JP2021043480A - Napping assistance system and napping assistance program - Google Patents

Abstract

To provide a napping assistance system capable of predicting the level of fatigue of a user and proposing timing for taking a nap.SOLUTION: A mapping assistance system is configured to receive information on a user when active, predict temporal change in at least either of fatigue level and action capability from the present time onwards, and derive a napping plan including at least recommended timing for taking a nap using a result of the prediction.SELECTED DRAWING: Figure 1

Description

The present invention relates to a system that assists in taking a nap to recover from fatigue.

In recent years, daytime naps have been attracting attention as they have the effect of relieving fatigue and can improve the efficiency of work and learning after naps.

For example, in Patent Document 1, an electrode for measuring a user's brain wave is arranged on the surface of a pillow and measured so that a short sleep can be efficiently taken regardless of time or place. A pillow-shaped type that detects the depth of sleep by analyzing brain wave signals and awakens the user by generating a stimulus such as an alarm when a predetermined time has passed after reaching the predetermined depth of sleep. Sleep aids are disclosed.

Japanese Unexamined Patent Publication No. 2013-31542

In recent years, it has been emphasized to secure healthy lives and promote welfare of people, and to realize sustainable economic growth and promotion of rewarding employment. Therefore, it is expected that a nap assistance system that maximizes daytime activity efficiency (performance) can be expected for those who are raising children or caring for home, or workers working in companies or groups, who can recover from fatigue by taking a nap in the gap time. Has been done.

A nap has the effect of relieving fatigue, but a nap during the day is ineffective if it is too short, and if it is too long, it may have the opposite effect, such as drowsiness remaining after awakening or poor sleep at night. In addition, fatigue caused by human activity changes depending on multiple factors such as the activity content of the day and the physical strength and physical condition of the individual. On the other hand, it is difficult for users to understand and predict how tired they are and when to take a nap to effectively recover from fatigue. At present, it is difficult to get the motivation to take a nap positively to recover from fatigue and improve the work and work efficiency after the nap.

The technique of Patent Document 1 can adjust the timing of awakening by measuring the brain wave signal during a nap, but the brain wave can be measured only when the user voluntarily takes a nap. is there.

An object of the present invention is to provide a nap assist system capable of predicting the degree of fatigue of a user and proposing the timing of taking a nap.

In order to achieve the above object, the present invention has a fatigue degree prediction unit and a fatigue degree prediction unit that receive information during a user's activity and predict a time change of at least one of the fatigue degree and the activity ability after the current time. Provided is a nap assist system having a nap plan calculation unit that calculates a nap plan including at least the recommended timing of nap using the prediction result of.

According to the present invention, it is possible to predict the degree of fatigue and the time change of activity ability, which are difficult for the user to notice, and recommend a nap. Therefore, the user can recover the fatigue by taking a nap. It is possible to improve the efficiency of work and work after a nap.

The block diagram which shows the structure of the nap assistance system of this embodiment. Explanatory drawing which shows the predicted value such as the degree of fatigue and the example of a nap plan displayed on the input / output device 70 of FIG. Explanatory drawing which shows the user who is taking a nap. It is explanatory drawing which shows an example of the machine learning model 10a of the fatigue degree prediction unit 10 of FIG. (A) and (b) Explanatory drawing which shows the screen example which accepts the evaluation of the nap plan displayed on the input / output device 70 of FIG. The block diagram which shows the configuration example of the nap assistance system of this embodiment. The flowchart which shows the operation of the nap assistance system of this embodiment.

Hereinafter, the nap assist system according to the embodiment of the present invention will be described with reference to the drawings.

The nap assist system 1 of the present embodiment includes at least a fatigue degree prediction unit 10 and a nap plan generation unit 20 as shown in the functional block diagram in FIG.

The fatigue degree prediction unit 10 receives one or more information at the time of the user's activity, and predicts a time change after the current time of at least one of a value representing a predetermined fatigue degree and a value representing the activity ability (performance) ( (See FIG. 2). Information during user activity includes the user's activity (acceleration, exercise and step count of 1 or more, or a value calculated from them), tension (or relaxation), heart rate (pulse), age, gender, and , One or more occupations can be used.

The value representing the degree of fatigue and the value representing the activity ability may be any value as long as they represent the degree of fatigue of the user and the height of the activity ability, and are predetermined using the user's information. It is desirable that the value can be calculated by a mathematical formula or the like, but it may be a value that cannot be calculated directly from this information. If the fatigue level or activity ability is a value that cannot be calculated directly from the information at the time of the user's activity, the user information is input in advance to the machine learning model 10a described later, and the value representing the fatigue level or activity ability is learned as teacher data. By doing so, it is possible to predict the degree of fatigue and activity ability from the information at the time of the user's activity. The machine learning model 10a will be described in detail later.

The nap plan generation unit 20 calculates a nap plan (see FIG. 2) 21 including at least the recommended nap start timing by using the prediction result of the fatigue degree prediction unit 10.

For example, when the fatigue degree prediction unit 10 predicts the time change of the fatigue degree, the time when the predicted fatigue degree reaches a predetermined value, the time when the fatigue degree starts to increase, or from these time points It is possible to set a nap plan 21 in which a nap (falling asleep) timing is recommended at a predetermined time point (range) such as a predetermined time range. Further, when the time change of the activity ability is predicted, a predetermined time range is set as the recommended nap timing 21 from the time when the activity ability decreases to a predetermined value or the time when the activity ability starts to decrease. be able to.

In addition to the recommended nap start timing, the nap plan 21 may include a recommended nap duration or a recommended nap end (wake-up) timing indicating how many minutes the nap is to be taken. The recommended nap duration may be a predetermined length or may be varied according to the value of one or more pieces of information during the user's activity. For example, the recommended nap duration may vary depending on the degree of relaxation.

The nap duration or awakening timing may be determined by calculation during the nap using the user's information (for example, an electroencephalogram) during the nap (after falling asleep) as described later.

As shown in FIG. 2, the predicted degree of fatigue and time change of activity ability (performance) and the nap timing of the nap plan are displayed on the display unit of the input / output device 70 such as a smartphone.

The user can grasp the prediction of the time change of his / her fatigue level or activity ability by looking at the display unit of the input / output device 70 or the like. At the same time, the recommended nap timing can be grasped.

As described above, the nap assist system of the present embodiment is recommended because it is possible to predict the degree of fatigue and the time change of the activity ability, which are usually difficult to be aware of by oneself, and recommend the user to take a nap. When the user takes a nap at the nap timing, the user's fatigue can be recovered and the work and work efficiency after the nap can be improved.

In addition, as shown in FIG. 2, the nap plan may include prediction of at least one time change of fatigue and activity ability after nap when taking a nap at the recommended nap timing. This allows the user to understand how much his / her fatigue level will be reduced or his / her activity ability will be improved if he / she takes a nap at the recommended nap timing. This motivates users to take an active nap to reduce fatigue and improve their activity. The fatigue level and / or the time change of the activity ability when taking a nap can be predicted by the fatigue level prediction unit 10.

Further, it is desirable that the nap assist system 1 of the present embodiment further includes an awakening timing determination unit 30 that receives information on the user during the nap and determines the timing for awakening the user by calculation. An alarm clock device 40 is connected to the awakening timing determination unit 30. The alarm clock device 40 awakens the user by performing a predetermined operation on the user at the timing for awakening the user determined by the awakening timing determination unit 30. As a predetermined operation, the user is awakened by generating a stimulus to the user such as sound, light, vibration, odor, heat, and lifting the body.

The awakening timing determination unit 30 determines the awakening timing using, for example, at least one of the user's brain wave and the time after taking a nap. Specifically, as shown in FIG. 3, the measurement data of the brain wave of the user during the nap is captured in real time from the brain wave measurement device 131 connected to the nap signal measurement device 64, and the spectrum of theta wave with respect to the α wave. The awakening timing is determined based on the ratio of (θ wave / α wave). For example, if the awakening timing determination unit 30 confirms that it has entered a predetermined sleep stage (a shallow sleep stage with awareness (consciousness)), it takes a nap for a predetermined time (for example, 10 minutes) from that point and enters deep sleep. Awaken before.

Further, even if the awakening timing determination unit 30 takes in the measurement data of the user's pulse rate from the pulse wave measuring device 132 connected to the nap signal measuring device 64 in real time and determines the awakening timing based on the pulse rate. Good. In addition, the awakening timing may be determined by combining the pulse rate and the brain wave. For example, if it is detected that the ratio of the θ wave / α wave spectrum of the brain wave is increased from the awake rest, and the pulse is reduced by a predetermined amount (about 10%) from the awake rest. , Take a nap for a predetermined time (for example, 10 minutes) and wake up before going into deep sleep.

As shown in FIG. 1, the fatigue degree prediction unit 10 described above can be configured to include a machine learning model 10a. Machine learning here includes deep learning. As an example of the machine learning model 10a, various neural networks can be used as shown in FIG. 4 as an example. Since the fatigue degree prediction unit 10 includes the machine learning model 10a, it is possible to receive various information at the time of the user's activity and accurately predict the time change of the fatigue degree / activity ability after the current time.

The machine learning model 10a has been trained in advance using input data and teacher data (correct answer data). As the input data, information on the activity of the user or another person is used. As the teacher data, data obtained by calculating the time change of the fatigue level and / or the activity ability of the user or another person who provided the information during the activity since the present time is used from the measured ability. The information during the activity of the user or another person as input data is the amount of activity of the user (one or more of acceleration, exercise amount and number of steps, or a value calculated from them), heart rate (pulse), age, gender, and degree of tension. One or more of (or degree of relaxation) can be used. The value representing the degree of fatigue and the value representing the activity ability as teacher data may be a value calculated by a mathematical formula or the like predetermined using one or more of these pieces of information, or calculated using information other than the above. It may be a value obtained or a value measured separately using a measuring device.

Further, the fatigue degree prediction unit 10 may be configured to include a re-learning unit 10b. The re-learning unit 10b uses the information at the time of the user's activity used for predicting the degree of fatigue as input data, and the degree of fatigue or activity ability calculated from the user's information actually measured after the prediction, or actually measured. The machine learning model is retrained using the time-varying changes in fatigue or activity ability as teacher data.

Here, the following formula (1) can be used as an example of a mathematical formula for calculating the degree of fatigue or activity ability from the actually measured user information.
(Number 1)
Fatigue (t) = Fatigue (t-1) + f1 (activity, heart rate) ... (1)
Here, the degree of fatigue (t) and the degree of fatigue (t-1) represent the degree of fatigue at time t and t-1, respectively. In addition, f1 (activity amount, heart rate) indicates a function of activity amount and heart rate. Generally, the higher the user's activity, the higher the degree of fatigue. In addition, for example, when the heartbeat is high, the degree of mental and physical tension is high and fatigue accumulates. Since the influence of the amount of activity and the heart rate on the degree of fatigue differs depending on the individual, the degree of fatigue can be calculated by weighting the function f1 by performing machine learning in comparison with the teacher data.

Further, the nap assist system 1 of the present embodiment may further include a nap evaluation receiving unit 80 that receives a subjective evaluation after the nap by the user. The re-learning unit 10b uses the user's subjective evaluation received by the nap evaluation receiving unit 80 for re-learning the machine learning model 10a as a part of the teacher data. As a result, the machine learning model of the fatigue degree prediction unit 10 can generate a nap plan with a high evaluation point. The items of the user's subjective evaluation include, for example, whether he / she was able to wake up comfortably from the nap as shown in FIGS. 5 (a) and 5 (b), whether the head and body were refreshed after the nap, and the timing of the nap (falling asleep). Examples include whether the (time) and nap length were appropriate, and whether the alarm clock was properly woken up.

Further, the awakening timing determination unit 30 can determine the awakening timing by using the subjective evaluation data received in the past by the nap evaluation reception unit 80.

<Operation of nap assistance system>
Hereinafter, the operation of each part of the nap assist system will be described with reference to the flow of FIG. Here, as shown in FIG. 6, the nap assist system 1 is arranged in the server 100 connected to the network 90, and the function of the nap assist system 1 is realized by software. explain.

That is, the functions of each part of the nap assist system 1 are realized by software when the CPU built in the server (computer) 100 reads and executes a program stored in advance in the memory also built in. However, some or all of the functions of each part of the nap assist system 1 can be realized by hardware such as an ASIC (application specific integrated circuit) or an FPGA (field-programmable gate array).

Further, the input / output device 70 and the database 50 are connected to the network 90. As the input / output device 70, a terminal device such as a smartphone or a personal computer is used here.

In addition, the active user wears an active information measuring device 62 that at least measures the amount of exercise, the number of steps, and the heart rate (pulse rate). The activity information measuring device 62 includes, for example, a combination of an activity meter and a pulse rate monitor that measure acceleration to calculate the amount of exercise and the number of steps, a function of optically measuring the pulse rate of a blood vessel in the arm, acceleration, and GPS. A sports watch or the like having a function of measuring the amount of exercise and the number of steps from position information or the like can be used. Further, body temperature, respiratory rate and the like may be measured by a device worn by the user.

Further, the activity information measuring device 62 has a function of calculating the degree of tension and the degree of relaxation from the fluctuation (RR interval, etc.) of the heart rate (pulse rate) by a mathematical formula obtained in advance.

The user at the time of activity inputs the age, gender, and occupation as at least basic information from the user basic information reception unit 61. Further, the user may input the sleep time of the previous day, the sleep depth of the previous day, the medication status, the satiety level, and the like from the user basic information reception unit 61. The user basic information receiving unit 61 may be configured to be shared by an input / output device 70 such as a smartphone.

Also, during a nap, the user wears at least one of the electroencephalogram measuring device 131 and the pulse wave measuring device 132. The output signals of these devices 31 and 132 are connected to the nap signal measuring device 64, and the brain wave and the heart rate (pulse rate) are measured. The electroencephalogram measuring device 131 may be any. For example, it may be worn on the head and applied to the temples, incorporated into a pillow, or incorporated into earphones. The pulse wave measuring device 132 may also serve as a pulse meter or a sports watch of the active information measuring device 62.

The alarm clock 40 may have any configuration, but at the awakening timing determined by the awakening timing determination unit 30, for example, a user such as sound, light, vibration, odor, temperature, and lifting the body. It is a device that generates a stimulus to. This awakens the user.

The electroencephalogram measuring device 131 and the pulse wave measuring device 132 may be arranged in the alarm clock device 40.

An environmental information acquisition device 63 having a function of measuring one or more of air temperature, atmospheric pressure, and humidity and reading a season from a calendar is arranged in a space where a user is active or takes a nap. The environmental information acquisition device 63 can be substituted by the temperature, atmospheric pressure, humidity, and calendar information of the weather forecast of the area published on the Internet.

The information acquired by the active information measuring device 62, the user basic information receiving unit 61, the nap signal measuring device 64, and the environmental information acquiring device 63 is taken into the input / output device 70 here, and is taken into the input / output device 70 and the network. It is stored in the database 50 via 90.

The nap assist system 1 is provided in the server 100 connected to the network 90. The operation of each part of the nap assist system 1 will be described with reference to the flow of FIG.

The fatigue degree prediction unit 10 of the nap assist system 1 includes a recurrent neural network as a machine learning model 10a. As shown in FIG. 4, the recurrent neural network includes an input layer 401, one or more intermediate layers 402, and an output layer 403, and learns by weight adjustment by diachronic error backpropagation using input data and teacher data. It's done. The input data at the time of learning is the basic information of the user or another person and the information at the time of activity (one or more such as age, gender and occupation, and the amount of exercise, the number of steps, the heart rate (pulse), and the degree of tension (or the degree of relaxation). 1 or more and 1 or more of season, temperature, barometric pressure and humidity) and the presence or absence of a nap and the time of falling asleep and the duration of the nap. The teacher data is a time change of a value representing the degree of fatigue of the user or another person and a value representing the activity ability, and a value calculated by a predetermined mathematical formula or the like using one or more of these pieces of information is used.

The fatigue degree prediction unit 10 of the nap assist system 1 uses the database as information on the user's activity, such as one or more of the user's age, gender, occupation, etc., and the amount of exercise, the number of steps, the heartbeat, and the degree of tension (or the degree of relaxation). ) And one or more of the season, temperature, atmospheric pressure, and humidity are taken in and input to the trained machine learning model 10a (step 601).

The machine learning model 10a outputs a time change after the current time of a value representing a predetermined fatigue degree and a value representing an activity ability (performance) corresponding to the input user information and the like. This predicts the time variation of fatigue and activity (step 602). The predicted values of the fatigue degree and the time change of the activity ability are displayed on the display unit of the input / output device 70 as shown in FIG.

The nap plan generation unit 20 generates a nap plan 21 indicating a recommended nap (falling asleep) timing based on the predicted degree of fatigue (step 603). For example, a nap plan 21 is generated with a nap (sleeping) timing that recommends a predetermined time range from the time when the predicted value of fatigue reaches a predetermined value, and the display unit of the input / output device 70 is shown as shown in FIG. It is displayed together with the predicted value of the time change of the degree of fatigue.

The fatigue degree prediction unit 10 adds the recommended nap timing of the nap plan 21 generated in step 603 to the input information used in step 602, and inputs it to the machine learning model 10a. As a result, the machine learning model 10a outputs a predicted value of the time change of the fatigue degree and the activity ability when the machine learning model 10a takes a nap at the recommended nap timing of the nap plan 21 (step 604). The fatigue degree prediction unit 10 displays the predicted value when taking a nap on the display unit of the input / output device 70 together with the predicted value of the time change of the fatigue degree as shown in FIG.

The user can grasp the prediction of the time change of his / her fatigue level or activity ability by looking at the display unit of the input / output device 70 or the like. At the same time, the recommended nap timing can be grasped. The user decides whether or not to take a nap to recover from fatigue, and when taking a nap, wears an electroencephalogram measuring device 131 or the like.

The awakening timing determination unit 30 takes in the user's brain waves and pulse stored in the nap signal measuring device 64 from the database 50 at real timing, and determines whether or not the user has entered the nap (has fallen asleep) from those values. (Step 605). Specifically, when the ratio of theta wave / α wave spectrum of the brain wave increases above a predetermined value and the pulse rate decreases by 10% or more before falling asleep, it becomes a light sleep with awareness (sleep stages 1 and 2). It is determined that the user has entered, and the process proceeds to step 606.

In step 606, the awakening timing determination unit 30 determines the awakening timing (time) after a predetermined time (for example, 10 minutes) after determining that the person has fallen asleep in step 605 in order to awaken before entering deep sleep. To do. This makes it possible for the user to take an appropriate nap.

When the time reaches the awakening timing (step 607), the awakening timing determination unit 30 instructs the alarm device 40 to operate (step 608). The alarm device 40 executes a predetermined alarm operation such as generating an alarm sound to awaken the user.

After awakening, the nap evaluation reception unit 80 displays the screens shown in FIGS. 5A and 5B on the display unit of the input / output device 70, and receives the evaluation of the nap plan 21 from the user (step 609).

The re-learning unit 10b takes in the information during the activity of the user before and after the nap from the database 50, and calculates the values of the changes in the degree of fatigue and the activity ability by using a predetermined mathematical formula (step 610). ..

The re-learning unit 10b uses the values of changes in the degree of fatigue and activity ability calculated in step 610, the timing of the nap, and the evaluation results received by the nap evaluation reception unit 80 as teacher data of the machine learning model 10a, and input data. Relearns the machine learning model 10a using the input data of step 601 (step 611). As a result, the machine learning model 10a after re-learning can output the predicted value of the change in the actual fatigue level and the activity ability and the predicted value of the nap evaluation. Therefore, in steps 602, 603, and 604, the fatigue The accuracy of predicting the degree is improved, the recovery of fatigue due to nap is larger, and it becomes possible to generate a nap plan with a higher predicted value of nap evaluation.

Further, the awakening timing determination unit 30 receives the evaluation result received by the nap evaluation reception unit 80 from the nap evaluation reception unit 80, and when determining the awakening timing next time, etc. so that the nap evaluation score becomes higher. (Step 612). For example, when it is evaluated that the length of the nap is too short as shown in FIG. 5B, the time from the determination that the person has fallen asleep in step 605 to the timing of awakening is set to a predetermined time (for example, 3). Minutes) Make it longer.

In step 612, not only the awakening timing but also the nap plan (nap timing) generated by the nap plan generation unit 20 may be adjusted so that the nap evaluation score is high.

As described above, according to the present embodiment, it is possible to predict the degree of fatigue and the time change of the activity ability, which are difficult for the user to notice, and recommend a nap. When the user takes a nap according to this recommendation, the user's fatigue can be recovered and the work after the nap and the efficiency of the work can be improved.

In the above-described embodiment, the nap assist system 1 is configured to receive information at the time of user activity or nap once stored in the database 50, but the present embodiment is not limited to this configuration. For example, the nap assist system 1 receives signals directly from the user basic information receiving unit 61, the active information measuring device 62, the environmental information acquiring device 63, and the nap signal measuring device 64, and stores the signals in the built-in memory for use. Of course, it is also possible.

Further, the nap assist system 1 can be configured to cooperate with the user's scheduler and the system that manages the user's labor. As a result, the nap assist system 1 can set the nap timing according to the user's schedule and awaken from the nap. In addition, the nap assist system 1 displays a message recommending a long nap to the user when the user is very tired, and in some cases, recommending the user to take a break from work, and works safely. It is also possible to help create an environment where people can work.

1 ... Nap assist system, 10 ... Fatigue prediction unit, 10a ... Machine learning model, 10b ... Re-learning unit, 20 ... Nap plan generation unit, 30 ... Awakening timing determination unit, 40 ... Awakening device, 50 ... Database, 61 ... User basic information reception unit, 62 ... Activity information measurement device, 64 ... Nap signal measurement device, 70 ... Input / output device, 80 ... Nap evaluation reception unit, 100 ... Server

Claims (10)

A fatigue level prediction unit that receives information on the user's activity and predicts the time change of at least one of a value representing a predetermined fatigue level and a value representing the activity ability after the current time.
A nap assist system comprising a nap plan generation unit that calculates a nap plan including at least a recommended nap timing using the prediction result of the fatigue degree prediction unit. The nap assist system according to claim 1, wherein the nap plan includes at least one of a value representing the degree of fatigue and a value representing the activity ability when taking a nap at the recommended nap timing after the nap. A nap assistance system characterized by including prediction of time changes. The nap assist system according to claim 1, wherein when the user enters a nap, the nap timing determination unit has a wake-up timing determining unit that receives information of the user during the nap and determines the timing for awakening the user by calculation. A nap assistance system. The nap assist system according to claim 1, wherein the fatigue degree prediction unit includes a machine learning model.
The machine learning model has been pre-trained using input data and teacher data, the input data includes information on the activity of the user or another person, and the teacher data is after the current time of the user or another person. A nap assist system comprising time-varying data of values representing the degree of fatigue and / or values representing activity ability. The nap assist system according to claim 4, wherein the fatigue degree prediction unit includes a re-learning unit.
The re-learning unit calculates at least one of a value representing the user's fatigue level and a value representing the activity ability at that time using the information at the time of the user's activity.
The machine learning model is relearned using the information at the time of the activity of the user used for the prediction as input data and the time change of at least one of the calculated value representing the degree of fatigue and the value representing the activity ability as teacher data. A nap assistance system characterized by letting you do it. The nap assist system according to claim 5, further comprising a nap evaluation reception unit that accepts a subjective evaluation after a nap by a user, and the re-learning unit receives the evaluation received by the nap evaluation reception unit. , A nap assist system characterized in that it is used as a part of the teacher data. The nap assist system according to claim 3, wherein the awakening timing determination unit captures measurement data of the user's brain wave during nap in real time, and the ratio of theta wave spectrum to the α wave included in the measured brain wave. A nap assist system characterized in that the timing of awakening is determined based on (θ wave / α wave). In the nap assist system according to claim 3, the awakening timing determination unit captures the measurement data of the pulse rate of the user during the nap in real time, and determines the timing of awakening based on the captured pulse rate. A nap assistance system characterized by doing. The nap assist system according to claim 3, further comprising a nap evaluation reception unit that receives a subjective evaluation after a nap by a user, and the awakening timing determination unit is the subjective evaluation received by the nap evaluation reception unit. A nap assist system characterized in that the time from falling asleep to awakening is adjusted when the timing of awakening is determined next time using the data of the evaluation. Computer,
Fatigue prediction means that receives information on the user's activity and predicts the time change of at least one of the value representing a predetermined fatigue degree and the value representing the activity ability after the current time.
A nap assistance program that functions as a nap plan generation means for calculating a nap plan including at least the recommended nap timing using the prediction result of the fatigue degree prediction unit.

Images