driving operation status written in the memory card 231 by the R/W 23 together with the driver ID. The drive recorder may also be equipped with a communication mechanism that communicates with the data server 30 via the digital network N and directly transmits the driving operation status and the driver ID.

The alcohol detector 25 detects the alcohol content of the driver who is the subject of the roll call. The driver's license reader 26 recognizes the contents of the letters, numbers, symbols, and images written on the surface of the driver's license held by the driver, for example. The roll call terminal 20 verifies the identity of the driver who is the subject of the roll call based on the driver's name recognized by the driver's license reader 26 and the driver's name obtained from the memory card 231 by the R/W 23. These peripheral devices for the roll call terminal 20 can themselves be publicly known.

The roll call terminal 20 is, for example, a personal computer, and operates by executing a predetermined computer program. The roll call terminal 20 transmits to the management server 10 imaging data including a facial image of the driver acquired from the imaging device 21 and speech sound data representing the speech content of the driver acquired from the sound collector 22. The roll call terminal 20 transmits to the data server 30 operation data representing the vehicle driving situation corresponding to the driving operation situation of the driver acquired from the memory card 231 by the R/W 23, together with the imaging data and speech sound data. The roll call terminal 20 also transmits to the management server 10 the detection result by the alcohol detector 25 and the identity confirmation result of the driver by the driver's license reader 26. The roll call terminal 20 displays an image on the display device 24 in response to an instruction from the management server 10.

<Data Server>
The data server 30 is a Web server and functions as a database that accumulates the imaging data, speech sound data, and operation data acquired from the roll call terminal 20 for each driver. Fig. 2 is an explanatory diagram of the configuration of data accumulated in the data server 30. The data server 30 accumulates the imaging data, speech sound data, and operation data by linking them to the driver ID. These data are transmitted from the roll call terminal 20 to the data server 30 at the time of roll call. The data server 30 stores the imaging data, speech sound data, and operation data by the date and time (here, date) at which they were acquired.

<Management Server>
The management server 10 is configured, for example, by a personal computer, and functions as a communication unit 11, an analysis unit 12, a detection unit 13, a diagnosis unit 14, and an output unit 15 by executing the computer program for operation management of the present invention. A display device 16 is connected to the management server 10. The management server 10 diagnoses the driving risk of the driver based on the image data and speech sound data at the time of roll call (detection date) acquired from the roll call terminal 20, and the image data, speech sound data, and operation data for a predetermined period starting from the time of roll call stored in the data server 30. The management server 10 outputs instruction information according to the diagnosed risk to the display device 16 or the roll call terminal 20. In this embodiment, a case where the image data and speech sound data are used to diagnose the risk level will be described, but the risk level may be diagnosed based on at least one of these data and the operation data.

The communication unit 11 is a communication interface for communicating between the roll call terminal 20 and the data server 30 via the digital network N. Various data used by the management server 10 for processing is input to the management server 10 via the communication unit 11. Furthermore, instruction information is transmitted to the roll call terminal 20 by the communication unit 11. Data is written from the management server 10 to the data server 30 via the communication unit 11.

Analysis Unit The analysis unit 12 performs a trend analysis of the driving operation of the driving crew based on the driving data of the driving crew acquired from the data server 30 by the communication unit 11. To this end, the analysis unit 12 acquires the driving data immediately before the driving start date of the driving crew from the data server 30. The analysis unit 12 performs a trend analysis of the braking operation, accelerator operation, steering operation, or a combination of these operations included in the driving data. The analysis unit 12 acquires the driving data of multiple driving crew from the data server 30 and analyzes the driving data of the multiple driving crew. The analysis unit 12 quantifies the analysis result of the driving data immediately before the driving start date of the driving crew based on the statistical value of the analysis result of the driving data of the multiple driving crew.

The analysis unit 12 performs the quantification using, for example, a scatter diagram of braking, accelerating, steering, or a combination of these operations. For example, there is a difference in the vehicle speed and the strength of braking when braking between a person with no accidents or violations and a person with an accident or violation history, with a person with no accidents or violations tending to perform strong braking at low speeds, and a person with an accident or violation history tending to perform strong braking at high speeds. The scatter diagram shows such operating tendencies between speed and braking strength.

Figure 3 is an example of a scatter diagram for braking operations. Comparing a scatter diagram of safe braking operations with a scatter diagram of dangerous braking operations that predict accidents, safe braking operations have less variation in the speed at which the brakes are applied, the strength of the brakes applied, and the gentleness of the brake operation. In the case of safe braking operations, braking operations are initiated at a slower speed than in the case of dangerous braking operations, the strength of the brakes applied is also lower, and the brake operation tends to be gentler. As a result, in the case of safe braking operations, the points representing the gentleness of the brake operation are concentrated at lower values than in the case of dangerous braking operations. In other words, safe braking operations are gentle braking operations performed at low speeds, and dangerous braking operations are sudden braking operations performed at high speeds.

Figure 4 is an example of a scatter diagram for steering. Comparing a scatter diagram for safe steering with a scatter diagram for dangerous steering that predicts an accident, safe steering has less variation in the maximum speed during steering, the angle of steering, and the gentleness of steering. In the case of safe steering, the maximum speed during operation is slower than in the case of dangerous steering, and the steering angle is smaller, so steering tends to be performed more gently. For this reason, in the case of safe steering, the points representing the gentleness of steering tend to be concentrated at lower values than in the case of dangerous steering. In other words, safe steering is gentle steering performed with sufficient deceleration, while dangerous steering is abrupt steering performed without deceleration.

Figure 5 is an example of a scatter diagram for right and left turn operations. Comparing a scatter diagram of a safe right and left turn operation with a scatter diagram of a dangerous right and left turn operation that predicts an accident, a safe right and left turn operation has less variation in the speed at the start of the right and left turn operation (e.g., the speed at the start of the turn operation), the steering angle, and the maximum speed during the right and left turn operation (e.g., the maximum speed at the intersection). In the case of a safe right and left turn operation, the speed at the start of the turn operation and the maximum speed during the turn tend to be lower than in the case of a dangerous right and left turn operation, and the steering angle also tends to be smaller. Therefore, in the case of a safe right and left turn operation, the points representing the centrifugal force generated by the right and left turn tend to be concentrated at lower values than in the case of a dangerous right and left turn operation. In other words, a safe right and left turn operation is, for example, a steering operation in which the vehicle decelerates sufficiently before entering an intersection, the maximum speed in the intersection is low, and the centrifugal force is low, while a dangerous right and left turn operation is, for example, a steering operation in which the vehicle enters an intersection at a high speed, the maximum speed at the intersection is high, and the centrifugal force is high.

The analysis unit 12 performs learning using the distribution of scatter plots based on the operation data of multiple drivers, the distribution of scatter plots based on the past operation data of the driver who is the subject of the roll call, and the presence or absence of an accident involving the driver, and quantifies the analysis results of the operation data immediately before the driver's start date of operation. At this time, the analysis unit 12 may analyze the operation data taking into account road conditions, weather, temperature, season, day of the week, traffic congestion forecast, etc.

The analysis unit 12 generates and outputs driving diagnosis data represented by multiple levels (five levels in this embodiment) of accident prediction levels and safety indexes through quantification. FIG. 6 is an explanatory diagram of the safety index and accident prediction level. The analysis unit 12 calculates the accident prediction rate, which is the analysis result of the quantified driving data for a predetermined period (one week in this embodiment) immediately before the start of operation of the driving crew, for each day, and calculates the average value of the accident prediction rate for the predetermined period. FIG. 6(a) illustrates the accident prediction rate for one week and its average value. The analysis unit 12 acquires the accident prediction level and the safety index according to the average value of the accident prediction rate for one week. FIG. 6(b) illustrates the relationship between the average value of the accident prediction rate, the accident prediction level, and the safety index. The accident prediction level is expressed in five levels, Level 1 to Level 5. The safety index is expressed in five levels, "100", "80", "60", "40", and "20", corresponding to the accident prediction level. In this example, the average value of the accident prediction rate for one week is 0.32, so the accident prediction level is Level 3 and the safety index is 60. The higher the accident prediction value, the higher the accident prediction level and the lower the safety index.

The detection unit 13 obtains at least one of the imaging data of the face image of the driver on that day and the speech sound data of the driver on that day from the roll call terminal 20 via the communication unit 11 before the start of driving, and detects an involuntary reaction of the driver using the obtained data. The detection unit 13 obtains at least one of the previous face image and speech sound data of the driver from the data server 30, and detects an involuntary reaction of the driver on that day by comparing the features of the face image or speech sound of the driver on that day with the features of the previous face image or speech sound.

Involuntary reaction detection by facial image When detecting an involuntary reaction using imaging data of a facial image, the detection unit 13 extracts the facial expression of the driver by a process of matching the imaging data with a predetermined reference data. The reference data is an image that serves as a criterion for extracting a facial expression from a facial image. The detection unit 13 compares the facial image with the reference data, and extracts the facial expression of the driver according to the difference. The detection unit 13 detects an involuntary reaction by comparing the facial expression extracted from the facial image acquired at the time of roll call with the facial expression of the driver extracted from the facial image before that. The detection unit 13 detects one or more characteristics of anger, contempt, disgust, fear, joy, sadness, and surprise that are different from those extracted and stored before that, as an involuntary reaction, by a matching process.

The detection unit 13 may extract the facial expression of the driver extracted from the previous facial image by, for example, acquiring previous image data of the driver from the data server 30 and performing a matching process between the acquired image data and reference data. In addition, the detection unit 13 may store the extraction result in the data server 30 each time the detection unit 13 extracts the facial expression of the driver from the facial image, and when detecting an involuntary reaction from the facial image, acquire the extraction result from the data server 30 as the facial expression of the driver extracted from the previous facial image.

The detection unit 13 performs emotion analysis from the facial expression by comparing the image data of the face image with the reference data, and extracts the facial expression of the driver. The emotion analysis is performed by scoring seven categories of emotion elements (anger, contempt, disgust, fear, joy, sadness, surprise) from the facial image of the captured data by image analysis of the facial expression. The detection unit 13 also detects the degree of drowsiness from the face image by eyelid behavior analysis, and grasps the sleeping state.
Fig. 7 is an explanatory diagram of face image analysis. In the example of Fig. 7(a), the scores of the seven emotion elements are "Happiness" is "100.00", "Surprise" is "05.00", and the rest are "00.00". The detection unit 13 detects a value representing an involuntary reaction by comparing each emotion element of the seven categories acquired from the face image at the time of roll call with the average score of each emotion element.

- Detection of involuntary reaction by speech sound When detecting an involuntary reaction using speech sound data, the detection unit 13 extracts speech sound characteristics specific to the driver by audio processing of the speech sound data. The detection unit 13 detects an involuntary reaction of the driver by comparing the extracted speech sound characteristics with previously extracted speech sound characteristics of the driver. For example, the detection unit 13 calculates a vitality whose value changes depending on the physical condition according to a comparison result between the speech sound characteristics extracted from the speech sound data acquired at the time of roll call and the speech sound characteristics of the driver extracted previously. The detection unit 13 detects the difference value between the previously calculated and stored vitality of the driver as an involuntary reaction.

The detection unit 13 may obtain previous speech sound data of the driver from the data server 30, for example, and extract the speech sound features of the driver extracted from the previous speech sound from the obtained speech sound data. In addition, the detection unit 13 may store the extraction result in the data server 30 each time it extracts speech sound features from previous speech sounds, and obtain the extraction result from the data server 30 when detecting an involuntary reaction from the speech sound data as the speech sound features of the driver extracted from the previous speech sounds.

"Vitality" is a value that represents the state of mind and level of vitality at the time of speaking, and is obtained by analyzing the frequency of the voice. The detection unit 13 accumulates and analyzes vitality over a certain period of time, and expresses the long-term trend of mental vitality (rising, stable, falling) as "activity value". Figure 8 is an explanatory diagram of vitality and activity value. Figure 8(a) shows an example of vitality over one week. Figure 8(b) shows an example of activity value over one week. The detection unit 13 detects the average value of vitality (G) and activity value (K) ((G+K)/2) as a value representing an involuntary reaction (Figure 8(c)).

Diagnosis Unit The diagnosis unit 14 diagnoses the driving risk of the driver on the detection day by comparing the quantified analysis result by the analysis unit 12 and the detection result of the involuntary reaction by the detection unit 13 with predetermined conditions. The diagnosis unit 14 outputs the risk, which is the diagnosis result, to the output unit 15. The output unit 15 outputs instruction information according to the risk by the display device 16. Note that the output unit 15 may not only display the instruction information on the display device 16, but also transmit the instruction information to the roll call terminal 20 via the communication unit 11 and display it on the display device 24 of the roll call terminal 20.

Figure 9 is an example diagram of instruction information output by display device 16. Figure 9(a) is an overall view of the roll call record table including instruction information. Figure 9(b) is an enlarged view of an item on the roll call record table. The roll call record table is displayed on display device 16. The operations manager checks the roll call record table displayed on display device 16 during a roll call, and uses the contents of the roll call record table as the basis for making decisions when calling and giving instructions to the driver. By receiving a roll call from the operations manager while checking the roll call record table displayed on display device 24 connected to roll call terminal 20, the driver can check the degree of danger while driving, etc.

The roll call record sheet is digitized and stored in the management server 10, and the results of the pre-drive roll call and the post-drive roll call are recorded for each vehicle, driver name, operation details, destination, and customer name. The results of the pre-drive roll call include the roll call method, roll call location, roll call time, daily inspection confirmation, risk prediction tool judgment, health condition (illness/fatigue/lack of sleep), alcohol detection results, roll call executor (operations manager name), and instructions. The results of the post-drive roll call include the roll call method, roll call location, roll call time, health condition (illness/fatigue), alcohol detection results, roll call executor (operations manager name), rest time the previous day, one month's working time, and instructions. After conducting a roll call based on the roll call record sheet, the operations manager can enter the necessary information into the roll call record sheet and use it at a future roll call.

The roll call record table displays the risk level, which is the result of the diagnosis by the diagnosis unit 14, in the "Risk Prediction Tool Judgment" item. The risk level is displayed as red, yellow, or green. If the risk level is displayed as red or yellow as a result of the diagnosis by the diagnosis unit 14, the "Instructions, etc." item displays that there are precautions according to the diagnosis result. In the example of FIG. 9(b), the risk level is yellow and a notice is displayed that there are precautions. When the operation manager selects the yellow judgment button for "Risk Prediction Tool Judgment," the diagnosis results based on the driving operation tendency and the involuntary reaction detection results that were the basis for the risk level judgment by the management server 10 are displayed on the display device 16. The operation manager can then provide driving guidance to the driver. In addition, the operation manager can check the previous day's rest time and the one-month working time during post-drive inspections and give instructions to the driver.

Figure 9 (c) is an example diagram of the details of the diagnosis result. The details of the diagnosis result include the result of the driving diagnosis, the diagnosis result based on the image data of the face image, and the diagnosis result based on the speech sound data. The diagnosis unit 14 assigns a score based on a predetermined standard to the analysis result of the quantified trend analysis (driving diagnosis data) obtained from the analysis unit 12. In the example of Figure 9 (c), the diagnosis unit 14 assigns a score of "60" to the analysis result of the quantified trend analysis (accident prediction level, safety index) obtained from the analysis unit 12. The diagnosis unit 14 tally up the scores assigned in this way according to the driving diagnosis data for each predetermined period starting from the detection date.

The diagnosis unit 14 assigns a score based on a predetermined criterion to the involuntary reaction detection result obtained from the detection unit 13. When an involuntary reaction is detected from the image data of the face image, the diagnosis unit 14 assigns a score based on the average score of the emotion analysis and the weekly emotion trend line. FIG. 7(b) illustrates a weekly emotion trend line. The diagnosis unit 14 calculates the average score for each emotion element over one week. The weekly emotion trend line is a line connecting the average scores of the emotion elements over a predetermined period in the past (e.g., one week). In the example of FIG. 9(c), the diagnosis unit 14 assigns a score of "60" to the involuntary reaction detected from the image data of the face image. The diagnosis unit 14 tally up the scores thus assigned according to the face image data for each predetermined period, counting from the date of detection.

When an involuntary reaction is detected from the speech sound data, the diagnosis unit 14 assigns a score based on vitality and activity value. In the example of FIG. 9(c), the diagnosis unit 14 assigns a score of "80" to the value representing the involuntary reaction detected from the speech sound data illustrated in FIG. 8(c). The diagnosis unit 14 tallys up the scores assigned in this way according to the speech sound data for each specified period, counting from the date of detection.

The diagnosis unit 14 determines which of the multiple risk levels the driving operation tendency on the detection date belongs to, based on the total value of the tabulated results of each score. In the example of FIG. 9(c), the diagnosis unit 14 determines that the risk level belongs to the range of "90 to 210" because the total value of each score is "200". If the risk level is in the range "300 to 230", the diagnosis unit 14 outputs instruction information to the output unit 15 to display a green color image. If the risk level is in the range "210 to 90", the diagnosis unit 14 outputs instruction information to the output unit 15 to display a yellow color image. If the risk level is in the range "90 to 20", the diagnosis unit 14 outputs instruction information to the output unit 15 to display a red color image.

By displaying color images and messages, the operation manager can grasp the driver's driving suitability for the day based on the driver's risk level. The operation manager makes a comprehensive judgment on whether the driver can operate the vehicle based on the roll call record sheet and messages displayed on the display device 16 and the labor management data indicating the driver's working hours and overtime hours. The operation manager can then provide driving guidance to the driver or instruct him or her to stop driving. In addition, if the alcohol detection result obtained from the roll call terminal 20 is equal to or higher than a specified value or if the driver's identity verification result based on the driver's license does not match the driver's name, the diagnosis unit 14 may display a message on the display device 16 or the display device 24 of the roll call terminal 20 that the driver is not allowed to drive without performing the diagnosis process. This allows the operation manager to prevent drunk driving and not carrying a driver's license in advance.

<Operational form>
10 is an explanatory diagram of the delivery roll call process by the management server 10. The management server 10 starts the delivery roll call process by accepting a request for the roll call process from the roll call terminal 20 at the time of the delivery roll call. When the management server 10 starts the delivery roll call process, it first acquires the operation data, at least one of the image data of the face image and the speech sound data, the alcohol detection result, and the driver's license identification result from the roll call terminal 20 through the communication unit 11 (S101). The management server 10 checks whether the alcohol detection result is equal to or higher than a prescribed value and whether the driver's license is the driver's own identity based on the acquired alcohol detection result and the driver's license identification result, and judges whether the delivery roll call process should be continued (S102).

If the alcohol detection result is equal to or exceeds a prescribed value, or if the driver's license recognition result does not match the driver, the management server 10 stops the departure roll call process (S102: N). The management server 10 displays a message on the display device 16 and the display device 24 of the roll call terminal 20 indicating that the driver is not permitted to drive, and ends the process (S103).

If the alcohol detection result is below the specified value and the driver's license recognition result matches the driver, the management server 10 continues the departure roll call process (S102: Y). The management server 10 analyzes the driving data using the analysis unit 12 and outputs the quantified analysis results (S104). The management server 10 detects an involuntary reaction of the driver based on at least one of the image data of the face image and the speech sound data using the detection unit 13, and outputs the detection results (S105).

The diagnosis unit 14 acquires the quantified analysis results output from the analysis unit 12 and the involuntary reaction detection results output from the detection unit 13, and diagnoses the driving risk of the driver based on the analysis results and the involuntary reaction detection results (S106). The diagnosis unit 14 transmits the diagnosis results to the output unit 15. The output unit 15 displays the roll call record table in which the diagnosis results have been input into the risk prediction tool item on the display unit 16 and the display unit 24 of the roll call terminal 20 (S107). The operation manager judges whether or not the driver can operate the vehicle based on the roll call record table displayed on the display unit 16.

Effects of the embodiment
In this way, by using the crew management system 1 of this embodiment, the operation manager can perform an advanced depot roll call that contributes to the safe driving of the driver by using comprehensive quantified data that takes into account the results of comparison and analysis with regular data obtained from big data acquired in the past and stored in the data server 30, rather than a subjective judgment alone. Specifically, the operation manager can judge the driving risk of the driver and the state of the driver at that time by referring to the roll call record table that includes the results of comparison and analysis between the data acquired at the time of the roll call and the data acquired in the past. As a result, the operation manager can easily detect signs indicating a lack of concentration that even the driver himself cannot recognize by using the crew management system 1, and can properly manage labor.

For example, the crew management system 1 quantifies the results of trend analysis of braking, accelerating, steering, or a combination of these operations immediately before the start of driving based on the statistical values of the analysis results of multiple drivers. For example, the crew management system 1 quantifies based on a scatter diagram of the analysis results of braking, accelerating, steering, or a combination of these operations of multiple drivers. The crew management system 1 assigns a score based on a predetermined standard to the results of the trend analysis, and tally up the scores for each predetermined period starting from the detection date, and determines which of multiple levels of risk the driving operation tendency on the detection date belongs to based on the tally up results. By using the statistical values of the analysis results of the driving operations of multiple drivers, such a crew management system 1 can appropriately determine the risk of accident occurrence from the driving operation tendency of the driver who is the subject of the roll call. The operation manager can determine the driving risk of the driving crew based on comprehensive quantified data and manage the work appropriately.

The crew management system 1 of this embodiment detects an involuntary reaction by extracting the driver's facial expression (anger, contempt, disgust, fear, joy, sadness, surprise) by comparing the captured image data with predetermined reference data, and comparing the extracted facial expression with the driver's facial expressions previously extracted. The crew management system 1 assigns a score based on a predetermined criterion to the detected feature, and tally up the scores for each predetermined period starting from the detection date, and determines which of multiple levels of risk the involuntary reaction on the detection date belongs to based on the tally up results. Such a crew management system 1 makes it possible to determine the risk level based on the involuntary reaction detected by comparing it with the driver's past facial expressions. Therefore, the operation manager can easily detect signs indicating a lack of concentration that even the driver himself cannot recognize using the crew management system 1, and can properly manage labor.

The crew management system 1 of this embodiment detects an involuntary reaction by comparing the speech sound characteristics unique to the driver with previously extracted speech sound characteristics of the driver. For example, the crew management system 1 calculates a vitality value that changes depending on the physical condition from the speech sound characteristics, and detects the difference value between the previously calculated vitality value of the driver as an involuntary reaction.
The crew management system 1 assigns a score based on a predetermined standard to the detected difference value, and tally up the scores for each predetermined period starting from the detection date, and determines which of multiple risk levels the involuntary reaction on the detection date belongs to based on the tally up results. Such a crew management system 1 makes it possible to determine the risk level based on the involuntary reaction detected by comparing it with the driver's past vitality. Therefore, the operation manager can easily detect signs of lack of concentration that even the driver himself cannot recognize using the crew management system 1, and can properly manage the work.

The crew management system 1 of this embodiment associates the diagnostic results with a red, yellow, or green color image and displays them on a specified display device. It also outputs message information corresponding to the diagnostic results. Because the diagnostic results are displayed as color images, the operations manager can easily visually confirm the diagnostic results. In addition, because message information corresponding to the diagnostic results is output, the operations manager can give accurate instructions to the driving crew.

1...crew management system, 10...management server, 20...roll call terminal, 30...data server, 21...imaging device, 22...sound collector, 23...card reader/writer (R/W), 16, 24...display device, 11...communication unit, 12...analysis unit, 13...detection unit, 14...diagnosis unit, 15...output unit, 231...memory card, N...digital network.

Claims (4)

an analysis means for analyzing a tendency of variability in a driver's driving operation based on driving data representing a vehicle driving state by the driver for a predetermined period before the start of driving , and quantifying a result of the analysis ;
A detection means for detecting an involuntary reaction of the driver before the start of driving by acquiring at least one of image data capturing an image of the face of the driver immediately before the start of driving and speech sound data representing the contents of the speech of the driver and comparing the acquired data with the data of the driver acquired in the past;
a diagnostic means for assigning a score based on a predetermined criterion to each of the quantified analysis results and the involuntary reaction detection results , storing the scores, and tallying up the scores for each predetermined period starting from the detection date by the detection means, and diagnosing the risk of the driver on the detection date based on the tallying up results before the start of driving;
an output means for outputting instruction information according to the risk level before starting driving;
An operation management device comprising: 2. The traffic management device according to claim 1, wherein the detection means detects the involuntary reaction by extracting the driver's facial expression by matching the imaging data with predetermined reference data, and comparing the extracted facial expression with a previously extracted and stored facial expression of the driver. 3. The traffic management device according to claim 2, wherein the detection means detects, through the matching process, one or more characteristics of anger, contempt, disgust, fear, joy, sadness, and surprise that are different from those previously extracted and stored as one of the involuntary reactions. A computer program for traffic management that causes a computer to function as the traffic management device according to any one of claims 1 to 3 .

Images