Method and apparatus for analyzing driver fatigue

Abstract

To more accurately detect the driver's fatigue level. [Solution] The driver fatigue analysis method acquires first data showing environmental information of the driver while driving, measured by sensors of a vehicle capable of delivering goods, and second data showing information on the driver's daily delivery plan by the vehicle. A driver fatigue prediction model is trained using multiple pairs of acquired first and second data as training data to predict the driver's fatigue level. Based on the first and second data acquired for each driver, the driver's fatigue level is predicted, and the predicted driver fatigue level is notified.

Description

Technical Field

[0001] The present disclosure relates to a driving fatigue degree analysis method and a driving fatigue degree analysis device.

Background Art

[0002] In Patent Document 1, in order to determine whether a driver is fatigued and detect the driver's fatigue, a driver fatigue degree estimation device is disclosed that takes fatigue reduction measures for the driver when the continuous driving time exceeds the allowable driving time. This driver fatigue degree estimation device stores a fatigue map in which high-fatigue areas are registered for each region, and calculates the fatigue degree according to the high-fatigue driving time spent driving in a high-fatigue area determined by referring to the fatigue map. The driver fatigue degree estimation device changes the allowable driving time according to the calculated fatigue degree, and determines that it is the timing to take fatigue reduction measures for the driver when the continuous driving time exceeds the allowable driving time.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] For example, in the transportation industry, it has begun to be said that it is necessary to visualize and grasp the fatigue degree of drivers in order to improve the safety of vehicle driving operations. Especially recently, with the shortage of young talents in the transportation industry and the aging of drivers, chronic labor shortages have become a problem. If the information on the fatigue degree of drivers is not sufficiently grasped, drivers with accumulated fatigue may be forced into long-distance or long-time driving operations. That is, since the safety of driving operations may not be ensured, urgent measures are required.

[0005] In the configuration of Patent Document 1, the driver's fatigue level is calculated based on the amount of time spent driving in high-fatigue areas determined by referring to a fatigue map. However, drivers may drive not only in high-fatigue areas but also in non-high-fatigue areas not based on the fatigue map. Furthermore, the driver's fatigue level is thought to be influenced by their driving plan for the day (for example, how long and how far they will drive). For this reason, there was room for improvement over the conventional technology in terms of technology for determining the driver's fatigue level.

[0006] This disclosure was devised in light of the conventional circumstances described above, and aims to more appropriately detect the degree of driver fatigue. [Means for solving the problem]

[0007] This disclosure provides a driving fatigue analysis method that acquires first data indicating environmental information of a driver while driving, measured by sensors of a vehicle capable of delivering goods, acquires second data indicating information on the driver's daily delivery plan using the vehicle, a driver fatigue prediction model that has been machine-trained to predict the driver's fatigue level using a plurality of pairs of acquired first and second data as training data, and predicts the driver's fatigue level based on the first and second data acquired in relation to the driver, and notifies the driver of the predicted fatigue level.

[0008] Furthermore, this disclosure provides a driving fatigue analysis device comprising a processor and a memory, wherein the processor, in cooperation with the memory, acquires first data indicating environmental information of a driver while driving, measured by sensors of a vehicle capable of delivering goods, and second data indicating information on the driver's daily delivery plan by the vehicle, and predicts the driver's fatigue level based on a driver fatigue prediction model that has been machine-trained to predict the driver's fatigue level using a plurality of pairs of acquired first data and second data as training data, and notifies the driver of the predicted driver fatigue level.

[0009] These comprehensive or specific embodiments may be implemented as systems, devices, methods, integrated circuits, computer programs, or recording media, or as any combination of systems, devices, methods, integrated circuits, computer programs, and recording media. [Effects of the Invention]

[0010] According to this disclosure, the degree of driver fatigue can be detected more appropriately. [Brief explanation of the drawing]

[0011] [Figure 1] This figure shows an example of the system configuration during operation of the driver fatigue analysis system according to this embodiment. [Figure 2] Block diagram showing a basic hardware configuration example of a computer device. [Figure 3] This diagram shows a data table illustrating a specific example of data stored in the data storage unit. [Figure 4] This diagram shows a feedback table illustrating specific examples of the content that is fed back from the fatigue management system. [Figure 5] Figure 1 is a sequence diagram showing an example of the overall operation procedure of the driver fatigue analysis system in chronological order. [Figure 6] A flowchart showing an example of the operation procedure for the data acquisition flow in Figure 5, in chronological order. [Figure 7] Figure 5 is a flowchart showing an example of the operation procedure of the judgment and notification flow in chronological order. [Figure 8] Figure 5 is a flowchart showing an example of the operation procedure of the warning / suggestion flow in chronological order. [Figure 9] Figure 5 is a flowchart showing an example of the operation procedure of the delivery plan update flow in chronological order. [Figure 10] A flowchart showing an example of the operation procedure for the fatigue level confirmation flow in Figure 5, in chronological order. [Figure 11] This figure shows an example of the system configuration during the learning phase of the AI ​​for driver fatigue analysis in the driver fatigue analysis system according to this embodiment. [Figure 12]Sequence diagram showing an example of the overall operation procedure by the driving fatigue degree analysis system of FIG. 11 in chronological order [Figure 13] Flowchart showing an example of the operation procedure of the annotation addition flow of FIG. 12 in chronological order [Figure 14] Flowchart showing an example of the operation procedure of the driving fatigue degree analysis AI generation flow of FIG. 12 in chronological order

Mode for Carrying Out the Invention

[0012] Hereinafter, with appropriate reference to the drawings, embodiments specifically disclosing the driving fatigue degree analysis method and the driving fatigue degree analysis apparatus according to the present disclosure will be described in detail. However, detailed descriptions more than necessary may be omitted. For example, detailed descriptions of well-known matters and duplicate descriptions for substantially the same configurations may be omitted. This is to avoid making the following description unnecessarily redundant and to facilitate the understanding of those skilled in the art. Note that the attached drawings and the following description are provided for those skilled in the art to fully understand the present disclosure, and it is not intended to limit the subject matter described in the claims thereby.

[0013] In the following embodiments, a use case will be exemplified and described in which the fatigue degree of a driver who makes a living by driving operations working for a transportation company is predicted, visualized, and managed, and feedback is given to the driver as necessary. However, the following embodiments are not limited to the above-described use case.

[0014] 1. Configuration of the driving fatigue degree analysis system (during actual operation) First, referring to FIGS. 1 to 4, the configuration during actual operation of the driving fatigue degree analysis system 100 according to this embodiment will be described. Actual operation means an operation of predicting the fatigue degree of a driver by inputting various data (described later) actually acquired for the driver after the driving fatigue degree analysis AI 54 for predicting the fatigue degree of the driver is generated or updated. FIG. 1 is a diagram showing an example of the system configuration during operation of the driving fatigue degree analysis system 100 according to this embodiment. FIG. 2 is a block diagram showing an example of the basic hardware configuration of the computer device 1. FIG. 3 is a diagram showing a data table TBL1 exemplifying a specific example of the data stored in the data storage unit 53. FIG. 4 is a diagram showing a feedback table TBL2 exemplifying a specific example of the content fed back from the fatigue degree management system 50.

[0015] The driving fatigue degree analysis system 100 includes at least a driving vehicle system 10, an external system 20, a health management system 30, a delivery management system 40, and a fatigue degree management system 50 (see FIGS. 1 and 11). The driving vehicle system 10, the external system 20, the health management system 30, the delivery management system 40, and the fatigue degree management system 50 are connected to each other via a network (not shown) so that data signals can be transmitted and received. The network may be, for example, a Wide Area Network (WAN), a Local Area Network (LAN), a Long Term Evolution (LTE), a mobile communication such as 4G or 5G, a power line communication, a short-range wireless communication (e.g., Bluetooth (registered trademark) communication), or a communication for mobile phones. Further, the external system 20, the health management system 30, the delivery management system 40, and the fatigue degree management system 50 are each constituted by the computer device 1 shown in FIG. 2. The relationship between each system and the computer device 1 will be described later.

[0016] The driving vehicle system 10 is configured as a computer device 1 and is installed in the vehicle VC1 in which the driver who is the subject of fatigue analysis by the driving fatigue analysis system 100 is riding. The driving vehicle system 10 includes at least the computer device 1 shown in Figure 2 (not shown in Figure 1), a driver monitoring system 11, a vital sensor 14, and an HMI 15. The driver monitoring system 11 includes at least a gyro sensor 12 and an in-vehicle camera 13, and extracts environmental information (an example of first data) of the driver while driving, measured by sensors in the vehicle VC1 capable of delivering cargo.

[0017] The gyro sensor 12 detects and extracts rotation information, orientation information, and vibration information of the vehicle VC1. Based on the rotation information, orientation information, and vibration information of the vehicle VC1, the gyro sensor 12 may also calculate and extract information indicating the driver's comfort level of the driving environment of the vehicle VC1 (see Figure 3). The data showing the various types of information extracted by the gyro sensor 12 is periodically sent from the computer device 1 (not shown in Figure 1) of the driving vehicle system 10 to the fatigue management system 50 via a network (not shown).

[0018] The in-vehicle camera 13 captures images of the driver operating the vehicle VC1, and analyzes and extracts various information about the driver's appearance based on the captured image data. This information may include, for example, the driver's eyelid movements, facial expressions, posture, actions (e.g., yawning), and at least one of the following based on combinations of these: the direction of the driver's gaze, the appropriateness of the direction of the gaze, fatigue level, and drowsiness information (see Figure 3). The data showing the various information extracted by the in-vehicle camera 13 is periodically sent from the computer device 1 of the driving vehicle system 10 (not shown in Figure 1) to the fatigue management system 50 via a network (not shown).

[0019] The vital sensor 14 is a sensor worn by the driver operating the vehicle VC1, and is, for example, a wristwatch type. However, the vital sensor 14 is not limited to a wristwatch type; for example, it may be a type placed inside the vehicle VC1. The vital sensor 14 extracts various information about the driver's biological information. This information may include, for example, the driver's pulse rate, body temperature, activity level, blood oxygen saturation, blood pressure, and at least one of the following based on combinations of these: the driver's sleep pattern such as sleep duration, sleep quality, and sleep depth, recovery ability after exercise, and stress level (see Figure 3). The data showing the various information extracted by the vital sensor 14 is periodically sent from the computer device 1 of the driving vehicle system 10 (not shown in Figure 1) to the fatigue management system 50 via a network (not shown).

[0020] HMI15 functions as a human-machine interface to enable interaction between the driver (human) operating the vehicle VC1 and the vehicle VC1, and is controlled by computer device 1 (not shown in Figure 1). In other words, HMI15 is placed inside the vehicle VC1 as an input / output device for computer device 1. HMI15 receives input (responses) from the driver in response to various information (e.g., warnings, suggestions) displayed on this input / output response device via computer device 1 (not shown in Figure 1), and displays feedback content (see Figure 4) from other systems (e.g., delivery management system 40, fatigue management system 50).

[0021] The external system 20 is configured as a computer device 1 and extracts weather information 21 and traffic information 22 for a delivery day when, for example, the driver drives vehicle VC1 to deliver goods based on a work instruction from the delivery management system 40. The weather information 21 and traffic information 22 are data stored in the memory 4 of the computer device 1, for example. Weather information 21 indicates weather information, temperature information, humidity information, pollen dispersion information, etc., at the location of vehicle VC1 (see Figure 3). Traffic information 22 indicates congestion information, traffic regulation information, etc., at the location of vehicle VC1 (see Figure 3). It is preferable that the setting of the delivery area necessary for extracting weather information 21 and traffic information 22 is performed in advance before the start of actual operation of the driver fatigue analysis system 100 according to this embodiment. In other words, the external system 20 extracts weather information 21 and traffic information 22 (an example of third data) for a delivery day when vehicle VC1 is driven, based on the setting of the delivery area necessary for extracting weather information 21 and traffic information 22. The weather information 21 and traffic information 22 extracted by the external system 20 are periodically sent from the external system 20 to the fatigue management system 50 via a network (not shown).

[0022] The health management system 30 is configured as a medical terminal (e.g., computer device 1) used by medical professionals (e.g., potential nurses) working at medical institutions (e.g., general hospitals or individual clinics) that have individual contracts with the delivery company. The health management system 30 manages information indicating the health status of each driver performing driving duties within the delivery company, based on medical diagnoses made by medical professionals to that driver (e.g., expert knowledge data 31, and personal data such as driver health-related data 32). The expert knowledge data 31 and driver health-related data 32 are data stored in the memory 4 of the computer device 1, for example. The driver health-related data 32 may include at least one of the following: age, sex, height, weight, vision, hearing, heart rate, examination date, physical examination findings, blood pressure, urinalysis, liver function, lipids, kidney function, uric acid, diabetes, blood, electrocardiogram, chest X-ray, gastric X-ray, fecal occult blood test, metabolic syndrome diagnosis, dental examination results, medical history, medication, sleep duration, smoking status, exercise time, alcohol consumption, nutritional balance, work style, stress diagnosis results, work history, etc. (see Figure 3). Based on the feedback from the fatigue management system 50 (see Figure 4), expert knowledge data 31, and driver health-related data 32, the health management system 30 corrects (modifies) the fatigue level of the target driver through operation by a medical professional working at a medical institution. In this case, the health management system 30 notifies the fatigue management system 50 of data showing the corrected fatigue level.

[0023] The delivery management system 40 is configured as a delivery management terminal (e.g., computer device 1) used by the delivery manager of the transportation company where the driver operating vehicle VC1 works. The delivery management system 40 includes at least a delivery planning / execution management unit 41 and delivery management information 42. The delivery planning / execution management unit 41 is configured, for example, by the processor 2 of computer device 1. The delivery management information 42 is data stored, for example, in the memory 4 of computer device 1. The delivery planning / execution management unit 41 determines and manages the daily delivery plan for each driver operating vehicle VC1. The daily delivery plan shows, for example, the delivery plan and delivery results for vehicle VC1 performing deliveries. Specifically, the delivery plan and delivery results may include at least one of the following: source information, transit point information, destination information, vehicle information, driver information, load capacity information, delivery deadline information, transportation route information, etc. (see Figure 3). The delivery management information 42 is data showing the delivery plan and delivery results determined by the delivery planning / execution management unit 41. The delivery management system 40 periodically sends data showing delivery management information 42 for each driver to the fatigue management system 50.

[0024] The fatigue management system 50 may be configured as, for example, a computer device 1 and deployed in a cloud-based manner, or deployed as an on-premise system at a specific business location (e.g., a transportation company). The fatigue management system 50 includes at least a data acquisition unit 51, a driver information management unit 52, a data storage unit 53, a driver fatigue analysis AI 54, a fatigue prediction unit 55, a notification unit 56, and a fatigue correction unit 57. The data acquisition unit 51, the driver information management unit 52, the driver fatigue analysis AI 54, the fatigue prediction unit 55, the notification unit 56, and the fatigue correction unit 57 are configured as, for example, a processor 2 of the computer device 1, or a combination of the processor 2 and a communication interface circuit 3. The data storage unit 53 is configured as, for example, a memory 4 of the computer device 1.

[0025] The data acquisition unit 51 receives and acquires data transmitted from the driving vehicle system 10, the external system 20, and the delivery management system 40, and sends it to the driver information management unit 52.

[0026] The driver information management unit 52 organizes and manages various data acquired by the data acquisition unit 51, linking them to the same driver, and stores the various data linked to the same driver in the data storage unit 53. Alternatively, the driver information management unit 52 and the data storage unit 53 may be integrated, and the data storage unit 53 may be omitted.

[0027] The data storage unit 53 serves as the memory for the fatigue management system 50, storing various data associated with the same driver by the driver information management unit 52. The data storage unit 53 also stores data from the driver fatigue analysis AI 54, which is generated for each driver. This driver-specific driver fatigue analysis AI 54 data is read out as needed by the fatigue prediction unit 55.

[0028] The Driving Fatigue Analysis AI 54 is a driving fatigue prediction model that predicts the fatigue level of a target driver, generated or updated by machine learning using training data (see Figures 11 to 13) consisting of various data accumulated for each driver. Specifically, the Driving Fatigue Analysis AI 54 can take at least one of the data from the driving vehicle system 10, the external system 20, and the delivery management system 40, all linked to the same driver, as input and predict the fatigue level of that driver as output. The Driving Fatigue Analysis AI 54 can also take at least one of the data from the driving vehicle system 10, the external system 20, the health management system 30, and the delivery management system 40, all linked to the same driver, as input and predict the fatigue level of that driver as output.

[0029] The fatigue prediction unit 55 reads a driving fatigue analysis AI 54 corresponding to the driver whose fatigue level is to be predicted from the data storage unit 53, and inputs at least one of the data from the driving vehicle system 10, the external system 20, the health management system 30, and the delivery management system 40 into the read-out driving fatigue analysis AI 54 to predict the driver's fatigue level. The fatigue prediction unit 55 sends the predicted result of the driver's fatigue level to the notification unit 56. If the fatigue prediction unit 55 obtains the corrected fatigue level from the fatigue correction unit 57, it sends that fatigue level to the notification unit 56 in place of the predicted result.

[0030] The notification unit 56 acquires the fatigue level prediction result from the fatigue level prediction unit 55 or the fatigue level correction result from the fatigue level correction unit 57, and notifies the acquired fatigue level to the driving vehicle system 10, the health management system 30, and the delivery management system 40, respectively.

[0031] The fatigue level correction unit 57 acquires the corrected result when the content of the fatigue level prediction result is corrected (modified) based on an operation based on a medical diagnosis by a medical professional as a result of the driver's fatigue level prediction result being notified to the health management system 30 from the notification unit 56. If the fatigue level correction unit 57 receives instructions from the health management system 30 for correction content to the fatigue level prediction result rather than the corrected fatigue level, it may correct (modify) the fatigue level prediction result by the fatigue level prediction unit 55 based on those instructions. The fatigue level correction unit 57 sends the corrected fatigue level to the fatigue level prediction unit 55.

[0032] Now, referring to Figure 4, we will explain the content of the feedback received from the fatigue management system 50.

[0033] (common) The notification unit 56 of the fatigue management system 50 sends the driver's fatigue prediction result from the fatigue prediction unit 55, or the fatigue level and the content of various warnings / alerts based on that fatigue level, to each system (i.e., the driving vehicle system 10, the health management system 30, and the delivery management system 40). The content of various warnings / alarms based on fatigue level is stored in the data storage unit 53 in a table (not shown) that is generated in advance according to the fatigue level, and the notification unit 56 can refer to this table (not shown) to obtain the content of warnings / alarms corresponding to the fatigue level.

[0034] (individual) (1) Feedback to the driving vehicle system 10 The notification unit 56 of the fatigue management system 50 transmits to the driving vehicle system 10 the content of a suggestion for a rest based on the fatigue level prediction result of the fatigue level prediction unit 55, and a suggestion for a driving pattern (for example, reducing speed) based on that fatigue level prediction result, and provides feedback. The content of the suggestion for a rest based on fatigue level and the suggestion for a driving pattern based on fatigue level are stored in the data storage unit 53 in a table (not shown) that has been generated in advance according to the fatigue level, and the notification unit 56 can refer to this table (not shown) to obtain the content of the suggestion for a rest corresponding to the fatigue level and the suggestion for a driving pattern corresponding to that fatigue level.

[0035] (2) Feedback to the delivery management system 40 The notification unit 56 of the fatigue management system 50 transmits to the delivery management system 40 the content of proposals for revising the current operation plan based on the fatigue prediction results of the fatigue prediction unit 55, proposals for changing future operation plans based on the fatigue prediction results, and proposals for incentives and rewards for taking action based on the fatigue prediction results, and provides feedback. Proposals for revising the current operation plan include, for example, taking rest time, changing arrival times due to reduced speed, or changing drivers. Proposals for changing future operation plans include, for example, changing future rest and working hours for drivers, or changing rest times. Proposals for incentives and rewards include, for example, adjusting the treatment of drivers who take appropriate action based on warnings and drive safely or reduce fatigue. The content of these proposals, incentives and rewards are based on the transportation company's policy, so it is not preferable to store them as data in the data storage unit 53, but it is also acceptable for them to be stored as data in the data storage unit 53 and for the notification unit 56 to select from that data and notify the delivery management system 40.

[0036] (3) Feedback to the health management system 30 The notification unit 56 of the fatigue management system 50 transmits the contents of the main input data used when the driving fatigue analysis AI 54 predicted the driver's fatigue level, as feedback to the health management system 30. The notification unit 56 can obtain this input data from the fatigue prediction unit 55. The feedback contents of this input data are positioned as reference information for double-checking, for example, whether the fatigue level prediction by the driving fatigue analysis AI 54 was appropriate from the perspective of medical professionals.

[0037] Here, with reference to Figure 2, an example of the hardware configuration of the computer device 1 that constitutes the driving vehicle system 10, external system 20, health management system 30, delivery management system 40, and fatigue management system 50 will be described. As shown in Figure 2, the computer device 1 includes a processor 2, a communication interface circuit 3, and memory 4. In the computer device 1, the processor 2, the communication interface circuit 3, and the memory 4 are connected via an internal bus 5 so that data signals can be input and output to and from each other.

[0038] Processor 2 is composed of at least one of the following: a Central Processing Unit (CPU), a Digital Signal Processor (DSP), or a Field Programmable Gate Array (FPGA). Processor 2 functions as a controller that manages the overall operation of computer device 1. Processor 2 performs control processing to coordinate the operation of each part of computer device 1, data input / output processing between each part of computer device 1, data calculation processing, and data storage processing. Processor 2 operates according to the program stored in memory 4. During operation, Processor 2 uses memory 4 to cooperate in executing various processes and temporarily stores data generated or acquired by Processor 2 in memory 4.

[0039] The communication interface circuit 3 consists of an interface circuit for wirelessly or via wired transmission and reception of data signals between computer device 1 and another computer device 1. The communication interface circuit 3 sends data signals received from the other computer device 1 to the processor 2 or memory 4. The communication interface circuit 3 transmits data signals acquired from the processor 2 or memory 4 to the other computer device 1.

[0040] Memory 4 is configured, for example, using Random Access Memory (RAM) and Read Only Memory (ROM), and temporarily stores programs necessary for the operation of computer device 1, as well as data acquired or generated during operation. RAM is, for example, work memory used during the operation of computer device 1. ROM stores, for example, programs for controlling computer device 1 in advance.

[0041] 2. Operation of the driver fatigue analysis system (during actual operation) Next, with reference to Figures 5 to 10, the operation of the driver fatigue analysis system 100 according to this embodiment during actual operation will be described. Figure 5 is a sequence diagram showing an example of the overall operation procedure of the driver fatigue analysis system 100 in Figure 1 in chronological order. Figure 6 is a flowchart showing an example of the operation procedure of the data acquisition flow in Figure 5 in chronological order. Figure 7 is a flowchart showing an example of the operation procedure of the judgment / notification flow in Figure 5 in chronological order. Figure 8 is a flowchart showing an example of the operation procedure of the warning / suggestion flow in Figure 5 in chronological order. Figure 9 is a flowchart showing an example of the operation procedure of the delivery plan update flow in Figure 5 in chronological order. Figure 10 is a flowchart showing an example of the operation procedure of the fatigue confirmation flow in Figure 5 in chronological order.

[0042] In Figure 5, the fatigue management system 50 repeatedly performs the series of processes shown in Figure 5 for each individual driver at predetermined time intervals in communication with the driving vehicle system 10, the delivery management system 40, the health management system 30, and the external system 20. Specifically, the driving vehicle system 10 transmits various data extracted for the target driver (see Figure 3) to the fatigue management system 50 (St1a). The external system 20 transmits various data extracted to match the delivery date and delivery area of ​​the target driver (see Figure 3) to the fatigue management system 50 (St1b). The delivery management system 40 transmits various data extracted for the target driver (see Figure 3) to the fatigue management system 50 (St1c).

[0043] The fatigue management system 50 executes a data acquisition flow process (St2) that organizes linked data sets (datasets) for the same driver based on the various data sent in steps St1a, St1b, and St1c. Details of the process in step St2 will be described later with reference to Figure 6.

[0044] The fatigue management system 50 uses the dataset organized in step St2 and linked to the same driver to perform a judgment and notification flow process to predict the driver's current fatigue level (St3). Details of the process in step St3 will be described later with reference to Figure 7. The fatigue management system 50 generates the predicted result of the driver's current fatigue level predicted in step St3 and various notification contents for the driving vehicle system 10 related to that prediction result and feeds them back to the driving vehicle system 10 (St4a). Similarly, the fatigue management system 50 generates the predicted result of the driver's current fatigue level predicted in step St3 and various notification contents for the delivery management system 40 related to that prediction result and feeds them back to the delivery management system 40 (St4b). The fatigue management system 50 generates the predicted result of the driver's current fatigue level predicted in step St3 and various notification contents for the health management system 30 related to that prediction result and feeds them back to the health management system 30 (St4c).

[0045] Based on the feedback received in step St4a, the driving vehicle system 10 executes a warning / suggestion flow process to issue warnings and suggestions regarding the driving operations of the single driver in question (St5). Details of the process in step St5 will be described later with reference to Figure 8. The driving vehicle system 10 notifies the delivery management system 40 of the results of the process in step St5 (St6). Note that the process in step St6 may be included as part of the process in step St5 (see Figure 8).

[0046] Based on the information received in step St6, the delivery management system 40 executes a delivery plan update flow process (St7) to revise and update the driving tasks (e.g., delivery plans) assigned to the driver in question. Details of the process in step St7 will be described later with reference to Figure 9. The delivery management system 40 notifies the driving vehicle system 10 of the results of the process in step St7 (St8). Note that the process in step St8 may be included as part of the process in step St7 (see Figure 9).

[0047] Based on the feedback received in step St4c, the health management system 30 executes a fatigue level confirmation flow process to check the fatigue level of the driver in question (St9). Details of the process in step St9 will be described later with reference to Figure 10. The health management system 30 notifies the fatigue level management system 50 of the results of the process in step St9 (for example, corrective instructions that show concern for the driver's physical condition) (St10). Note that the process in step St10 may be included as part of the process in step St9 (see Figure 8).

[0048] In Figure 6, the processor 2 of the fatigue management system 50 (see Figure 2) acquires data about the driver extracted by the driver monitoring system 11 and the vital sensor 14, respectively (see Figure 3) (St11). The processor 2 of the fatigue management system 50 (see Figure 2) acquires data showing weather information 21 and traffic information 22 extracted by an external system 20 (e.g., a cloud computer) (see Figure 3) (St12). The processor 2 of the fatigue management system 50 (see Figure 2) acquires data showing delivery management information 42 related to the driver's driving duties extracted by the delivery management system 40 (see Figure 3) (St13). The processor 2 of the fatigue management system 50 (see Figure 2) stores each of the data acquired in each of steps St11 to St13 in the data storage unit 53, linking it to the same driver information (St14).

[0049] In Figure 7, the processor 2 of the fatigue management system 50 (see Figure 2) selects all or some of the data about the driver obtained in the data acquisition flow of step St2 (see Figure 6) (St21). The processor 2 of the fatigue management system 50 (see Figure 2) inputs the data selected in step St21 into the driver fatigue analysis AI 54, and the driver fatigue analysis AI 54 infers (predicts) the fatigue level of the target driver (St22). The processor 2 of the fatigue management system 50 (see Figure 2) notifies the driver vehicle system 10 (i.e., the driver), the delivery management system 40 (i.e., the transportation company), and the health management system 30 (i.e., the medical personnel) of the predicted driver fatigue level obtained in step St22 (St23).

[0050] In Figure 8, the processor 2 of the driving vehicle system 10 (see Figure 2) determines in step St4a whether the notification content (e.g., the driver's fatigue level prediction result) fed back from the fatigue management system 50 exceeds a predetermined threshold X (St31). If it is determined that the driver's fatigue level prediction result does not exceed the predetermined threshold X (St31, NO), the processing of the driving vehicle system 10 shown in Figure 8 is terminated.

[0051] On the other hand, if the processor 2 of the driving vehicle system 10 (see Figure 2) determines that the predicted driver fatigue level exceeds a predetermined threshold X (St31, YES), it notifies the driver operating the vehicle VC1 via the HMI 15 that the fatigue level exceeds the predetermined threshold X (St32). The predetermined threshold X is a preset value that has been determined by medical professionals to be a value that is likely to have adverse physical effects if the driving work continues beyond this point, and has been set in the driving vehicle system 10 (e.g., memory 4 of computer device 1). Furthermore, the processor 2 of the driving vehicle system 10 (see Figure 2) makes suggestions to the driver operating the vehicle VC1 via the HMI 15 according to the current fatigue level (e.g., take a break, reduce speed, etc.) (St33). The processor 2 of the driving vehicle system 10 (see Figure 2) detects a response (e.g., input operation) from the driver indicating acceptance or acknowledgment of the suggestion according to the current fatigue level displayed on the HMI 15 (St34). The processor 2 of the driving vehicle system 10 (see Figure 2) notifies (transmits) the content of the warning in step St32, the content of the suggestion in step St33, and the content of the response detected in step St34 to the delivery management system 40 (St35).

[0052] In Figure 9, the processor 2 of the delivery management system 40 (see Figure 2) determines in step St4b whether the notification received as feedback from the fatigue management system 50 (for example, the predicted driver fatigue level) exceeds a predetermined threshold X (St41). If it is determined that the predicted driver fatigue level does not exceed the predetermined threshold X (St41, NO), the processing of the delivery management system 40 shown in Figure 9 is terminated.

[0053] Meanwhile, if the processor 2 of the delivery management system 40 (see Figure 2) determines that the predicted driver fatigue level exceeds a predetermined threshold X (St41, YES), it waits for a predetermined time for a notification from the vehicle driving system 10 (see, for example, step St35) (St42). Based on the notification from the vehicle driving system 10 received while waiting in step St42 (see, for example, step St35), the processor 2 of the delivery management system 40 (see Figure 2) detects an input for reviewing (correcting) the current or future transportation plan (delivery plan) by an operator of the transportation company (e.g., the driver's supervisor) (St43). The processor 2 of the delivery management system 40 (see Figure 2) notifies the vehicle driving system 10 of the input result detected in step St43 (i.e., the result of reviewing the current or future delivery plan) (St44).

[0054] In Figure 10, the processor 2 of the health management system 30 (see Figure 2) determines in step St4c whether the notification content (e.g., the predicted driver fatigue level) fed back from the fatigue management system 50 exceeds a predetermined threshold Y (< threshold X) (St51). Here, the threshold Y used by the health management system 30 is smaller than the threshold X used by other systems (specifically, the driving vehicle system 10 and the delivery management system 40). This allows healthcare professionals to more appropriately correct any potential misjudgments in the fatigue level inference (prediction) by the driving fatigue analysis AI 54. If it is determined that the predicted driver fatigue level does not exceed the predetermined threshold Y (St51, NO), the processing of the driving vehicle system 10 shown in Figure 10 is terminated.

[0055] If the processor 2 of the health management system 30 (see Figure 2) determines in step St4c that the notification received as feedback from the fatigue management system 50 (for example, the predicted result of the driver's fatigue level) exceeds a predetermined threshold Y (< threshold X) (St51, YES), it waits for input from a medical professional (St52). Specifically, the medical professional verifies whether the predicted result of the driver's fatigue level and the data used for the prediction (for example, the data selected in step St21) are correct, and inputs the verification results into the health management system 30. If, as a result of this input, there is a problem with the predicted result of the driver's fatigue level (for example, there was an error in some of the data selected in step St21) (St53, YES), the processor 2 of the health management system 30 (see Figure 2) sends an instruction to the fatigue management system 50 to correct (modify) the predicted fatigue level by a specified value based on the medical professional's actions (St54). Furthermore, if there are no problems with the predicted driver fatigue level as a result of the input (for example, there are no errors in the data selected in step St21) (St53, NO), the processing of the driving vehicle system 10 shown in Figure 10 will be terminated.

[0056] The processor 2 of the fatigue management system 50 (see Figure 2) performs a process to correct the fatigue level of the relevant driver by a specified value based on the instructions sent from the health management system 30 in step St54 (St55). Preferably, this specified value is a value specified by a medical professional. If the corrected fatigue level is sent directly from the health management system 30 to the fatigue management system 50, the fatigue management system 50 may use the corrected fatigue level sent from the health management system 30 without modification. Furthermore, if the fatigue levels of a certain number of drivers have been corrected, the processor 2 of the fatigue management system 50 (see Figure 2) may use multiple corrected fatigue levels based on the instructions sent from the health management system 30 as training data and perform additional machine learning to update the driving fatigue analysis AI corresponding to that driver (St55).

[0057] 3. Configuration of the driver fatigue analysis system (when generating the driver fatigue analysis AI) Next, referring to Figure 11, the configuration of the driver fatigue analysis AI 54 of the driver fatigue analysis system 100 according to this embodiment will be described. The generation of the driver fatigue analysis AI 54 refers to the operation of generating the driver fatigue analysis AI 54 for each driver using machine learning with multiple training data. Figure 11 is a diagram showing an example of the system configuration during the training of the driver fatigue analysis AI 54 of the driver fatigue analysis system 100 according to this embodiment. In the configuration of the driver fatigue analysis system 100 shown in Figure 11, parts that overlap with the configuration of the driver fatigue analysis system 100 shown in Figure 1 are given the same reference numerals, simplifying or omitting the explanation, while different parts are explained.

[0058] The only structural difference between the driver fatigue analysis system 100 shown in Figure 11 and the driver fatigue analysis system 100 shown in Figure 1 is the fatigue management system 50. In other words, the configuration and operation of the driving vehicle system 10, external system 20, health management system 30, and delivery management system 40 are virtually identical between the driver fatigue analysis system 100 shown in Figure 11 and the driver fatigue analysis system 100 shown in Figure 1. However, processing by the driver fatigue analysis system 100 shown in Figure 11 begins before processing by the driver fatigue analysis system 100 shown in Figure 1 begins during actual operation.

[0059] In the driver fatigue analysis system 100 shown in Figure 11, the fatigue management system 50 is similarly configured as a computer device 1 (see Figure 2) and includes at least a data acquisition unit 51, a driver information management unit 52, a data storage unit 53, a driver fatigue analysis AI 54, a data processing unit 58, annotated training data 59, and an AI generation and update unit 60. The data processing unit 58 and the AI ​​generation and update unit 60 are configured, for example, as a processor 2 of the computer device 1, or a combination of the processor 2 and a communication interface circuit 3.

[0060] The health management system 30 acquires various data associated with the driver stored in the data storage unit 53 of the fatigue management system 50, and adds annotations of medical opinions from healthcare professionals regarding the driver's physical condition, etc., based on these various data. The various data are data sent from the driving vehicle system 10, the external system 20, and the delivery management system 40 (see Figure 3). The health management system 30 sends the added annotation data (i.e., the content of the current medical opinion regarding the driver) to the fatigue management system 50.

[0061] The data processing unit 58 acquires annotation data corresponding to drivers sent from the health management system 30 and generates annotated training data 59 by processing the data to add annotation data to various data associated with the driver (see above). This annotated training data 59 includes not only various data associated with the driver (i.e., data sent from the driving vehicle system 10, the external system 20, and the delivery management system 40 (see Figure 3)), but also annotation data sent from the health management system 30 for those various data. The data processing unit 58 stores the annotated training data 59 obtained through processing in the data storage unit 53.

[0062] Here, the annotation data includes not only the driver's fatigue level, which is determined by the medical professional's judgment based on their experience, etc., from various data linked to the driver (i.e., data sent from the driving vehicle system 10, the external system 20, and the delivery management system 40 (see Figure 3)), but also various parameter values ​​used to calculate that fatigue level. The annotation data may be added manually by the medical professional, or it may be added automatically based on a mechanical algorithm executable in the health management system 30, and then manually reviewed, corrected, or added to by the medical professional.

[0063] The AI ​​generation and update unit 60 generates or updates a driver fatigue analysis AI 54 for each driver by performing machine learning using annotated learning data 59 linked to drivers stored in the data storage unit 53 as input. The AI ​​generation and update unit 60 stores the driver fatigue analysis AI 54 in the data storage unit 53.

[0064] 4. Operation of the driver fatigue analysis system (when generating the driver fatigue analysis AI) Next, with reference to Figures 12 to 14, the operation of the driver fatigue analysis system 100 according to this embodiment during the generation of the driver fatigue analysis AI 54 will be explained. Figure 12 is a sequence diagram showing a time-series example of the overall operation procedure of the driver fatigue analysis system 100 in Figure 11. Figure 13 is a flowchart showing a time-series example of the operation procedure of the annotation addition flow in Figure 12. Figure 14 is a flowchart showing a time-series example of the operation procedure of the driver fatigue analysis AI generation flow in Figure 12.

[0065] In Figure 12, the fatigue management system 50 repeatedly performs the series of processes shown in Figure 12 for each individual driver at predetermined time intervals in communication with the driving vehicle system 10, the delivery management system 40, and the external system 20. Specifically, the driving vehicle system 10 transmits various data extracted for the target driver (see Figure 3) to the fatigue management system 50 (St61a). The external system 20 transmits various data extracted to match the delivery date and delivery area of ​​the target driver (see Figure 3) to the fatigue management system 50 (St61b). The delivery management system 40 transmits various data extracted for the target driver (see Figure 3) to the fatigue management system 50 (St61c).

[0066] The fatigue management system 50 executes a data acquisition flow process (St62) that organizes linked data sets (datasets) for the same driver based on the various data sent in steps St61a, St61b, and St61c, respectively. The process in step St62 is the same as the flow shown in Figure 6, so its explanation is omitted here.

[0067] After step St62, the health management system 30 repeatedly performs the series of processes shown in Figure 12 for each individual driver at predetermined time intervals in cooperation with the fatigue management system 50.

[0068] Specifically, the health management system 30 references and acquires various data obtained from the fatigue management system 50 in the data acquisition flow of step St62 (St63), and executes an annotation addition flow to add medical opinion results (annotation) for one driver (St64). Details of the process in step St64 will be described later with reference to Figure 13.

[0069] The health management system 30 sends the driver annotation data added in step St64 to the fatigue management system 50 (St65). The fatigue management system 50 generates annotated training data 59 (see Figure 11) with the annotation data sent from the health management system 30 in step St65 added. The fatigue management system 50 uses the annotated training data 59 to execute the processing of the driving fatigue analysis AI generation flow (St66).

[0070] In Figure 13, the processor 2 of the health management system 30 (see Figure 2) selects all or part of the various data about the driver obtained in the data acquisition flow of step St62 at the operation of a medical professional (St71). The processor 2 of the health management system 30 (see Figure 2) adds annotation data based on the expert knowledge data 31 and driver health-related data 32 stored in the health management system 30, which were added by the medical professional, to the data selected in step St71 (St72). The processor 2 of the health management system 30 (see Figure 2) notifies the fatigue management system 50 of the completion of annotation addition, including the annotation data added in step St72 (St73).

[0071] In Figure 14, the processor 2 of the fatigue management system 50 (see Figure 2) receives and acquires the annotation completion report sent from the health management system 30 in step St73 (St81). Based on the annotation completion report acquired in step St81, the processor 2 of the fatigue management system 50 (see Figure 2) generates annotated training data 59. When the processor 2 of the fatigue management system 50 (see Figure 2) has accumulated N (N: an integer greater than or equal to 2) annotated training data 59 over a certain period of time, it determines the fatigue level corresponding to the annotated training data 59 at each timing (for example, time T, T+1, T+2, ..., T+N) (St82). The processor 2 of the fatigue management system 50 (see Figure 2) performs machine learning using the annotated training data 59 at each timing so that the fatigue level at each timing determined in step St82 becomes the output for the annotated training data 59 as input (St83). As a result, the processor 2 (see Figure 2) of the fatigue management system 50 can generate a driving fatigue analysis AI 54 that is specific to the target driver.

[0072] (Summary of this disclosure) The above description of embodiments discloses the technical concepts corresponding to the following items.

[0073] (Item 1) The driver fatigue analysis method according to this disclosure acquires first data showing environmental information of the driver while driving, measured by sensors (gyro sensor 12, in-vehicle camera 13, vital sensor 14) of a vehicle (VC1) capable of delivering goods, acquires second data showing information of the driver's daily delivery plan using the vehicle, and uses a driver fatigue prediction model (driving fatigue analysis AI 54) that has been machine-trained to predict the driver's fatigue level using multiple pairs of the first and second data obtained at different timings as training data, and predicts the driver's fatigue level based on the first and second data acquired in relation to the driver, and notifies the driver of the predicted fatigue level. Therefore, according to the driver fatigue analysis method, it is possible to predict the current fatigue level of a driver with high accuracy by obtaining at least environmental information during driving and the daily delivery plan included in the driving work, and then using a fatigue prediction model that has been pre-trained on each driver. Consequently, the driver fatigue analysis method can more appropriately detect the fatigue level of drivers.

[0074] (Item 2) In the driver fatigue analysis method described in item 1, the predicted driver fatigue level is notified to a medical terminal (health management system 30) that stores driver health-related data (32) indicating the health status of each driver. When the driver fatigue level is corrected based on the driver health-related data by a medical professional using the medical terminal, the corrected driver fatigue level is obtained and stored in memory (data storage unit 53). Therefore, according to the driver fatigue analysis method, if the driver's fatigue level is notified to a medical terminal and corrected by a medical professional, the corrected driver's fatigue level can be obtained, thus providing an appropriate driver fatigue level.

[0075] (Item 3) In the prediction method described in item 2, the corrected fatigue level of the driver is notified to the delivery management terminal (delivery management system 40) of the transportation company where the driver works. As a result, according to the driver fatigue analysis method, the corrected driver fatigue level is shared with the delivery management terminal of the transportation company where the driver works, enabling the transportation company to appropriately consider the driver's physical condition when creating delivery plans.

[0076] (Item 4) In the driver fatigue analysis method described in any one of items 1 to 3, the predicted fatigue level of the driver is notified to the delivery management terminal (delivery management system 40) of the transportation company where the driver works, and when the second data is modified based on the driver's fatigue level by an operator using the delivery management terminal, the modified second data is acquired and stored in memory (data storage unit 53). As a result, according to the driver fatigue analysis method, the predicted driver fatigue level is shared with the delivery management terminal of the transportation company where the driver works. Therefore, if the transportation company is able to appropriately revise the delivery plan considering the driver's physical condition, it will be possible to obtain the revised delivery plan.

[0077] (Item 5) In the driver fatigue analysis method described in any one of items 1 to 4, a third data is obtained that shows at least one of the weather information and traffic information for the day of delivery of the cargo. The driver fatigue prediction model is generated by machine learning using multiple pairs of the first data, second data, and third data obtained at different times as training data. The driver fatigue is predicted based on the driver fatigue prediction model and the first data, second data, and third data obtained on the same day corresponding to the driver. This means that, according to the driver fatigue analysis method, the AI ​​for driver fatigue analysis can be trained on machine learning while also taking into account a third data point, which is at least one of the weather information and traffic information on the day of delivery, thereby improving the accuracy of predicting driver fatigue.

[0078] (Item 6) In the driver fatigue analysis method described in any one of items 1 to 5, the sensor includes at least one of a gyro sensor (12), an in-vehicle camera (13), and a vital sensor (14) worn by the driver, and the first data includes at least one of data relating to the rotation, orientation, and vibration of the vehicle measured by the gyro sensor, image data of the driver captured by the in-vehicle camera, and biological data of the driver measured by the vital sensor. As a result, according to the driver fatigue analysis method, various data can be acquired that show the driver's environmental conditions while driving or the driver's biometric information, measured by at least one of the gyro sensor, in-vehicle camera, and vital sensor. Using this data, the driver's current level of fatigue can be appropriately predicted.

[0079] (Item 7) In the driver fatigue analysis method described in item 2, the first data and the second data are transmitted to the medical terminal, and the fatigue prediction model is generated or updated by machine learning when a fourth data (annotated training data 59) that suggests the driver's health status based on a medical diagnosis by the medical professional is added by the medical professional through an operation of the medical professional, and the fourth data added in accordance with the driver is included in the training data. As a result, according to the driving fatigue analysis method, if a fourth data point indicating the driver's health status based on a medical diagnosis by a healthcare professional is added by the healthcare professional, the driving fatigue analysis AI can be generated taking this added data into account, enabling a more accurate prediction of the driver's fatigue level.

[0080] (Item 8) In the driver fatigue analysis method described in item 3, if the predicted driver's fatigue level is above a predetermined value, the delivery management terminal is notified of an instruction to revise the driver's daily delivery plan using the vehicle. This means that, according to the driver fatigue analysis method, if a driver's fatigue level exceeds a predetermined value, the transportation company employing that driver can be warned and urged to revise the day's delivery plan.

[0081] (Item 9) In the driver fatigue analysis method described in any one of items 1 to 8, if the predicted driver fatigue level is above a predetermined value, an output device (HMI15) mounted on the vehicle is notified of an improvement suggestion to reduce at least one driver fatigue level corresponding to that fatigue level. This allows the driver fatigue analysis method to explicitly show the driver, when their fatigue level exceeds a predetermined value, suggestions for improvement (such as encouraging them to take a break) that may alleviate their fatigue, in accordance with their fatigue level.

[0082] (Item 10) The driver fatigue analysis device (fatigue management system 50) according to this disclosure comprises a processor and a memory, wherein the processor, in cooperation with the memory, acquires first data indicating environmental information of the driver while driving, measured by sensors (gyro sensor 12, in-vehicle camera 13, vital sensor 14) of a vehicle (VC1) capable of delivering goods, acquires second data indicating information of the driver's daily delivery plan using the vehicle, and predicts the driver's fatigue level based on a driver fatigue prediction model (driving fatigue analysis AI 54) that has been machine-trained to predict the driver's fatigue level using a plurality of pairs of the first data and the second data obtained at different timings as training data, and the first data and the second data acquired in relation to the driver, and notifies the driver of the predicted fatigue level. As a result, the driver fatigue analysis device can accurately predict the driver's current fatigue level by acquiring at least environmental information during driving and the daily delivery plan included in the driving work, and then using a fatigue prediction model that has been pre-trained for each driver. Therefore, the driver fatigue analysis device can more appropriately detect the driver's fatigue level.

[0083] While embodiments have been described above with reference to the attached drawings, this disclosure is not limited to such examples. It is clear to those skilled in the art that various modifications, alterations, substitutions, additions, deletions, and equivalents can be conceived within the scope of the claims, and these are also understood to fall within the technical scope of this disclosure. Furthermore, the components of the embodiments described above can be combined in any way without departing from the spirit of the invention. [Industrial applicability]

[0084] The technology disclosed herein is useful as a driving fatigue analysis method and a driving fatigue analysis device for more appropriately detecting the degree of driver fatigue. [Explanation of symbols]

[0085] 1. Computer device 2 processors 3. Communication Interface Circuit 4 memory 5. Internal bus 10. Driving Vehicle System 11. Driver Monitoring System 12 Gyro Sensors 13 In-car camera 14 Vital Sensors 15 HMI 20 External Systems 21 Weather Information 22 Traffic information 30 Health Management Systems 31 Expert Knowledge Data 32. Driver health-related data 40 Delivery Management System 41 Delivery Planning and Execution Management Department 42 Delivery management information 50. Fatigue Management System (Driving Fatigue Analysis Device) 51 Data Acquisition Unit 52 Driver Information Management Department 53 Data Storage Unit 54. AI for analyzing driver fatigue levels 55 Fatigue level prediction unit 56 Notification Department 57 Fatigue Correction Unit 58 Data Processing Department 59 Annotated training data 60 AI generation update section 100 Driving Fatigue Analysis System TBL1 Data Table TBL2 Feedback Table

Claims

1. First data is obtained showing the driver's environmental information while driving, measured by sensors in a vehicle capable of delivering cargo. Second data is obtained showing information on the driver's daily delivery plan using the vehicle. A driver fatigue prediction model, which has been trained to predict the driver's fatigue level using multiple pairs of acquired first and second data as training data, and the first and second data acquired in relation to the driver, are used to predict the driver's fatigue level. Notify the predicted level of fatigue of the driver. Method for analyzing driver fatigue levels.

2. The predicted level of fatigue of the driver is notified to a medical terminal that stores driver health-related data indicating the health status of each driver. When the driver's fatigue level is corrected based on the driver's health-related data by a medical professional using the medical terminal, the corrected driver's fatigue level is acquired and stored in memory. The method for analyzing driver fatigue according to claim 1.

3. The corrected fatigue level of the driver is notified to the delivery management terminal of the transportation company where the driver works. The method for analyzing driver fatigue according to claim 2.

4. The predicted fatigue level of the driver is notified to the delivery management terminal of the transportation company where the driver works. If the second data is modified by an administrator using the delivery management terminal based on the driver's fatigue level, the modified second data is acquired and stored in memory. The method for analyzing driver fatigue according to claim 1.

5. A third data set is obtained that shows at least one of the weather information and traffic information for the delivery date of the aforementioned package. The driver fatigue prediction model is generated by machine learning using multiple pairs of the acquired first data, second data, and third data as training data. Based on the driver-specific fatigue prediction model and the first, second, and third data acquired on the same day corresponding to the driver, the driver's fatigue level is predicted. The method for analyzing driver fatigue according to claim 1.

6. The sensor includes at least one of a gyro sensor, an in-vehicle camera, and a vital sensor worn by the driver. The first data includes at least one of the following: data relating to the rotation, orientation, and vibration of the vehicle measured by the gyro sensor; image data of the driver captured by the in-vehicle camera; and biometric data of the driver measured by the vital sensor. The method for analyzing driver fatigue according to claim 1.

7. The first data and the second data are transmitted to the medical terminal. The fatigue prediction model is generated or updated by machine learning when a fourth data indicating the driver's health status based on a medical diagnosis by the medical professional is added by the medical professional, corresponding to the driver, and this fourth data is included in the training data. The method for analyzing driver fatigue according to claim 2.

8. If the predicted fatigue level of the driver exceeds a predetermined value, the delivery management terminal is notified of an instruction to revise the driver's daily delivery plan using the vehicle. The method for analyzing driver fatigue according to claim 3.

9. If the predicted driver fatigue level is above a predetermined value, the system notifies the output device mounted on the vehicle of improvement suggestions to reduce the fatigue of at least one driver, corresponding to the fatigue level. The method for analyzing driver fatigue according to claim 1.

10. Equipped with a processor and memory, The aforementioned processor, in cooperation with the memory, First data is obtained showing the driver's environmental information while driving, measured by sensors in a vehicle capable of delivering cargo. Second data is obtained showing information on the driver's daily delivery plan using the vehicle. A driver fatigue prediction model, which has been trained to predict the driver's fatigue level using multiple pairs of acquired first and second data as training data, and the first and second data acquired in relation to the driver, are used to predict the driver's fatigue level. Notify the predicted level of fatigue of the driver. Driver fatigue analysis device.

Images