Fatigue state determination system and fatigue state determination method

Abstract

A fatigue state determination system (1) comprises: a means for acquiring time-series sensor data measured by a wearable sensor worn by a user; a means for acquiring user input data input by the user; and a determination means for quantitatively determining a mental fatigue state of the user on the basis of the sensor data and the user input data. The determination means includes a means for calculating a transient fatigue degree (E2), a chronic fatigue degree (E1), and a fatigue allowance degree (E3) as quantitative information of the mental fatigue state.

Description

Fatigue state determination system and fatigue state determination method

[0001] The present invention relates to a technique for determining a person's mental fatigue state.

[0002] There are technologies for determining a user's stress level based on biological data measured using a wearable sensor attached to the user's body. For example, Patent Document 1 listed below discloses a method in which a medical wearable device attached to a patient's body is provided with a pulse wave sensor, a body movement sensor (acceleration sensor), a body temperature sensor, and a skin potential sensor, and stress information for the patient is generated based on one or more of the pulse wave data, body movement data, body temperature data, and skin potential data obtained from the respective sensors.

[0003] Patent Document 2 listed below discloses a method for acquiring biological data such as a subject's heart rate, heart rate waveform, body temperature, audio output level, blood pressure, pulse rate, etc. from a wearable biological sensor, and generating information for visualizing the subject's stress level based on heart rate variability parameters obtained from the biological data. Furthermore, Patent Document 3 listed below discloses a method for estimating a subject's stress based on the subject's acceleration information, and Patent Document 4 listed below discloses a method for determining a subject's stress using the subject's pulse rate.

[0004] JP 2023-5063 A JP 2023-65808 A Japanese Patent No. 7136341 A Japanese Patent No. 7255670 A

[0005] We sometimes feel fatigued due to our daily lives, work, etc. Fatigue can be mainly divided into physical fatigue and mental fatigue, and accumulation of fatigue can lead to mental and physical abnormalities such as illness. A stressful state is a state in which external stimuli cause strain or tension in the mind and body, and it is known that a stressful state delays recovery from fatigue and is one of the factors that cause fatigue (mainly mental fatigue). As such, fatigue and stress are closely related but not completely identical, so it would be useful to visualize a person's fatigue state in addition to existing stress assessment methods.

[0006] The present invention has been made in view of the above circumstances, and provides a technique for visualizing a person's mental fatigue state in an easily understandable manner.

[0007] According to the present invention, a fatigue state determination system is provided, comprising a first acquisition means for acquiring time-series sensor data measured by a wearable sensor worn by a user, a second acquisition means for acquiring user input data entered by the user, and a determination means for quantitatively determining the user's mental fatigue state based on the sensor data and the user input data, wherein the determination means includes a calculation means for calculating transient fatigue level, chronic fatigue level, and fatigue margin level as quantitative information of the mental fatigue state.

[0008] Furthermore, according to the present invention, there is provided a fatigue state determination method that can be executed in a system having one or more processors and memory, the fatigue state determination method including the steps of: acquiring time-series sensor data measured by a wearable sensor worn by a user; acquiring user input data entered by the user; and judging the user's mental fatigue state quantitatively based on the sensor data and the user input data, wherein the judgment step includes a calculation step of calculating transient fatigue level, chronic fatigue level, and fatigue margin level as quantitative information of the mental fatigue state.

[0009] The present invention also provides a computer program for causing one or more computers equipped with a processor and memory to execute the fatigue state determination method, and a computer-readable storage medium storing such a computer program. This storage medium includes a non-transitory tangible medium.

[0010] According to the present invention, it is possible to provide a technique for visualizing a person's mental fatigue state in an easily understandable manner.

[0011] FIG. 1 is a diagram conceptually showing an example of the hardware configuration of a fatigue state determination system (this system) according to this embodiment; FIG. 2 is a diagram conceptually showing an example of the software configuration of a fatigue state determination device (this determination device) according to this embodiment; FIG. 3 is a diagram conceptually showing a method for calculating chronic fatigue level, transient fatigue level, and fatigue margin in this embodiment; FIG. 4 is a diagram illustrating an example of a display screen presenting transient fatigue level, chronic fatigue level, and fatigue margin; FIG. 5 is a diagram illustrating an example of a display screen presenting temporal changes in transient fatigue level, chronic fatigue level, and fatigue margin; FIG. 6 is a diagram illustrating an example of a user notification screen; and FIG. 7 is a flowchart showing a fatigue state determination method (this determination method) according to this embodiment.

[0012] Hereinafter, embodiments of the present invention (hereinafter, sometimes referred to as the present embodiment) will be described. Note that the following embodiments are merely examples, and the present invention is not limited to the configurations of the following embodiments.

[0013] [Overview] First, an overview of the fatigue state determination system and fatigue state determination method according to the present embodiment will be described. In the present embodiment, time-series sensor data measured by a wearable sensor worn by a user is acquired, user input data input by the user is acquired, and the user's mental fatigue state is quantitatively determined based on the acquired sensor data and user input data.

[0014] A wearable sensor is attached to a user's body and includes one or more sensor devices that measure one or more types of information related to the user's body. For example, the wearable sensor may include one or more sensor devices selected from the group consisting of a temperature sensor, a pulse wave sensor, an electrodermal activity (EDA) sensor, an acceleration sensor, and other biosensors. In this embodiment, the wearable sensor may include one or more sensor devices that can measure information related to a person's mental fatigue state, and the type and number of sensor devices that make up the wearable sensor are not limited.

[0015] The sensor data measured by such a wearable sensor indicates various information measured depending on the configuration of the wearable sensor. The first acquisition means for acquiring the sensor data may be a wearable device equipped with a wearable sensor, as in the specific example described below, or may be a device (the fatigue state assessment device 10 described below) that indirectly acquires the sensor data via wireless communication from the wearable device via a user terminal. The method for acquiring the sensor data is also not limited. The sensor data may be acquired directly based on a sensor signal from the wearable device, or may be acquired indirectly via communication from a wearable device equipped with a wearable sensor via another computer, as in the specific example described below.

[0016] The user input data is data input by the target user wearing the wearable sensor. For example, the target user's responses to a questionnaire about psychological states such as motivation and frustration, or stress levels, are acquired as the user input data. The user input data may include information related to the target user's psychological state, and the specific content is not limited. Furthermore, the method for acquiring the user input data is not limited in any way. The user input data may be acquired through the target user's own input operation on a questionnaire screen, or may be generated by a third party, such as an operator, entering information related to the target user's psychological state that is obtained from the target user in some way.

[0017] In this embodiment, the quantitative information of the user's mental fatigue state obtained by the assessment is calculated as a transient fatigue level, a chronic fatigue level, and a fatigue margin. The transient fatigue level indicates the level of mental fatigue that can be recovered by resting (hereinafter referred to as transient fatigue), while the chronic fatigue level indicates the level of mental fatigue that is difficult to recover from even by resting (hereinafter referred to as chronic fatigue). The fatigue margin indicates the level of mental fatigue margin until the mental fatigue level reaches its maximum, i.e., the level of mental fatigue that is still tolerable even when transient fatigue and chronic fatigue are combined. The fatigue margin can be determined, for example, from the relative relationship between the transient fatigue level and the chronic fatigue level through a sensory evaluation using a population of multiple subjects. The transient fatigue level, chronic fatigue level, and fatigue margin may each be expressed as a percentage with the total being 100%, or may be expressed as a ratio to a set mental fatigue margin, or may be expressed as a value such as an aggregate score, as described below.

[0018] In this manner, in this embodiment, the user's mental fatigue state is quantified using three index values, namely, transient fatigue level, chronic fatigue level, and fatigue margin level, based on time-series sensor data and user input data. Therefore, this embodiment makes it possible to visualize a person's mental fatigue state in an easy-to-understand manner.

[0019] Hereinafter, a fatigue state determination system (hereinafter sometimes abbreviated as this system) and a fatigue state determination method (hereinafter sometimes abbreviated as this determination method) according to this embodiment will be described in detail.

[0020] [System Details] <<Hardware Configuration>> Fig. 1 is a diagram conceptually illustrating an example of the hardware configuration of a fatigue state determination system (this system) 1 according to this embodiment. This system 1 includes a fatigue state determination device 10, a user terminal 5, and a wearable device 20. The user terminal 5 and the wearable device 20 are used by the same user, and this system 1 determines the mental fatigue state of this user. The fatigue state determination device 10 and the user terminal 5 are connected to each other via a communication network 3 so as to be able to communicate with each other, and the user terminal 5 is connected to the wearable device 20 so as to be able to communicate with each other via short-range wireless communication such as Bluetooth (registered trademark) or Wi-Fi.

[0021] The communication network 3 is composed of one or more of a public network such as the Internet, a wide area network (WAN), a local area network (LAN), a wireless communication network, etc. However, in the present embodiment, the communication form between the determination device 10 and the user terminal 5 and between the user terminal 5 and the wearable device 20 is not limited. Furthermore, the wearable device 20 may be communicatively connected to the fatigue state determination device 10 via the communication network 3, with or without being communicatively connected to the user terminal 5.

[0022] The fatigue state determination device (hereinafter sometimes abbreviated as the determination device) 10 is a so-called computer, and includes a processor 11, a memory 12, an input / output interface (I / F) 13, a communication unit 14, etc. The processor 11 is composed of one or more of a CPU (Central Processing Unit), an MPU (Micro Processing Unit), an application specific integrated circuit (ASIC), a DSP (Digital Signal Processor), a GPU (Graphics Processing Unit), etc. The memory 12 is a RAM (Random Access Memory), a ROM (Read Only Memory), or an auxiliary storage device (such as a hard disk).

[0023] The input / output I / F 13 can be connected to user interface devices such as a display device 15 and an input device 16. The display device 15 is a device, such as an LCD (Liquid Crystal Display) or a CRT (Cathode Ray Tube) display, that displays a screen corresponding to drawing data processed by the processor 11 or the like. The input device 16 is a device, such as a keyboard or mouse, that accepts user input. However, the present determination device 10 does not necessarily have to include the display device 15 or the input device 16. The communication unit 14 communicates with other computers via the communication network 3 and exchanges signals with other devices such as printers. A portable recording medium or the like can also be connected to the communication unit 14. In this embodiment, the communication unit 14 is communicably connected to the user terminal 5 or the like via the communication network 3.

[0024] The hardware configuration of the present determination device 10 is not limited to the example in Fig. 1. The present determination device 10 may include other hardware components not shown. Furthermore, the number of each hardware component is not limited to the example in Fig. 1. For example, the present determination device 10 may have multiple processors 11. Furthermore, the present determination device 10 may be realized by multiple computers each having multiple housings.

[0025] The wearable device 20 is a device that can be attached to a specific part of a person's body, such as the head, arm, leg, or waist. The wearable device 20 includes a processor 21, a memory 22, an input / output I / F 23, a communication unit 24, a sensor 25, etc. The processor 21, memory 22, input / output I / F 23, and communication unit 24 are as described for the processor 11, memory 12, input / output I / F 13, and communication unit 14 of the present determination device 10, but differ from those of the present determination device 10 in the following points.

[0026] The input / output I / F 23 is connected to the sensor 25 and processes signals from the sensor 25. In this embodiment, the wearable device 20 does not necessarily have to include the display device 15 and the input device 16 as user interfaces, and therefore the input / output I / F 23 does not necessarily have to be connected to a user interface device. The communication unit 24 is connected to the user terminal 5 via short-range wireless communication so as to be able to communicate with the user terminal 5.

[0027] The sensor 25 corresponds to the wearable sensor described in the overview, since it is mounted on the wearable device 20 that is worn on the user's body. There is no limit to the number and types of sensors 25 that the wearable device 20 includes, but for ease of understanding, this embodiment will exemplify a case in which the wearable device 20 includes a temperature sensor, a pulse wave sensor, an electrodermal activity (EDA) sensor, and an acceleration sensor as the sensors 25.

[0028] The wearable device 20 samples the sensor signals received from the sensor 25 at a predetermined sampling period and wirelessly transmits the obtained time-series sensor data to the user terminal 5. In this embodiment, body temperature data, pulse wave data, EDA data, and acceleration data based on the sensor signals from the temperature sensor, pulse wave sensor, EDA sensor, and acceleration sensor are transmitted to the user terminal 5. There are no restrictions on the timing of transmission of sensor data from the wearable device 20 to the user terminal 5. The wearable device 20 may transmit sensor data to the user terminal 5 sequentially every time sensor data is acquired, or may hold the acquired sensor data and transmit a predetermined period of or a predetermined number of pieces of sensor data together to the user terminal 5.

[0029] The user terminal 5 is a so-called computer, and like the determination device 10, includes a processor, a memory, an input / output I / F, a communication unit, etc. The user terminal 5 is preferably a portable computer such as a smartphone or a tablet terminal, but may also be a stationary computer.

[0030] The user terminal 5 is capable of outputting information provided by the fatigue state assessment device 10 and displays a screen that serves as a user interface for the fatigue state assessment device 10. The output format of the user terminal 5 will be described later. The user terminal 5 also establishes pairing with the wearable device 20 via short-range wireless communication and receives the above-mentioned sensor data from the wearable device 20 via wireless communication. The user terminal 5 transmits the time-series sensor data received from the wearable device 20 to the assessment device 10.

[0031] The user terminal 5 is required to be capable of operating as a user interface for the fatigue state determination device 10 and to be capable of receiving sensor data from the wearable device 20 and transmitting it to the determination device 10, and there are no particular restrictions on the hardware and software configuration of the user terminal 5.

[0032] While the present system 1 can be used to assess a user's mental fatigue state in their everyday lives, the following description will use an example in which the present system 1 is used to assess the mental fatigue state of an office worker user for ease of understanding. In this example, the wearable device 20 is worn by the user at least while at work (while working), and the determination device 10 begins quantitatively assessing the user's mental fatigue state as the wearable device 20 is worn. However, the use of the present system 1 is not limited to this example.

[0033] <<Software configuration of the fatigue state determination device (this determination device)>> This determination device 10 quantitatively determines the user's mental fatigue state using the software configuration shown in Figure 2, and outputs the transient fatigue level, chronic fatigue level, fatigue margin level, etc. calculated as quantitative information to the user terminal 5.

[0034] 2 is a diagram conceptually illustrating an example of the software configuration of a fatigue state determination device (the present determination device) 10 according to this embodiment. As shown in FIG. 2, the present determination device 10 includes an acquisition unit 61, a user interface (UI) processing unit 62, a determination unit 63, a data storage unit 65, and the like. Each of these processing modules is realized, for example, by the processor 11 executing a computer program stored in the memory 12. This computer program may be installed, for example, from a portable recording medium such as a CD (Compact Disc) or a memory card, or from another computer on a network, via the input / output I / F 13 or the communication unit 14, and stored in the memory 12.

[0035] The acquisition unit 61 acquires time-series sensor data received from the user terminal 5 via the communication unit 14. The acquisition unit 61 corresponds to a first acquisition means. In this embodiment, the acquisition unit 61 acquires time-series body temperature data, pulse wave data, EDA data, and acceleration data. For example, acceleration data can also be considered data indicating the user's behavior (behavioral data), so the acquisition unit 61 corresponds to a third acquisition means for acquiring the user's time-series behavior data. Note that if the sampling period differs for each type of sensor data, the number of time-series sensor data pieces for the same time period (e.g., 3 minutes) may differ for each type of sensor data. Furthermore, the time series of sensor data that should be acquired may be interrupted due to a communication error between the user terminal 5 and the wearable device 20, for example.

[0036] The UI processing unit 62 outputs a display screen serving as a user interface to the user terminal 5 and performs processing in response to user operations on the display screen. Specifically, the UI processing unit 62, as will be described in detail later, causes the user terminal 5 to display a screen showing the transient fatigue level, chronic fatigue level, and fatigue tolerance level, a screen guiding the user to take a specific action such as taking a break, and the like. The UI processing unit 62 also causes the user terminal 5 to display a screen for prompting the user to input some information (such as responses to a questionnaire or sleep time) or for prompting the user to perform some operation (such as pressing a button to start work or start a break), and acquires information in response to the user operation on the screen. Hereinafter, the screen displayed on the user terminal 5 by the UI processing unit 62 may be collectively referred to as a UI screen. The UI processing unit 62 acquires user input data through user operations on such UI screens and therefore corresponds to second acquisition means.

[0037] The determination unit 63 quantitatively determines the mental fatigue state of the user using at least the sensor data acquired by the acquisition unit 61 and the user input data acquired by the UI processing unit 62. The determination unit 63 includes a calculation unit 70 that calculates the transient fatigue level, chronic fatigue level, and fatigue margin as quantitative information on the mental fatigue state. In other words, the determination unit 63 corresponds to the determination means, and the calculation unit 70 corresponds to the calculation means.

[0038] The calculation unit 70 includes multiple types of calculation models and a counting unit 77. The calculation models are software elements that apply specific processing for each model to input data to obtain output data, and are configured using function expressions, conditional branching, algorithms, numerical conversion, AI (Artificial Intelligence) models, etc. In this embodiment, each calculation model receives at least one of sensor data and / or user input data as input, and outputs each output value.

[0039] 3 is a diagram conceptually illustrating a method for calculating the chronic fatigue level, transient fatigue level, and fatigue margin level in this embodiment. As shown in FIG. 3, the counting unit 77 periodically calculates the transient fatigue intermediate score and the chronic fatigue intermediate score using the output scores from each calculation model, and calculates the transient fatigue summary score and the chronic fatigue summary score based on the periodically calculated transient fatigue intermediate score and the chronic fatigue intermediate score. In other words, the counting unit 77 corresponds to a counting means.

[0040] The calculation unit 70 calculates the transient fatigue level, the chronic fatigue level, and the fatigue margin as respective ratios of the transient fatigue level, the chronic fatigue level, and the fatigue margin based on the transient fatigue total score and the chronic fatigue total score calculated by the calculation unit 77. Specifically, the calculation unit 70 calculates the chronic fatigue level based on the chronic fatigue total score, calculates the transient fatigue level based on the transient fatigue total score from the ratio remaining after subtracting the chronic fatigue level ratio from the total, and calculates the fatigue margin as the ratio remaining after subtracting each of the chronic fatigue level and the transient fatigue level ratio from the total. In this embodiment, the calculation unit 70 calculates the chronic fatigue level by converting the chronic fatigue total score to an integer so that it takes a value between 0 and 100, calculates the transient fatigue level by rounding the transient fatigue total score to a value between 0 and 1 and multiplying it by the value obtained by subtracting the chronic fatigue level from the total (100), and calculates the margin of error as the remainder after subtracting the chronic fatigue level and the transient fatigue level from the total (100).

[0041] In this embodiment, the transient fatigue level, chronic fatigue level, and fatigue margin are each calculated as a ratio to their total, and the remaining chronic fatigue level is divided into the transient fatigue level and fatigue margin. Chronic fatigue is characterized by being more difficult to overcome than transient fatigue, even with rest, etc. Therefore, according to this embodiment, by calculating the chronic fatigue level based on the ratio and dividing the remainder into the transient fatigue level and fatigue margin, it is possible to quantify the mental fatigue state in accordance with the characteristics of a person's actual mental fatigue.

[0042] In this embodiment, the calculation model and the calculation method of the calculation unit 77 are set so that when the mental fatigue level is not at its maximum, the sum of the chronic fatigue level and the transient fatigue level is equal to or less than the total value (100) and the fatigue margin indicates a perceptually appropriate value. However, the "total" value, which can be said to be the tolerable amount or the maximum amount of mental fatigue, does not have to be a fixed value (100) but may be a variable that can change depending on the mental state or health state, such as the level of morale. In the latter case, the variable value indicating the total may be changed depending on the user's responses to a questionnaire (user-input data), etc. For example, when morale is at its maximum, a value above 100 is set as the total value, and when the user is feeling a bit under the weather or when hay fever is in full swing, a value below 100 is set as the total value.

[0043] In this embodiment, the calculation unit 70 includes, as calculation models, a fatigue model 71, a sleep model 72, a behavior model 73, a motivation model 74, a transient stress model 75, and a chronic stress model 76.

[0044] <Fatigue Model> The fatigue model 71 is a calculation model that estimates the degree of fatigue over time. In this embodiment, the fatigue model 71 estimates the degree of fatigue over working hours of a target user who is an office worker. In this embodiment, the wearable device 20 is worn during work, so the fatigue model 71 inputs the elapsed time from the time the wearable device 20 is worn and outputs a score ranging from 0 to 1. Because fatigue increases with the length of work hours, the fatigue model 71 outputs a larger score the longer the elapsed time. However, the fatigue model 71 may output a score of zero (0) if the elapsed time is less than one hour. The output score from the fatigue model 71 is referred to as a time fatigue score.

[0045] The time when the wearable device 20 is attached can be determined by various methods. For example, whether the wearable device 20 is attached may be determined based on the acquisition status of sensor data by the acquisition unit 61 and the content of the acquired sensor data, and the timing at which it is determined that the wearable device 20 is attached may be determined as the attachment time. For example, when the wearable device 20 is not attached, the body temperature data, pulse wave data, and EDA data do not show normal values. Therefore, it is possible to determine whether the wearable device 20 is attached based on the content of one or more of the sensor data, namely, the body temperature data, pulse wave data, and EDA data. Alternatively, the attachment time may be determined as the timing at which the UI processing unit 62 detects a user operation indicating the start of work on a screen displayed on the user terminal 5. Alternatively, the wearable device 20 may be provided with an operation button, and the timing at which the operation button is pressed may be determined as the attachment time.

[0046] <Sleep Model> The sleep model 72 is a computational model that estimates the degree of recovery from mental fatigue due to last night's sleep. In this embodiment, the sleep model 72 inputs last night's sleep time and outputs a score ranging from 0 to 1 (hereinafter referred to as the sleep recovery score). In other words, a higher sleep recovery score indicates a greater degree of recovery from mental fatigue. If the input sleep time is within a set time, the sleep model 72 outputs a sleep recovery score of 1 or close to 1. If the input sleep time is outside the set time, the sleep model 72 outputs a smaller sleep recovery score the greater the deviation from the set time. The set time may be an amount of sleep known as the ideal amount of sleep for each age group, and may be set, for example, within a range of 7 to 9 hours. Alternatively, a time that is considered to be appropriate for each individual may be input via the UI screen, and the input time may be used as the set time.

[0047] Because it is known that not only the amount of sleep but also the quality of sleep is important for fatigue recovery, the sleep model 72 may calculate the sleep recovery score using a coefficient a corresponding to the quality of sleep. In this case, the sleep model 72 outputs a sleep recovery score in the range of 0 or more and the coefficient a or less. The coefficient a corresponding to the quality of sleep can be determined using various methods. Because a correlation between the number of steps taken per day and the quality of sleep (e.g., sleep efficiency) is known, the sleep model 72 inputs the number of steps taken yesterday along with the amount of sleep last night, and outputs a sleep recovery score in the range of 0 or more and the coefficient a or less. For example, the sleep model 72 sets the coefficient a to 1 or more if the input number of steps is equal to or greater than an ideal number of steps (e.g., 8,300 steps) that can improve sleep efficiency, and sets the coefficient a to less than 1 if the input number of steps is less than the ideal number of steps, and calculates the sleep recovery score by multiplying the score determined as described above according to the input amount of sleep by the coefficient a.

[0048] The sleep duration and number of steps input to the sleep model 72 may be acquired by the UI processing unit 62 in response to a user operation on a UI screen displayed on the user terminal 5, or the sleep duration and number of steps measured last night and yesterday may be transmitted and acquired by the acquisition unit 61. Moreover, the sleep duration and number of steps may be calculated and stored based on sensor data (especially acceleration data) from last night or yesterday acquired by the acquisition unit 61.

[0049] <Behavior Model> The behavior model 73 is a calculation model that estimates the degree of fatigue associated with the user's behavior. In this embodiment, the behavior model 73 estimates the type of behavior of the user at that time using sensor data (especially acceleration data) acquired by the acquisition unit 61 as input, and outputs a score corresponding to the estimated type of behavior (hereinafter referred to as behavioral fatigue score).

[0050] Various existing methods can be used to estimate the activity type from acceleration data. For example, the activity model 73 estimates the activity type from the input acceleration data using a trained model. The trained model may be a regression equation obtained by regression analysis, or a neural network model obtained by principal component analysis, deep learning, or the like, and the data structure, learning algorithm, etc. of the model are not limited. However, the activity model 73 can also estimate the activity type using other sensor data in addition to or instead of the acceleration data.

[0051] Examples of estimated activity types include doing nothing, working on a personal computer (PC), meeting, eating, drinking, writing, napping, chatting, stretching, walking, jogging, walking up stairs, and traveling by train. The activity model 73 identifies an activity type ID (activity type identifier) ​​corresponding to the input acceleration data. Furthermore, there are no limitations on the method for obtaining an activity fatigue score from an activity type. For example, the activity model 73 holds a conversion table between activity type IDs and activity fatigue scores, and can convert an activity type ID into an activity fatigue score by referring to this conversion table.

[0052] Human behavior includes behaviors that increase transient fatigue, behaviors that can reduce transient fatigue, and behaviors that can reduce chronic fatigue. For example, behaviors such as computer work, meetings, and writing increase transient fatigue, while behaviors such as eating, drinking, and chatting can reduce transient fatigue. Furthermore, behaviors known as active rest, such as stretching, walking, jogging, and climbing stairs, can reduce chronic fatigue. Napping can reduce both transient and chronic fatigue. Therefore, in this embodiment, the behavioral model 73 outputs behavioral fatigue scores divided into a transient fatigue score and a chronic fatigue score. Hereinafter, these scores are referred to as the transient fatigue behavioral fatigue score and the chronic fatigue behavioral fatigue score. The transient fatigue behavioral fatigue score and the chronic fatigue behavioral fatigue score correspond to the first behavioral fatigue score and the second behavioral fatigue score, respectively. The behavioral model 73 corresponds to a third estimation means that periodically estimates the first behavioral fatigue score and the second behavioral fatigue score corresponding to the behavior type indicated by the acquired behavioral data based on the behavioral data.

[0053] In the conversion table, each behavioral type ID is associated with a behavioral fatigue score for transient fatigue and a behavioral fatigue score for chronic fatigue. In this embodiment, a behavioral type ID corresponding to a behavior that increases transient fatigue is associated with a behavioral fatigue score for transient fatigue (+1) and a behavioral fatigue score for chronic fatigue (0). A behavioral type ID corresponding to a behavior that can reduce transient fatigue is associated with a behavioral fatigue score for transient fatigue (-1) and a behavioral fatigue score for chronic fatigue (0). A behavioral type ID corresponding to a behavior that can reduce chronic fatigue is associated with a behavioral fatigue score for transient fatigue (0) and a behavioral fatigue score for chronic fatigue (-1). A behavioral type ID corresponding to a behavior that can reduce both transient fatigue and chronic fatigue (e.g., napping) is associated with a behavioral fatigue score for transient fatigue (-1) and a behavioral fatigue score for chronic fatigue (-1). In this manner, the behavioral fatigue scores for transient fatigue and chronic fatigue output from the behavior model 73 in this embodiment are set to "-1," "0," or "+1."

[0054] As described above, the behavioral types estimated by the behavioral model 73 may include behavioral types that belong to breaks during working hours (task hours), such as eating, drinking, chatting, napping, stretching, walking, jogging, etc. According to the behavioral model 73 of this embodiment, among these behavioral types that belong to breaks, the behavioral fatigue scores output for behavioral types such as eating, drinking, and chatting are output so as to indicate a decrease in fatigue only in transient fatigue (a behavioral fatigue score for transient fatigue of "-1" is output), the behavioral fatigue scores output for behavioral types called active rest, such as stretching, walking, and jogging, are output so as to indicate a decrease in fatigue only in chronic fatigue (a behavioral fatigue score for chronic fatigue of "-1" is output), and the behavioral fatigue score output for the behavioral type of nap is output so as to indicate a decrease in fatigue in both transient fatigue and chronic fatigue (a behavioral fatigue score for transient fatigue of "-1" and a behavioral fatigue score for chronic fatigue of "-1" are output). That is, among the estimated behavior types, the behavior types belonging to breaks during working hours can be expressed as including a behavior type estimated by the third estimation means (behavior model 73) such that only the first behavioral fatigue score (behavior fatigue score for transient fatigue) indicates a decrease in fatigue, and a behavior type estimated by the third estimation means (behavior model 73) such that only the second behavioral fatigue score (behavior fatigue score for chronic fatigue) indicates a decrease in fatigue. Furthermore, when behavior data indicating a behavior type of nap is acquired, the third estimation means (behavior model 73) can be expressed as estimating such that both the first behavioral fatigue score (behavior fatigue score for transient fatigue) and the second behavioral fatigue score (behavior fatigue score for chronic fatigue) indicate a decrease in fatigue.

[0055] <Motivation Model> The motivation model 74 is a calculation model that estimates the degree of reduction in mental fatigue associated with the user's motivation. The motivation model 74 in this embodiment receives user input data related to motivation and outputs a score corresponding to motivation (hereinafter referred to as motivation score). The motivation model 74 in this embodiment outputs a motivation score of "1" when motivation is high, a motivation score of "0.7" when motivation is medium, and a motivation score of "0" when motivation is average. However, the motivation score to be output is not limited to these three values; it may be set such that the higher the motivation, the larger the value, and the lower the motivation, the smaller the value.

[0056] The user input data related to motivation is acquired by the UI processing unit 62 in response to a user operation on a UI screen displayed on the user terminal 5. For example, the UI screen displays a five-level selection of motivation, and the UI processing unit 62 acquires a numerical value (1 to 5) indicating the level of motivation selected by the user operation as the user input data. However, the user input data may not be a numerical value, but may be text data entered by the user as a tweet. In this case, the motivation model 74 may estimate the level of motivation by applying existing language processing or the like to the text data as the user input data.

[0057] <Transient Stress Model> The transient stress model 75 is a calculation model that estimates the degree of transient stress experienced by the user using at least the sensor data acquired by the acquisition unit 61. The transient stress model 75 in this embodiment inputs pulse wave data, EDA data, and body temperature data acquired as sensor data, and outputs a score indicating the degree of transient stress (hereinafter referred to as a transient stress score). The transient stress model 75 in this embodiment estimates the presence or absence of transient stress, and outputs a transient stress score of "1" if transient stress is present, and outputs a transient stress score of "0" if transient stress is not present.

[0058] However, the transient stress model 75 may use an existing method such as the method disclosed in Patent Document 1, etc., described as the background art, and the output transient stress score may not be a binary value, but may be configured so that the higher the degree of transient stress, the higher the output value, and the lower the degree of transient stress, the lower the output value. Furthermore, the transient stress model 75 may also use data other than sensor data, such as user input data, as input. In other words, the transient stress model 75 corresponds to a first estimation means that periodically estimates a first stress score (transient stress score) indicating the degree of transient stress using at least the sensor data.

[0059] <Chronic Stress Model> The chronic stress model 76 is a calculation model that estimates the degree of chronic stress experienced by a user using at least user input data acquired by the UI processing unit 62. The chronic stress model 76 in this embodiment inputs user input data related to chronic stress and outputs a score indicating the degree of chronic stress (hereinafter referred to as the chronic stress score). The chronic stress model 76 in this embodiment outputs a chronic stress score of "1" when chronic stress is high, a chronic stress score of "0.7" when chronic stress is medium, and a chronic stress score of "0" when chronic stress is low or absent. However, the output chronic stress score is not limited to these three values; it may be set to a larger value as the chronic stress increases and a smaller value as the chronic stress decreases.

[0060] User input data related to chronic stress is acquired by the UI processing unit 62 in response to user operations on the UI screen displayed on the user terminal 5. For example, the UI screen displays the degree of perception of one or more symptoms known to be caused by chronic stress, selectable in multiple stages (e.g., four stages: very strong, strong, not strong, not strong at all). Various specific symptoms are known to be symptoms of chronic stress, such as fatigue, irritability, tension, depression, lack of concentration, loss of appetite, and difficulty falling asleep, and it is sufficient if one or more of these symptoms are presented on the UI screen.

[0061] The UI processing unit 62 acquires, as user input data, a numerical value (e.g., 1 to 4) indicating the degree of sensation of each symptom selected by a user operation on the UI screen. However, the user input data may not be a numerical value corresponding to such an option, but may be text data input by the user as a tweet. In this case, the chronic stress model 76 may search for words related to chronic stress by applying existing language processing or the like to the text data as user input data, and estimate the degree of chronic stress. Furthermore, the chronic stress model 76 may receive input of data other than user input data, such as sensor data. In other words, the chronic stress model 76 corresponds to a second estimation means that estimates a second stress score (chronic stress score) indicating the degree of chronic stress using at least the user input data.

[0062] Each of the above-described computational models is executed at a predetermined timing. For example, the behavior model 73 and the transient stress model 75, which use sensor data as input, may be executed in the cycle in which the acquisition unit 61 acquires sensor data, or may be executed in a predetermined cycle different from that. The fatigue model 71 uses the elapsed time from the time the wearable device 20 is worn as input, and is therefore executed in a predetermined cycle after the wearable device 20 is detected to be worn. The sleep model 72 is executed only once when the inputted sleep time of the previous night is acquired. Furthermore, the motivation model 74 and the chronic stress model 76, which use user input data as input, are executed each time user input data is acquired.

[0063] The aggregation unit 77 periodically (e.g., every 3 minutes) calculates the transient fatigue intermediate score and the chronic fatigue intermediate score using the scores output from each calculation model. The calculation formulas for the transient fatigue intermediate score and the chronic fatigue intermediate score used in this embodiment are as follows: In the following calculation formulas, c1 represents the transient stress score, c2 represents the chronic stress score, cya represents the behavioral fatigue score for transient fatigue, cra represents the behavioral fatigue score for chronic fatigue, c7 represents the sleep recovery score, c8 represents the time fatigue score, and c9 represents the motivation score. wy1, wya, wy7, wy8, wy9, wr2, wra, wr7, wr8, and wr9 represent weighting coefficients for each score, respectively. Transient fatigue intermediate score = wy1 x c1 + wya x cya - wy7 x c7 + wy8 x c8 - wy9 x c9 Chronic fatigue intermediate score = wr2 x c2 + wra x cra - wr7 x c7 + wr8 x c8 - wr9 x c9

[0064] That is, in this embodiment, the aggregation unit 77 calculates the transient fatigue intermediate score using at least the transient stress score (c1) but not the chronic stress score (c2), whereas the aggregation unit 77 calculates the chronic fatigue intermediate score using at least the chronic stress score (c2) but not the transient stress score (c1). Therefore, the transient fatigue level reflects the transient stress score estimated using the sensor data, and the chronic fatigue level reflects the chronic stress score determined using the user input data. Therefore, according to this embodiment, the transient fatigue level can objectively reflect the stress sensation that varies from person to person using the sensor data, and the chronic fatigue level, which is difficult to measure objectively, can be appropriately quantified using the user input data.

[0065] In this embodiment, the tallying unit 77 uses a transient fatigue behavioral fatigue score (CYA) to calculate the transient fatigue intermediate score, and a chronic fatigue behavioral fatigue score (CRA) to calculate the chronic fatigue intermediate score. This allows the calculation of the transient fatigue intermediate score and the chronic fatigue intermediate score to reflect the fact that a person's behaviors include behaviors that increase transient fatigue, behaviors that may reduce transient fatigue, and behaviors that may reduce chronic fatigue. For example, even when a person performs behaviors that fall under the same rest period, behaviors that only reduce transient fatigue (e.g., chatting, eating) will reduce the transient fatigue intermediate score, while behaviors that only reduce chronic fatigue (e.g., active rest) will reduce the chronic fatigue intermediate score. Therefore, this embodiment allows for accurate assessment of the mental fatigue state.

[0066] Furthermore, the weighting coefficients multiplied by the output scores from each calculation model in the calculation formulas for the transient fatigue intermediate score and the chronic fatigue intermediate score are appropriately determined according to the degree of influence of each output score on the transient fatigue level or the chronic fatigue level. For example, the weighting coefficient wy1 for the transient stress score has a large influence on the transient fatigue level and is therefore set larger than the weighting coefficients wya, wy7, wy8, and wy9 for the other output scores used in calculating the transient fatigue intermediate score. Furthermore, the weighting coefficient wca for the behavioral fatigue score for transient fatigue is set larger than the weighting coefficient wra for the behavioral fatigue score for chronic fatigue. This is because increases or decreases in mental fatigue associated with behavior have a greater effect on transient fatigue than on chronic fatigue. Therefore, although both the behavioral fatigue scores for transient fatigue and chronic fatigue calculated for the behavior type of nap indicate a decrease in fatigue, the weighting coefficients wca and wra cause the transient fatigue intermediate score to decrease to a greater extent than the chronic fatigue intermediate score.

[0067] The tallying unit 77 stores the transient fatigue intermediate score and the chronic fatigue intermediate score calculated as described above in the data storage unit 65. Then, the tallying unit 77 calculates a transient fatigue total score by accumulating the multiple transient fatigue intermediate scores stored in the data storage unit 65. For example, the tallying unit 77 calculates the transient fatigue total score by summing the multiple transient fatigue intermediate scores calculated within a predetermined time period (e.g., two hours) and stored in the data storage unit 65. In this way, the transient fatigue total score can be said to be the sum of multiple transient fatigue intermediate scores calculated periodically.

[0068] The tallying unit 77 also stores the calculated transient fatigue summary score and chronic fatigue summary score in the data storage unit 65. The tallying unit 77 then calculates the latest chronic fatigue summary score based on the transient fatigue intermediate score and chronic fatigue intermediate score stored in the data storage unit 65 and the immediately preceding chronic fatigue summary score. For example, the tallying unit 77 calculates the latest chronic fatigue summary score by adding the latest chronic fatigue intermediate score stored in the data storage unit 65 with the previously calculated chronic fatigue summary score and the transient fatigue intermediate score calculated a predetermined time ago and stored in the data storage unit 65. At this time, the transient fatigue intermediate score may be multiplied by a weighting coefficient for the transient fatigue intermediate score. In this way, the latest chronic fatigue summary score is calculated using the latest chronic fatigue intermediate score as well as the previously calculated chronic fatigue summary score, and the previously calculated chronic fatigue summary score also includes the chronic fatigue intermediate score calculated at that time. Therefore, the chronic fatigue summary score can be considered to be an accumulation of multiple periodically calculated chronic fatigue intermediate scores.

[0069] In this embodiment, the transient fatigue intermediate score calculated a predetermined time before is used to calculate the chronic fatigue summary score. However, multiple transient fatigue intermediate scores may also be used to calculate the chronic fatigue summary score. Because transient fatigue can become chronic fatigue if it accumulates without being resolved, this embodiment appropriately reflects the relationship between transient fatigue and chronic fatigue in the transient fatigue summary score, allowing for appropriate quantification of the mental fatigue state. Preferably, the transient fatigue intermediate score used to calculate the chronic fatigue summary score is one or more transient fatigue intermediate scores calculated within a predetermined time (e.g., two hours) before the transient fatigue intermediate score that is the subject of the transient fatigue summary score calculation. In this way, the chronic fatigue summary score can more appropriately reflect the characteristics of transient fatigue that can turn into chronic fatigue due to its accumulation.

[0070] In this embodiment, the chronic fatigue score is calculated using the previously calculated chronic fatigue score, but the first chronic fatigue score of the day is calculated using the chronic fatigue score calculated at the end of the previous day. This allows the chronic fatigue score to appropriately reflect the fact that chronic fatigue persists from day to day and is relieved by sleep that night, thereby improving the accuracy of the chronic fatigue level.

[0071] As described above, in this embodiment, the transient fatigue summary score is calculated using multiple transient fatigue intermediate scores calculated within a predetermined time period (e.g., two hours), while the chronic fatigue summary score is calculated using the most recent chronic fatigue intermediate score. This difference in calculation method allows the chronic fatigue summary score and transient fatigue summary score to more appropriately reflect the characteristics of chronic fatigue, which changes slowly but transient fatigue, which changes easily. However, the chronic fatigue summary score may also be calculated by using the sum of multiple chronic fatigue intermediate scores calculated within a predetermined time period (e.g., two hours) and stored in the data storage unit 65.

[0072] In this embodiment, the calculation unit 70 truncates the transient fatigue total score and chronic fatigue total score calculated by the counting unit 77 to "0" if they are 0 or less, and truncates them to "1" if they are 1 or greater, and then multiplies the chronic fatigue total score by 100 to obtain the chronic fatigue level. Next, the counting unit 77 subtracts the chronic fatigue level from 100, multiplies the result by the transient fatigue total score, and obtains the fatigue margin by subtracting the chronic fatigue level and the transient fatigue level from 100. In this way, in this embodiment, the chronic fatigue level, transient fatigue level, and fatigue margin are calculated so that the total of the chronic fatigue level, transient fatigue level, and fatigue margin is 100, and each represents a ratio.

[0073] Fig. 4 is a diagram showing an example of a display screen presenting the transient fatigue level, chronic fatigue level, and fatigue margin. The UI processing unit 62 causes the user terminal 5 to display a display screen presenting the transient fatigue level, chronic fatigue level, and fatigue margin calculated by the calculation unit 70, as shown in Fig. 4. In the display screen shown in Fig. 4, the transient fatigue level, chronic fatigue level, and fatigue margin are displayed as circles of corresponding sizes, and each value is displayed as a percentage inside each circle. In the example of Fig. 4, a circle E1 indicates the chronic fatigue level, a circle E2 indicates the transient fatigue level, and a circle E3 indicates the fatigue margin.

[0074] It is preferable that each circular display be colored in a different color. For example, circular display E1 indicates a chronic fatigue level that is difficult to recover from even by resting, and is therefore displayed in red, which suggests danger; circular display E2 indicates a transient fatigue level that is easily recovered from by resting, and is therefore displayed in yellow, which suggests caution; and circular display E3, which indicates a fatigue margin, is preferably displayed in blue, which suggests safety. In this manner, the balance of the chronic fatigue level, transient fatigue level, and fatigue margin can be grasped at a glance, and the user's mental fatigue state can be presented in an easy-to-see manner. In this embodiment, chronic fatigue level, transient fatigue level, and fatigue margin are referred to as mental batteries to make them more familiar to the user.

[0075] The UI processing unit 62 may cause the user terminal 5 to display a display screen presenting changes over time in transient fatigue level, chronic fatigue level, and fatigue margin, instead of or in addition to the display screen illustrated in FIG. 4 . FIG. 5 is a diagram illustrating an example of a display screen presenting changes over time in transient fatigue level, chronic fatigue level, and fatigue margin. The display screen illustrated in FIG. 5 includes a graph display GD in which the horizontal axis represents time and the vertical axis represents the magnitude of transient fatigue level, chronic fatigue level, and fatigue margin. In the graph display GD, the changes over time in transient fatigue level, chronic fatigue level, and fatigue margin are shown as broken lines, and each region of transient fatigue level, chronic fatigue level, and fatigue margin is color-coded. In the example of FIG. 5 , graph region R1 represents chronic fatigue level, graph region R2 represents transient fatigue level, and graph region R3 represents fatigue margin.

[0076] In this case, the calculation unit 70 periodically calculates the transient fatigue level, chronic fatigue level, and fatigue margin and stores them in chronological order in the data storage unit 65, and the UI processing unit 62 generates the graph display GD shown in Fig. 5 based on the transient fatigue level, chronic fatigue level, and fatigue margin in chronological order stored in the data storage unit 65. In this way, it is possible to grasp at a glance the time transition of the percentage balance of the transient fatigue level, chronic fatigue level, and fatigue margin, and the user's mental fatigue state can be presented in an easy-to-read manner.

[0077] 6 is a diagram showing an example of a user notification screen. The UI processing unit 62 displays a corresponding notification screen on the user terminal 5 based on one or more (or all) of the transient fatigue level, chronic fatigue level, and fatigue margin calculated by the calculation unit 70. The notification screen may include display content that guides the user to an action that can reduce one or both of the transient fatigue level and chronic fatigue level. Therefore, the UI processing unit 62 corresponds to a display processing means. This allows the user to automatically be guided to an action that can reduce mental fatigue when the user's mental fatigue state increases, thereby maintaining the user's mental fatigue state at a normal level.

[0078] 6 includes display content that prompts the user to take a break, close their eyes, and take a deep breath. For example, the UI processing unit 62 causes the user terminal 5 to display the user notification screen TG when the fatigue tolerance level falls below a predetermined value, or when one or both (total) of the transient fatigue level and the chronic fatigue level exceed a predetermined value. This is because taking a break is expected to have the effect of reducing at least the transient fatigue level. Furthermore, by providing the user notification screen TG with an operation button TB indicating the start of a break, the UI processing unit 62 can detect the taking of a break in response to detection of a user operation of pressing the operation button TB.

[0079] [Method Details] The fatigue state determination method according to this embodiment (hereinafter sometimes abbreviated as this determination method) will be described in detail below. Fig. 7 is a flowchart showing the fatigue state determination method according to this embodiment (this determination method). This determination method can be executed by a system having one or more processors and memories, such as the above-described system 1. That is, this determination method may be executed only by the determination device 10, or may be executed by the determination device 10 and the wearable device 20, or may be executed by the determination device 10, the wearable device 20, and the user terminal 5.

[0080] In the following description, an example will be given in which the present determination method is executed by the present determination device 10. In this case, the present determination device 10 can execute the present determination method by having the processor 11 execute a computer program stored in the memory 12. This computer program is installed from a portable recording medium such as a CD (Compact Disc) or a memory card, or from another computer on a network, via the input / output I / F 13 or the communication unit 14, and stored in the memory 12.

[0081] As shown in Fig. 7, this determination method includes steps (S71) to (S81). In the following, the determination device 10 will be described as the entity that executes each step, but each step can also be described as being executed by one or more computers or processors, or each of the above-mentioned processing modules that the determination device 10 has. Since each step is similar to the processing content of each processing module of the determination device 10, details of each step will be omitted as appropriate.

[0082] In step (S71), the determination device 10 (acquisition unit 61) acquires time-series sensor data measured by the sensor 25 of the wearable device 20. In this embodiment, the sensor data is wirelessly transmitted from the wearable device 20 equipped with the sensor 25 to the user terminal 5, and then transmitted by communication from the user terminal 5 to the determination device 10. In this embodiment, the acquired sensor data includes time-series body temperature data, pulse wave data, EDA data, and acceleration data (behavioral data).

[0083] In step (S72), the determination device 10 (UI processing unit 62) acquires user input data. In this embodiment, one or more of the following is acquired: user input data indicating the user's sleep time last night and the number of steps taken yesterday; user input data indicating the user's level of motivation; and user input data indicating the user's level of awareness of each of multiple symptoms that may occur due to chronic stress. Because such user input data is acquired when the user performs an input operation on the UI screen, there may be cases where the user input data cannot be acquired in step (S72). Furthermore, in order to acquire periodically required user input data, the determination device 10 may cause the user terminal 5 to display a UI screen for obtaining user input data at a predetermined interval.

[0084] In step (S73), the determination device 10 executes the computational models. In this embodiment, the determination device 10 includes a fatigue model 71, a sleep model 72, a behavioral model 73, a motivational model 74, a transient stress model 75, and a chronic stress model 76 as computational models. In step (S73), the computational models for which data to be input to the computational models was acquired in steps (S71) and (S72) are executed. For example, the behavioral model 73 and the transient stress model 75 use sensor data as input data, and are therefore executed periodically as the sensor data is acquired. Furthermore, the fatigue model 71 uses the time the wearable device 20 is worn or the elapsed time since the start of work as input data, and is therefore executed periodically after the wearable device 20 is worn or the start of work is detected. On the other hand, the sleep model 72, the motivational model 74, and the chronic stress model 76 use user input data as input, and are therefore executed when the user input data required for each model is acquired.

[0085] In steps (S74) and (S75), the present assessment device 10 (counting unit 77) calculates the transient fatigue intermediate score and the chronic fatigue intermediate score using the scores output from each calculation model in step (S73). The calculation formulas for the transient fatigue intermediate score and the chronic fatigue intermediate score are as described above. The output score of a calculation model that was not executed because the necessary user input data was not acquired is substituted into the calculation formula as no data or "0". The execution timing of steps (S74) and (S75) may be synchronized with the acquisition timing of step (S71) as shown in FIG. 7, or may be set at a cycle longer than the acquisition cycle of step (S71), as shown in FIG. 7. In step (S76), the present assessment device 10 stores the transient fatigue intermediate score and the chronic fatigue intermediate score calculated in steps (S74) and (S75) in the data storage unit 65.

[0086] The assessment device 10 repeatedly executes steps (S71) to (S76) (S77; NO) to periodically calculate and store the transient fatigue intermediate score and the chronic fatigue intermediate score, while timing the calculation. When the calculation timing arrives (S77; YES), the assessment device 10 (calculation unit 77) calculates a transient fatigue calculation score by accumulating the multiple transient fatigue intermediate scores stored in the data storage unit 65 (S78). In this embodiment, the assessment device 10 (calculation unit 77) calculates the transient fatigue calculation score by summing the multiple transient fatigue intermediate scores calculated within a predetermined time period (e.g., 2 hours) and stored in the data storage unit 65 (S78).

[0087] Next, the assessment device 10 (the counting unit 77) calculates the latest chronic fatigue summary score based on the transient fatigue intermediate score and the chronic fatigue intermediate score stored in the data storage unit 65 and the most recent chronic fatigue summary score (S79). In this embodiment, the assessment device 10 (the counting unit 77) calculates the latest chronic fatigue summary score by adding the latest chronic fatigue intermediate score stored in the data storage unit 65, the chronic fatigue summary score calculated previously and stored in the data storage unit 65, and the transient fatigue intermediate score calculated a predetermined time ago and stored in the data storage unit 65 (S79). However, in step (S79), in addition to the latest chronic fatigue intermediate score, the sum of multiple chronic fatigue intermediate scores calculated within a predetermined time (e.g., 2 hours) and stored in the data storage unit 65 may also be used.

[0088] In step (S80), the present determination device 10 stores the transient fatigue total score and the chronic fatigue total score calculated in step (S78) and step (S79) in the data storage unit 65.

[0089] Next, in step (S81), the determination device 10 (calculation unit 70) calculates the transient fatigue level, chronic fatigue level, and fatigue margin based on the transient fatigue summary score and chronic fatigue summary score calculated in steps (S78) and (S79). In this embodiment, the determination device 10 (calculation unit 70) rounds the transient fatigue summary score and chronic fatigue summary score to "0" if they are 0 or less, and rounds them to "1" if they are 1 or greater, and then multiplies the chronic fatigue summary score by 100 to determine the chronic fatigue level. Furthermore, the determination device 10 multiplies the value obtained by subtracting the chronic fatigue level from 100 by the transient fatigue summary score to determine the transient fatigue level, and subtracts the chronic fatigue level and transient fatigue level from 100 to determine the fatigue margin.

[0090] The present determination device 10 repeatedly executes the flowchart shown in Fig. 7 to periodically calculate the transient fatigue level, chronic fatigue level, and fatigue margin, and stores the calculated transient fatigue level, chronic fatigue level, and fatigue margin in chronological order in the data storage unit 65. The present determination device 10 (UI processing unit 62) can display a display screen such as that shown in Fig. 4 or 5 on the user terminal 5 using the transient fatigue level, chronic fatigue level, and fatigue margin stored in chronological order in the data storage unit 65.

[0091] 7 shows a plurality of steps in order for ease of understanding, but the execution order of the steps in this determination method is not limited to the example in Fig. 7. In this determination method, some steps may be executed in parallel, or the execution order of some steps may be changed.

[0092] [Modifications] The above-described embodiment is merely an example and may be partially modified as appropriate. For example, in the above-described embodiment, the system 1 includes the wearable device 20 and the user terminal 5. However, the user terminal 5 may be omitted by providing the wearable device 20 with the functions of the user terminal 5 (e.g., UI function, communication function, etc.). In this case, the determination device 10 may exchange data with the wearable device 20 via wireless communication. Also, in the above-described embodiment, the determination device 10 acquires sensor data via the user terminal 5. However, the determination device 10 may acquire sensor data directly from the wearable device 20. Furthermore, all or some of the functions of the determination device 10 may be provided in the user terminal 5. If all of the functions of the determination device 10 are provided in the user terminal 5, the system 1 does not need to include the determination device 10, and the above-described determination method may be executed by the user terminal 5.

[0093] Furthermore, the present determination device 10 does not need to have all of the fatigue model 71, sleep model 72, behavioral model 73, motivation model 74, transient stress model 75, and chronic stress model 76 as calculation models, and may omit the fatigue model 71, behavioral model 73, or motivation model 74. The present determination device 10 may further include other calculation models not described above, and may calculate the intermediate scores, aggregate scores, transient fatigue level, chronic fatigue level, and fatigue margin using output scores from the other calculation models. Furthermore, with regard to the data input to each calculation model, other data may be input in addition to the data described above, and the specific configuration of each calculation model is not limited to the above and may be modified as appropriate.

[0094] Some or all of the above-described embodiments and modifications can be specified as follows: However, the above-described embodiments and modifications are not limited to the following descriptions.

[0095] <1> A fatigue state determination system comprising: a first acquisition means for acquiring time-series sensor data measured by a wearable sensor attached to a user; a second acquisition means for acquiring user input data input by the user; and a determination means for quantitatively determining the mental fatigue state of the user based on the sensor data and the user input data, wherein the determination means includes a calculation means for calculating a transient fatigue level, a chronic fatigue level, and a fatigue margin level as quantitative information of the mental fatigue state. <2> The fatigue state determination system described in <1>, wherein the calculation means includes a calculation means for calculating a transient fatigue summary score and a chronic fatigue summary score based on one or both of the sensor data and the user input data, and calculates the transient fatigue level, the chronic fatigue level, and the fatigue margin as respective ratios to the sum of the transient fatigue level, the chronic fatigue level, and the fatigue margin, calculates the chronic fatigue level based on the chronic fatigue summary score, calculates the transient fatigue level based on the transient fatigue summary score from the remaining ratio obtained by subtracting the ratio of the chronic fatigue level from the total, and calculates the fatigue margin as the remaining ratio obtained by subtracting each ratio of the chronic fatigue level and the transient fatigue level from the total. <3> The fatigue state determination system of <2>, wherein the determination means further includes first estimation means for periodically estimating a first stress score indicating a level of transient stress using at least the sensor data, and second estimation means for estimating a second stress score indicating a level of chronic stress using at least the user input data, and the aggregation means: periodically calculates a transient fatigue intermediate score using at least the first stress score, periodically calculates a chronic fatigue intermediate score using at least the second stress score, calculates the transient fatigue summary score by accumulating a plurality of periodically calculated transient fatigue intermediate scores, and calculates the chronic fatigue summary score by accumulating a plurality of periodically calculated chronic fatigue intermediate scores. <4> The fatigue state determination system of <3>, wherein the aggregation means uses one or more of the transient fatigue intermediate scores calculated previously a predetermined time in calculating the chronic fatigue summary score.<5> The fatigue state determination system described in <3> or <4>, further comprising: a third acquisition means for acquiring time-series behavioral data of the user, wherein the acquired behavioral data indicates the user's behavior type at each time point; the determination means further includes a third estimation means for periodically estimating a first behavioral fatigue score and a second behavioral fatigue score corresponding to the behavior type indicated by the acquired behavioral data based on the behavioral data; the aggregation means further uses the first behavioral fatigue score to calculate the transient fatigue intermediate score and further uses the second behavioral fatigue score to calculate the chronic fatigue intermediate score; and the behavior types that belong to breaks during working hours among the behavioral types that can be indicated by the behavioral data include a behavior type for which the third estimation means estimates that only the first behavioral fatigue score indicates a decrease in fatigue, and a behavior type for which the third estimation means estimates that only the second behavioral fatigue score indicates a decrease in fatigue. <6> The fatigue state determination system of <5>, wherein, when behavioral data indicating a nap behavior type is acquired, the third estimation means estimates that both the first behavioral fatigue score and the second behavioral fatigue score will indicate a decrease in fatigue, and the aggregation means calculates the transient fatigue intermediate score and the chronic fatigue intermediate score based on the estimated first behavioral fatigue score and second behavioral fatigue score so that the transient fatigue intermediate score indicates a greater degree of fatigue decrease than the chronic fatigue intermediate score.<7> The fatigue state determination system of any one of <1> to <6>, further comprising: a display processing means that processes to display a corresponding notification screen based on one or more of the transient fatigue level, the chronic fatigue level, and the fatigue margin, wherein the notification screen includes display content that guides the user to an behavior that can reduce one or both of the transient fatigue level and the chronic fatigue level.<8> A fatigue state determination method that can be executed by a system having one or more processors and memories, comprising: a step of acquiring time-series sensor data measured by a wearable sensor worn by a user; a step of acquiring user input data input by the user; and a determination step of quantitatively determining the mental fatigue state of the user based on the sensor data and the user input data, wherein the determination step includes a calculation step of calculating a transient fatigue level, a chronic fatigue level, and a fatigue margin level as quantitative information of the mental fatigue state. <9> The fatigue state determination method according to <8>, wherein the transient fatigue level, the chronic fatigue level, and the fatigue margin are each calculated as a ratio to a sum of the transient fatigue level, the chronic fatigue level, and the fatigue margin, and the calculation step includes: calculating a transient fatigue summary score and a chronic fatigue summary score based on one or both of the sensor data and the user input data; calculating the chronic fatigue level based on the chronic fatigue summary score; calculating the transient fatigue level based on the transient fatigue summary score from a ratio remaining after subtracting the ratio of the chronic fatigue level from the total; and calculating the fatigue margin as a ratio remaining after subtracting each ratio of the chronic fatigue level and the transient fatigue level from the total. <10> The fatigue state assessment method according to <9>, wherein the assessment step further includes a step of periodically estimating a first stress score indicating a level of transient stress using at least the sensor data, and a step of estimating a second stress score indicating a level of chronic stress using at least the user input data, and the compilation step includes a step of periodically calculating a transient fatigue intermediate score using at least the first stress score, a step of periodically calculating a chronic fatigue intermediate score using at least the second stress score, a step of calculating the transient fatigue summary score by accumulating the periodically calculated multiple transient fatigue intermediate scores, and a step of calculating the chronic fatigue summary score by accumulating the periodically calculated multiple chronic fatigue intermediate scores. <11> The fatigue state assessment method according to <10>, wherein the calculation of the chronic fatigue summary score in the compilation step uses one or more transient fatigue intermediate scores calculated previously a predetermined time.<12> The fatigue state determination method according to <10> or <11>, further comprising: a step of acquiring time-series behavioral data of the user; the acquired behavioral data indicates the user's behavior type at each time point; the determination step further comprises a step of periodically estimating a first behavioral fatigue score and a second behavioral fatigue score corresponding to the behavior type indicated by the acquired behavioral data, based on the acquired behavioral data; the first behavioral fatigue score is further used to calculate the transient fatigue intermediate score, and the second behavioral fatigue score is further used to calculate the chronic fatigue intermediate score; and among the behavioral types that can be indicated by the behavioral data, behavior types belonging to breaks during working hours include behavior types for which only the first behavioral fatigue score is estimated to indicate a decrease in fatigue, and behavior types for which only the second behavioral fatigue score is estimated to indicate a decrease in fatigue. <13> The fatigue state assessment method according to <12>, wherein, when behavioral data indicating a nap behavior type is acquired, both the first behavioral fatigue score and the second behavioral fatigue score are estimated to indicate a decrease in fatigue, and the transient fatigue intermediate score and the chronic fatigue intermediate score are calculated based on the first behavioral fatigue score and the second behavioral fatigue score so that the transient fatigue intermediate score indicates a greater degree of fatigue decrease than the chronic fatigue intermediate score. <14> The fatigue state assessment method according to any one of <8> to <13>, further comprising: a processing step of displaying a corresponding notification screen based on one or more of the transient fatigue level, the chronic fatigue level, and the fatigue margin, wherein the notification screen includes display content that guides the user to an action that can reduce one or both of the transient fatigue level and the chronic fatigue level. <15> A computer program capable of causing one or more computers having a processor and a memory to execute the fatigue state assessment method according to any one of <8> to <14>. <16> A computer-readable storage medium storing a computer program that causes one or more computers having a processor and a memory to execute the fatigue state determination method according to any one of <8> to <14>.

[0096] This application claims priority based on Japanese Patent Application No. 2024-96607, filed on June 14, 2024, the disclosure of which is incorporated herein in its entirety by reference.

Claims

1. A fatigue state determination system comprising: a first acquisition means for acquiring time-series sensor data measured by a wearable sensor attached to a user; a second acquisition means for acquiring user input data entered by the user; and a determination means for quantitatively determining the mental fatigue state of the user based on the sensor data and the user input data, wherein the determination means includes a calculation means for calculating a transient fatigue level, a chronic fatigue level, and a fatigue margin level as quantitative information of the mental fatigue state.

2. The fatigue state determination system of claim 1, wherein the calculation means includes an aggregation means for calculating a transient fatigue summary score and a chronic fatigue summary score based on one or both of the sensor data and the user input data, and calculates the transient fatigue level, the chronic fatigue level, and the fatigue margin as respective ratios to the sum of the transient fatigue level, the chronic fatigue level, and the fatigue margin, calculates the chronic fatigue level based on the chronic fatigue summary score, calculates the transient fatigue level based on the transient fatigue summary score from the remaining ratio obtained by subtracting the ratio of the chronic fatigue level from the total, and calculates the fatigue margin as the remaining ratio obtained by subtracting the respective ratios of the chronic fatigue level and the transient fatigue level from the total.

3. The fatigue state determination system of claim 2, wherein the determination means further includes a first estimation means for periodically estimating a first stress score indicating the degree of transient stress using at least the sensor data, and a second estimation means for estimating a second stress score indicating the degree of chronic stress using at least the user input data, and the aggregation means periodically calculates a transient fatigue intermediate score using at least the first stress score, periodically calculates a chronic fatigue intermediate score using at least the second stress score, calculates the transient fatigue aggregate score by accumulating multiple transient fatigue intermediate scores calculated periodically, and calculates the chronic fatigue aggregate score by accumulating multiple chronic fatigue intermediate scores calculated periodically.

4. The fatigue state determination system according to claim 3, wherein the calculation means uses one or more of the transient fatigue intermediate scores calculated a predetermined time before in calculating the chronic fatigue calculation score.

5. A fatigue state determination system as described in claim 3 or 4, further comprising: a third acquisition means for acquiring time-series behavioral data of the user, wherein the acquired behavioral data indicates the user's behavior type at each time point; the determination means further includes a third estimation means for periodically estimating a first behavioral fatigue score and a second behavioral fatigue score corresponding to the behavior type indicated by the acquired behavioral data based on the behavioral data; the aggregation means further uses the first behavioral fatigue score to calculate the transient fatigue intermediate score and further uses the second behavioral fatigue score to calculate the chronic fatigue intermediate score; and the behavior types that belong to breaks during working hours among the behavioral types that can be indicated by the behavioral data include behavior types for which the third estimation means estimates that only the first behavioral fatigue score indicates a decrease in fatigue, and behavior types for which the third estimation means estimates that only the second behavioral fatigue score indicates a decrease in fatigue.

6. The fatigue state determination system described in claim 5, wherein, when behavioral data indicating a nap behavior type is obtained, the third estimation means estimates that both the first behavioral fatigue score and the second behavioral fatigue score indicate a decrease in fatigue, and the aggregation means calculates the transient fatigue intermediate score and the chronic fatigue intermediate score based on the estimated first behavioral fatigue score and second behavioral fatigue score so that the transient fatigue intermediate score indicates a greater degree of fatigue decrease than the chronic fatigue intermediate score.

7. A fatigue state determination system as described in any one of claims 1 to 6, further comprising a display processing means for processing to display a corresponding notification screen based on one or more of the transient fatigue level, the chronic fatigue level, or the fatigue margin level, wherein the notification screen includes display content that guides the user to take an action that can reduce either or both of the transient fatigue level or the chronic fatigue level.

8. A fatigue state determination method that can be executed by a system having one or more processors and memory, comprising: a step of acquiring time-series sensor data measured by a wearable sensor worn by a user; a step of acquiring user input data entered by the user; and a determination step of quantitatively determining the user's mental fatigue state based on the sensor data and the user input data, wherein the determination step includes a calculation step of calculating transient fatigue level, chronic fatigue level, and fatigue margin level as quantitative information of the mental fatigue state.

Images