WO2019159252A1 - Stress estimation device and stress estimation method using biosignal - Google Patents

Abstract

This stress estimation device is provided with: a biosignal composing means 12 which connects, over the entire period in which stress is to be estimated, biosignal data collected from a stress estimation subject to generate an entire period biosignal, and which connects, over each of a plurality of short periods shorter than the entire period, the biosignal data to generate a plurality of short period biosignals; a stress feature amount calculation means 15 which calculates a stress feature amount from the entire period biosignal to take the calculated stress feature amount as an entire period feature amount, and calculates stress feature amounts from the plurality of short period biosignals to take the calculated stress feature amounts as short period feature amounts; and a stress score estimation means 16 which estimates a stress score from the entire period feature amount and the short period feature amounts.

Description

Stress estimation apparatus and stress estimation method using biological signal

The present invention relates to a stress estimation apparatus and a stress estimation method using a biological signal.

Long-term stress is a stress that accumulates when a person is exposed to various stressors over a long period of time. Long-term stress can lead to mental illness such as depression. Occupational stress is an example of long-term stress. Occupational stress is stress that accumulates especially when workers are exposed to various stressors during work. Depression caused by occupational stress decreases worker productivity. Therefore, early detection and prevention of depression and the like are important. Various long-term stress estimation techniques have been proposed for early detection and prevention of depression and the like.

An example of a long-term stress estimation system is described in Non-Patent Documents 1, 2, 3, and 4. For example, Non-Patent Document 1 and Non-Patent Document 2 disclose a technique for identifying the degree of stress using a feature value obtained by statistically processing a biological signal of a subject over a long period of one month. The feature amount is obtained by statistically processing the biological signal for each action time such as sitting, walking, running, and sleeping. A person with a high score and a person with a low score in a PSSPS (Perceived Stress Scale) questionnaire often used as an index of long-term stress is identified.

Non-Patent Document 3 discloses a technique for estimating the happiness level and productivity of a user based on the frequency distribution of the duration of human physical activity. Non-Patent Document 4 includes statistics (average, standard deviation) of behaviors such as walking, conversation, and desk work, and their behavior time (the behavior time itself, the ratio of the behavior time in one day, and the number of behaviors). And median) and the personality and depression indicators.

The technique described in each non-patent document performs statistical processing over the entire period (for example, one month) during which stress is to be estimated when calculating the feature value for stress estimation from the biological signal data. Therefore, a sufficiently high estimation accuracy cannot be obtained.

The reason for this is that in the PSS や questionnaire, which is psychologically recognized to reflect long-term stress, and the stress response items in the Occupational Stress Simple Questionnaire, the overall stress during the period in which stress should be estimated (1 month) This is because there is no question of experience, and whether or not there is a short-term stress experience during that period and the frequency of the experience. As described above, in the process of estimating the long-term stress, the biological signal information in the time unit having a finer granularity than the period in which the stress is to be estimated is not sufficiently utilized. As a result, it is difficult to accurately estimate the scores of long-term stress indicators such as PSS questionnaires and simple occupational stress questionnaires.

An object of the present invention is to provide a stress estimation device and a stress estimation method capable of estimating long-term stress with high accuracy.

The stress estimation device according to the present invention generates a whole-period biological signal by connecting the biological signal data collected from the subject of stress estimation over the whole period in which stress is estimated, A biosignal forming means for generating a plurality of short-term biosignals by connecting them over a plurality of short-terms shorter than the period, and calculating a stress feature quantity from the whole-period biosignal as a full-period feature quantity, A stress feature quantity calculating means for calculating a stress feature quantity from the inter-vivo signal and making it a short-term feature quantity; and a stress score estimating means for estimating a stress score from the whole-period feature quantity and the short-term feature quantity.

In the stress estimation method according to the present invention, the biological signal data collected from the subject of the stress estimation is connected over the entire period in which the stress is to be estimated to generate a biological signal for the entire period. Connected over each of a plurality of short periods shorter than the period to generate a plurality of short-term biosignals, calculate stress feature values from the whole-period biosignals as full-time feature values, and stress from the short-term biosignals The feature amount is calculated as a short-term feature amount, and a stress score is estimated from the whole-period feature amount and the short-term feature amount.

The stress estimation program according to the present invention generates a whole-period biological signal by connecting the biological signal data collected from the subject of stress estimation to the computer over the whole period in which the stress is to be estimated. Are connected over each of a plurality of short periods shorter than the whole period to generate a plurality of short-term biological signals, and a stress characteristic quantity is calculated from the whole-period biological signal to be a whole-period characteristic quantity. A process of calculating a stress feature quantity from the inter-vivo signal and making it a short-term feature quantity and a process of estimating a stress score from the whole-period feature quantity and the short-term feature quantity are executed.

According to the present invention, long-term stress can be estimated with high accuracy.

It is a block diagram which shows 1st Embodiment of a stress estimation apparatus. It is a flowchart which shows operation | movement of the stress estimation apparatus of 1st Embodiment. It is a block diagram which shows 2nd Embodiment of a stress estimation apparatus. It is a flowchart which shows operation | movement of the stress estimation apparatus of 2nd Embodiment. It is a block diagram which shows 3rd Embodiment of a stress estimation apparatus. It is a flowchart which shows operation | movement of the stress estimation apparatus of 3rd Embodiment. It is explanatory drawing which shows the example of a whole structure of the stress estimation apparatus of a 1st Example. It is a block diagram which shows the structural example of an analysis server. It is explanatory drawing which shows the performance parameter | index of a 1st Example model and a reference model. It is a block diagram which shows the example of a whole structure of the stress estimation apparatus of 2nd Example. It is a block diagram which shows the principal part of a stress estimation apparatus. It is a block diagram which shows the principal part of the other aspect of a stress estimation apparatus. It is a block diagram which shows the principal part of the further another aspect of a stress estimation apparatus. It is a block diagram which shows the principal part of another aspect of a stress estimation apparatus.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

Embodiment 1. FIG.
FIG. 1 is a block diagram showing a first embodiment of a stress estimation apparatus. In the first embodiment, the stress estimation apparatus is realized by the information processing server 100. The information processing server 100 includes a biological signal storage unit 101, a biological signal configuration unit 102, a whole-period biological signal storage unit 103, a short-term biological signal storage unit 104, a stress feature amount calculation unit 105, and a stress score estimation unit. 106 and a stress score output unit 107.

Note that the biological signal constituting unit 102, the stress feature quantity calculating unit 105, and the stress score estimating unit 106 are, for example, 1 in the information processing server 100 according to a stress estimation program stored in a storage unit (not shown) in the information processing server 100. One or a plurality of processors (for example, CPU: Central Processing Unit) can execute the processing. The biological signal storage unit 101, the whole-period biological signal storage unit 103, and the short-term biological signal storage unit 104 can be realized by a storage device (not shown) in the information processing server 100.

The storage unit in which the stress estimation program is stored is a non-transitory nonvolatile memory such as a ROM (Read Only Memory), a flash memory, or a hard disk. The storage device realized by the biological signal storage unit 101, the whole-period biological signal storage unit 103, and the short-term biological signal storage unit 104 is, for example, a hard disk, a flash memory, or an SSD (Solid State Drive).

In this embodiment, the stress estimation apparatus is realized by the information processing server 100, but the constituent elements of the stress estimation apparatus can also be realized by a hardware circuit.

The biological signal storage unit 101 stores biological signal data collected from the subject of stress estimation. The biological signal configuration unit 102 uses the biological signal data stored in the biological signal storage unit 101. The biological signal configuration unit 102 connects biological signal data over a predetermined period. In the present embodiment, the predetermined period includes an entire period for which a stress score is to be calculated and a period shorter than the entire period (short period: for example, one day). For example, when the entire period is one month, the biological signal configuration unit 102 connects the biological signal data for one month to generate a biological signal for the entire period. When the short period is one day, the biological signal data of each day from the first day to the last day of January is connected to obtain a short-term biological signal for each day.

Hereinafter, as an example, the whole period is 1 month and the short period is 1 day. The numerical sequence _{PN represented} by the expression (1) is a biological signal for one day on the Nth day. The “certain biological signal” may be any one-dimensional signal such as a heartbeat, a pulse wave, a numerical value of skin conductivity, and an acceleration in the X-axis direction. In the numerical sequence, _pn represents the nth biological signal data itself. D is the number of biological signals for one day. For example, assuming that the sampling rate is SR [Hz], D = 24 * 3600 * SR except in exceptional cases where a leap second is inserted.

　

The whole period biological signal (biological signal for one month) Q is expressed by equation (2).

　

In the formula (2), M represents the number of days in one month. M is 28, 29, 30, or 31 depending on the situation. The biological signal data stored in the biological signal storage unit 101 is _RL which is a set of fragmented _pn as follows. Here, E _L > S _L.

　

For example, it is assumed that the biological signal data on a specific N'th day is R _{L '} , R _{L' +1} , R _{L '+2} . At this time, the biological signal _{PN ′} of the N′th day is generated by sequentially connecting RL ′, RL ′ + 1, and RL ′ + 2.

There may be an empty period (period in which the collection of R _L ceases) between the time each R _L is collected. In that case, information indicating the existence of an empty period is also recorded in the biological signal storage unit 101. In this case, the number of components of P _{N ′} is less than D.

Further, the N 'date of the first data p _{(N' - 1)} and _D, even when the air-period between the _'first data p _{SL of'} R _L is present, in the biological signal storage unit 101, Information indicating the existence of such an empty period is recorded. Even when an empty period exists between the last data on the _N'th day and the last data of R _{L '+2} , information indicating the existence of such an empty period is recorded in the biological signal storage unit 101. Is done.

In this way, the information of the missing data is recorded together with the data _{PN ′} for one day on the N′th day. _Let the record be NPN _' . When the day break position is in the middle of _RL , the biological signal constituting unit 102 performs a process of cutting _RL .

As described above, the biological signal configuration unit 102 can configure _{PN ′} and NP _{N ′} from an appropriate combination of _RL . Similarly, the biological signal constructing unit 102 can construct Q from an appropriate combination of R _L. As for Q, the information NQ of the missing data is recorded in the biological signal storage unit 101. In this case, the number of components of Q is less than MD.

The process described above is referred to as “concatenation” in this specification. In this embodiment, the entire period is one month and the short period is one day. However, as long as twice the short period is equal to or less than the entire period, these two periods may have any length.

The whole-period biological signal storage unit 103 stores the whole-period biological signal (all-period biological signal) output from the biological signal configuration unit 102. The short-term biological signal construction unit 104 stores a short-term biological signal (short-term biological signal) output from the biological signal construction unit 102.

The stress feature quantity calculation unit 105 calculates a stress feature quantity for each of the whole-period biosignal and the short-term biosignal, and outputs the stress feature quantity as a full-period feature quantity and a short-term feature quantity. As the stress feature amount, as described in Non-Patent Documents 1, 2, 3, and 4, the feature amount statistically processed for each action time such as sitting, walking, running, deskwork, dialogue, and sleep (for example, average) , Variance, median, power spectral density, and components of a histogram of the number of peaks for a certain period of time such as 30 seconds). Moreover, the frequency distribution of the duration of the activity more than a predetermined threshold value etc. is used suitably.

The stress score estimation unit 106 receives the whole-period feature value and the short-term feature value output from the stress feature value calculation unit 105. The stress score estimation unit 106 estimates a stress score from the whole-period feature value and the short-term feature value. The stress score is a score of stress accumulated over a long period of time. For example, it is a score recognized to reflect psychological stress such as the score of a PSS questionnaire. In order to accurately estimate the stress score, the stress score estimation unit 106 uses a database in which biosignal data acquired from many subjects in advance and the stress score that is teacher data are stored as a set. The machine learning model learned in the above is used.

Note that the stress score estimation unit 106 may estimate the PSS score itself as the stress score, but may perform class classification using the stress score or the like. For example, the stress score estimation unit 106 sets two or more classes by setting threshold values for one or more stress scores. Then, the stress score estimation unit 106 classifies the stress score using a threshold value. In this case, the calculated stress score is a number specifying a classification class.

The stress score estimation unit 106 outputs the stress score estimation value to the stress score output unit 107.

In the present embodiment, by realizing the function as described above, a stress score that reflects the overall tendency of stress and the tendency of finer changes in time at the same time is estimated. In other words, not only the stress experience for the entire period for which stress should be estimated, but also information on the frequency of psychologically important stress experiences for each short period can be obtained, so a more accurate stress score can be estimated. is there.

Next, the operation of this embodiment will be described with reference to the flowchart of FIG.

First, the biological signal construction unit 102 acquires time-series biological signal data from the biological signal storage unit 101 (step A1). The biological signal construction unit 102 connects the biological signal data based on the time-series information included in the biological signal data (step A2).

Furthermore, the biological signal construction unit 102 determines from the time-series information of the biological signal data the short-term biological signal separation position (for example, 24 hours that is the end of the day when the short period is one day, one week is the short period). When the data is read (at 24 o'clock on Sunday, which is the end of one week) (step A3), the data connection is stopped (step A6). Then, the biological signal configuration unit 102 stores the connected biological signal data in the short-term biological signal storage unit 104 as a short-term biological signal (step A8).

In addition, when the biosignal forming unit 102 reads the final position of all the stress estimation periods (for example, 24 hours on the 30th or 31st day when the entire period is one month) (step A4), data The connection is stopped (step A7). Then, the biological signal configuration unit 102 stores the connected biological signal data in the whole-period biological signal storage unit 103 as a whole-period biological signal (step A9).

The whole-period biological signal and the short-term biological signal stored in the whole-period biological signal storage unit 103 and the short-term biological signal storage unit 104 are input to the stress feature amount calculation unit 105. The stress feature amount calculation unit 105 calculates a stress feature amount for each of the whole-period biological signal and the short-term biological signal. For example, as described in Non-Patent Documents 1, 2, 3, and 4, the stress feature quantity calculation unit 105 statistically processes stresses by action time such as sitting, walking, running, deskwork, dialogue, and sleep. A feature amount (for example, a component of a histogram of the average, variance, median, power spectral density, and peak number for a fixed period such as 30 seconds) is calculated (steps A10 and A11). Note that the stress feature amount calculation unit 105 may calculate a frequency distribution or the like of the duration of an activity exceeding a certain threshold as the stress feature amount.

Finally, as described above, the stress score estimation unit 106 estimates the stress score using the calculated all-period feature value and short-term feature value. In order to accurately estimate the stress score, the stress score estimation unit 106 estimates the stress score using the machine learning model as described above (step A12). The stress score output unit 107 outputs the estimated score.

The effect of this embodiment will be described.

In the present embodiment, the stress score is estimated using both the feature amount based on the whole-period biological signal (long-term feature amount) and the feature amount based on the short-term biological signal (short-term feature amount). That is, not only the stress experience of the whole period for which stress should be estimated, but also information on the frequency of stress experience for each short period, which is psychologically important. Therefore, long-term stress is estimated with higher accuracy.

Embodiment 2. FIG.
Next, a second embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 3 is a block diagram showing a second embodiment of the stress estimation apparatus. In the stress estimation apparatus of the second embodiment, a short-term feature amount difference calculation unit 108 is added to the stress estimation apparatus of the first embodiment shown in FIG. In the second embodiment, a case where the stress estimation apparatus is realized by the information processing server 100 is taken as an example.

Components other than the short-term feature amount difference calculation unit 108 in FIG. 2, that is, the biological signal storage unit 101, the biological signal configuration unit 102, the whole-period biological signal storage unit 103, the short-term biological signal storage unit 104, and the stress feature amount calculation The configurations and functions of the unit 105, the stress score estimation unit 106, and the stress score output unit 107 are the same as the configurations and functions of the corresponding elements shown in FIG.

The short-term feature value difference calculation unit 108 performs difference calculation on a part of the feature value based on the short-term biological signal calculated by the stress feature value calculation unit 105. The difference calculation is effective in the following cases.

For example, when a user is a worker and long-term stress estimation is performed as a system of a company at work, a biological signal may be acquired only on the working day. In such a case, the stress accumulation situation at the weekend, which is presumed to be important as a period during which stress accumulation is alleviated, cannot be obtained from the biological signal. Therefore, the short-term feature amount difference calculation unit 108 calculates the difference between Friday at the weekend and Monday at the beginning of the week. The short-term feature amount difference calculation unit 108 outputs the difference to the stress score estimation unit 106 as a feature amount.

That is, the short-term feature quantity difference calculation unit 108 calculates, for example, a short-term feature quantity difference before and after the biosignal data non-acquisition period, and the calculation result is added to the stress score estimation unit 106 as an additional short-term feature quantity. Output.

The short-term feature value difference calculation unit 108 also outputs the short-term feature value input from the stress feature value calculation unit 105 to the stress score estimation unit 106. However, the stress feature quantity calculation unit 105 may be configured to output the short-term feature quantity directly to the stress score estimation unit 106.

Next, the operation of this embodiment will be described with reference to the flowchart of FIG.

The processing in steps A1 to A12 shown in FIG. 4 is the same as the processing shown in FIG.

In step B1, the short-term feature amount difference calculation unit 108 performs the difference calculation as described above on the feature amount based on the short-term biological signal calculated by the stress feature amount calculation unit 105.

In this embodiment, the stress score estimation unit 106 also considers the difference input from the short-term feature amount difference calculation unit 108 when estimating the stress score. It is considered that the difference reflects the degree of reduction or increase in stress during the period when the biological signal data is not acquired. Based on the difference, the stress score estimation unit 106 can estimate, for example, the frequency related to recovery from the stress experience. The stress score estimation unit 106 can perform more accurate stress estimation by adding a difference to the stress score.

In this embodiment, in addition to the feature quantity based on the whole-period biosignal and the feature quantity based on the short-term biosignal, information on the difference regarding the feature quantity based on the short-term biosignal can be obtained. Provides information on frequency of recovery from stress experiences. That information is also used to achieve higher estimation accuracy.

Embodiment 3. FIG.
Next, a third embodiment of the present invention will be described with reference to the drawings.

FIG. 5 is a block diagram showing a third embodiment of the stress estimation apparatus. In the stress estimation apparatus of the third embodiment, a feature amount calculation unit 109 is added to the stress estimation apparatus of the second embodiment shown in FIG. Note that, in the third embodiment, a case where the stress estimation apparatus is realized by the information processing server 100 is taken as an example.

Components other than the feature amount calculation unit 109 in FIG. 5, that is, the biological signal storage unit 101, the biological signal configuration unit 102, the whole-period biological signal storage unit 103, the short-term biological signal storage unit 104, the stress feature amount calculation unit 105, The stress score estimation unit 106, the stress score output unit 107, and the short-term feature amount difference calculation unit 108 are configured in the same manner as the elements shown in FIG.

The feature quantity calculation unit 109 calculates all or part of the short-term feature quantity calculated by the stress feature quantity calculation unit 105 and all or part of the short-term feature quantity difference calculated by the short-term feature quantity difference calculation unit 108. As an object, feature quantity (feature value) calculation or feature value detection is executed. Examples of the feature value include a maximum value, a minimum value, a second maximum value, a second minimum value, a first quartile, a third quartile, and a standard deviation.

The stress score estimator 106 not only calculates the difference between the short-term feature value and the short-term feature value, but also all or a part of the short-term feature value or all or a part of the difference value of the short-term feature value (maximum Value, minimum value, second maximum value, second minimum value, first quartile, third quartile, standard deviation, etc.) Information about experience and frequency can be obtained. Note that the stress score estimation unit 106 may consider the feature amount of both the whole or part of the short-term feature amount and the whole or part of the difference of the short-term feature amount.

Next, the operation of this embodiment will be described with reference to the flowchart of FIG.

6 is the same as the process shown in FIG. 4 in steps A1 to A12 and B1.

In steps C <b> 1 and C <b> 2, the feature quantity calculation unit 109 targets the short-term feature quantity calculated by the stress feature quantity calculation unit 105 and the short-term feature quantity difference calculated by the short-term feature quantity difference calculation unit 108. The feature value is calculated for all or a part of each.

Explain the significance of calculating feature values. A case where the feature amount (feature value) is proportional to the intensity of a certain stress experience is taken as an example. Assume that the user A and the user B have the same feature amount in the entire period in which long-term stress is to be estimated. Moreover, the case where each short period is (1/8) of the whole period for which long-term stress should be estimated is taken as an example.

As an example, the short-term feature value of the user A is “20, 1, 19, 2, 18, 3, 17, 4”, and the short-term feature value of the user B is “11, 10, 11, 10, 11, Suppose that it was 10, 11, 10 ". From these figures, it can be said that the user A frequently experienced a short-term stress experience. Moreover, it can be said that the user B has few strong stress experiences as the user A experienced. Therefore, in this example, the score of the PSSＡ questionnaire is expected to be higher for user A. Here, as an example, the entire period is divided into eight periods to make it a short period, but conditions such as how many short periods the entire period is divided can be arbitrarily set according to the situation.

The maximum value, minimum value, second maximum value, and second minimum value of the short-term feature amount are used to estimate how much stress has been experienced during the entire period in which long-term stress should be estimated. To help. In the example of the user A and the user B, the maximum value of the user A is “20”, and the second maximum value is “19”. The maximum value of user B is “11”, and the second maximum value is also “11”. The reason why not only the maximum value but also the minimum value is used is that each feature quantity includes not only those proportional to the stress experience but also those that are inversely proportional or inversely proportional. .

The 1st quartile and the 3rd quartile help to estimate the frequency of strong stress experiences during the entire period in which long-term stress should be estimated. In the above example of user A and user B, the first quartile of user A is “19”, and the first quartile of user B is “11”. The reason why not only the first quartile but also the third quartile is used as the feature amount is the same as the case of the maximum value and the minimum value. This is because it includes not only those that do, but also those that are inversely proportional or inversely proportional.

Standard deviation (variation of stress experience) helps to estimate the difference between weak and strong stress experiences over the entire period in which long-term stress should be estimated. The fact that there is a large difference between a weak stress experience and a strong stress experience, even though the whole-period feature amount is the same, means that a relatively strong stress experience frequently existed. For example, in the example of user A and user B, the standard deviation of user A is 8.64, and the standard deviation of user B is 0.53.

Further, when the difference calculation is performed on the four short-term feature values adjacent to each other for the short-term feature values of the user A and the user B, for the user A, “19, 17, 15, 13 ”,“ 1, 1, 1, 1 ”for user B. In these cases, when the maximum value, the second maximum value, the first quartile, the standard deviation, and the like are calculated, the user A is higher in all cases.

In the present embodiment, the feature amount calculation is performed on the difference between the short-term feature amount and the short-term feature amount. By using feature values for short-term feature quantities and short-term feature quantities along with information on all-period feature quantities, short-term feature quantities, and short-term feature quantity differences, it becomes more clearly psychologically important. Can be obtained.

Specifically, the stress score estimation unit 106 also takes into account the value (feature value) calculated by the feature amount calculation unit 109 when estimating the stress score. That is, feature values (maximum value, minimum value, second maximum value, second minimum value, first fourteenth) regarding all or part of the short-term feature value, or all or part of the difference of the short-term feature value. Quantile, third quartile, standard deviation, etc.) are also considered. The stress score estimation unit 106 can perform more accurate stress estimation by taking them into consideration.

Example 1.
Next, the first embodiment will be described with reference to FIGS. The first example is an example corresponding to the first embodiment. In the following description, reference is also made to specific conditions in an experiment actually performed in order to confirm the effect of the first embodiment.

FIG. 7 is an explanatory diagram showing an example of the overall configuration of the stress estimation apparatus of the first embodiment. In the example shown in FIG. 7, the analysis server 400 as an information processing server is configured to be able to communicate with the biological signal sensors 420A and 420B and the information display devices 430A and 430B via the communication units 410a, 410b, 410c, and 410d. Has been.

The biological signal sensors 420A and 420B acquire a biological signal of a user who is a worker during working hours. Non-patent documents 1, 2, 3, and 4 as skin signals of the user, the skin surface electrical activity that reflects the user's perspiration, the triaxial acceleration that reflects the user's body movement, the user's body temperature ( Skin surface temperature) pulse wave, heartbeat, voice and the like.

Biosignal sensors 420A and 420B are wristband type sensors as described in Non-Patent Documents 1 and 2, as an example. However, a sensor such as a badge type or an employee card type as described in Non-Patent Documents 3 and 4 can also be suitably used as the biological signal sensors 420A and 420B.

In actual experiments, Empatica's E4 sensor was used. The E4 sensor acquires skin conductivity, triaxial acceleration, pulse wave, and skin surface temperature data at sampling rates of 4 Hz, 32 Hz, 64 Hz, and 1 Hz そ れ ぞ れ, respectively. The acquired data is stored in the built-in memory of the biological signal sensors 420A and 420B.

The communication means 410a, 410b, 410c transmit the biological signal data acquired by the biological signal sensors 420A, 420B to the analysis server 400.

In an actual experiment, the E4 sensor is connected to a personal computer with an attached USB (Universal Serial Bus) cable as the communication means 410a and 410b. The personal computer uploads the biological signal data to Empatica Cloud via a wireless LAN (Local Area Network) as the communication means 410c and the Internet as the communication means 410d by the installed dedicated software. The analysis server 400 downloads biosignal data from Empatica Cloud.

FIG. 8 is a block diagram illustrating a configuration example of the analysis server. The analysis server 400 includes a communication interface 111, a biological signal storage unit 101, a biological signal configuration unit 102, a whole-period biological signal storage unit 103, a short-term biological signal storage unit 104, a stress feature amount calculation unit 105, and a stress score estimation unit 106. The stress score output unit 107 and the stress score storage unit 110 exist.

Note that the biological signal constituting unit 102, the stress feature amount calculating unit 105, and the stress score estimating unit 106 are, for example, a processor (for example, a CPU) in the analysis server 400 according to a program stored in a storage unit (not shown) in the analysis server 400. ) Can be realized by executing the process. The biological signal storage unit 101, the whole-period biological signal storage unit 103, the short-term biological signal storage unit 104, and the stress score storage unit 110 can be realized by a storage device (not shown) in the analysis server 400.

The biological signal storage unit 101 stores biological signal data received by the communication interface 111. The biological signal configuration unit 102 generates a biological signal for the entire period using the biological signal data stored in the biological signal storage unit 101. In addition, the biological signal configuration unit 102 also generates a short-term biological signal. The whole-period biological signal and the short-term biological signal are stored in the whole-period biological signal storage unit 103 and the short-term biological signal storage unit 104.

In the actual experiment, the whole period was set to 1 month and the short period was set to 1 week.

The stress feature quantity calculation unit 105 calculates the stress feature quantity from the whole period biosignal for one month. The stress feature amount calculation unit 105 calculates a stress feature amount from a short-term biosignal for each week. As the stress feature amount, as described above, the feature amount statistically processed for each action time such as sitting, walking, running, deskwork, dialogue and sleep (for example, average, variance, median, power spectral density, and 30 A component of a histogram of the number of peaks in a certain period such as seconds) or a frequency distribution of the duration of an activity exceeding a certain threshold value is preferably used.

In actual experiments, as stress feature quantities, the average, variance, median value, power spectral density, and a certain period such as 30 seconds disclosed in Non-Patent Document 1, such as sweating, body movement, skin surface temperature, etc. All the components of the histogram of the number of peaks were calculated. Moreover, about these, calculation was performed about the three activity states of a sitting position, a walk, and driving | running | working, and the whole which put them together. In Non-Patent Document 1, it is assumed that the feature amount is calculated for the whole sleep and the first, second, third, and fourth quarters during sleep. However, in this embodiment, since the subject is a worker and only data during working hours is acquired, data during sleep is not used. Further, in this example, the heart rate that Empatica E4 can acquire was not used. Further, the power spectral density, the histogram element of the number of peaks in a certain period, and the like depend on the data acquisition time. Therefore, they are normalized by dividing them by the actual data acquisition time (pure data acquisition time excluding all of the time when data could not be acquired properly due to a mounting error or the time when it was not mounted).

The stress score estimation unit 106 estimates long-term stress using the inputted long-term feature value for one month and short-term feature value for each week.

In actual experiments, PSS was estimated by regression analysis as long-term stress. The score calculated from the PSS questionnaire conducted at the end of the experiment period (one month) is used as teacher data for the user, and the linear regression model is learned using the feature amount calculated by the stress feature amount calculation unit 105. The model was used to estimate PSS. In the estimation, the Leave-One-person-Out-Cross-Validation method was used. That is, in order to estimate the PSS score of one user, all the other users are used as training data, and a learned model using the PSS score and the feature amount of those users is used.

Furthermore, the performance of the model was also evaluated. At that time, the regularization coefficient regarding the hyperparameter of the linear regression model was fixed to 0.2. In addition, training samples were increased by Over Sampling method for users with high and low PSS scores, which tend to have a small number of samples.

Specifically, for the top 20% and the bottom 20% of users in the PSS score, we introduced a process to increase the number of training samples by Over Sampling. Regarding Over Sampling ratio, one-to-one (no Over Sampling), 10: 1 (when training samples of top 20% and bottom 20% are increased 10 times), and 1-0 (top 20% and bottom) The processing in the three cases (when not learning the middle 60% sample other than 20%) was executed. Regarding the selection of feature values, processing was performed in three cases, up to the top five, the top ten, and the top 20 of the correlation coefficient with PSS. The correlation coefficient between the PSS） score estimated by linear regression and the actual score was calculated for three cases (3 × 3) of the correlation coefficient with PSS. Also, the error of the estimated PSS score with respect to the actual score, specifically RMSE (Rooted Mean Square Error), was calculated as a performance index.

Furthermore, the users with the top 20% of PSS scores were defined as “high stress group”, and a recall curve and relevance rate curve were created for finding high stress group users. The area under these curves (Area Under Curve) was also calculated as a performance index. A model including the above performance index is referred to as a “first embodiment model”.

In addition, in order to confirm the effect of the first embodiment, a performance index of a model (referred to as “reference model”) using only the whole-period feature quantity for one month without using the short-term feature quantity is also calculated. did. FIG. 9 shows the results when the correlation coefficient is maximum for each of the first embodiment model and the reference model. In the example shown in FIG. 9, the first example model exceeds the reference model for all performance indexes.

The stress score estimated by the stress score estimation unit 106 is recorded in the stress score storage unit 110 via the stress score output unit 107.

The stress score can be supplied to a personal computer via the communication interface 111, the Internet, and a wireless LAN according to a user request. In the personal computer, the stress score is displayed on the information display devices 430A and 430B.

In this embodiment, the functions of the information display devices 430A and 430B and the functions of the biological signal sensors 420A and 420B are realized by different components, but these functions may be realized by one terminal. .

Example 2
Next, a second embodiment will be described with reference to FIG. The second example is also an example corresponding to the first embodiment. FIG. 10 is a block diagram illustrating an overall configuration example of the stress estimation apparatus according to the second embodiment. In the second embodiment, a stress estimation device is constructed in the wearable terminal 500.

The biological signal sensor 420 corresponding to the biological signal sensors 420A and 420B shown in FIG. 7 is built in the wearable terminal 500. In addition, an information display device 430 corresponding to the information display devices 430A and 430B shown in FIG.

Wearable terminal 500 further incorporates analysis device 510. The analysis device 510 includes a biological signal storage unit 101, a biological signal configuration unit 102, a whole-period biological signal storage unit 103, a short-term biological signal storage unit 104, a stress feature amount calculation unit 105, a stress score estimation unit 106, and a stress score output unit. 107 and a stress score storage unit 110.

Compared with the first embodiment, in the second embodiment, there is no process for transmitting and receiving signals via components corresponding to the communication means 410a, 410b, 410c, 410d. Therefore, the function of the stress estimation device is realized even in a situation where communication means cannot be obtained.

The operation of the stress estimation apparatus other than the process of transmitting and receiving signals is the same as that in the first embodiment.

In FIG. 10, a wristwatch type terminal is illustrated as the wearable terminal 500, but the wearable terminal 500 is not limited to a wristwatch type terminal, and a spectacle type terminal or the like can also be suitably used.

As described above, in the above embodiment, long-term stress can be estimated with high accuracy. The reason for this is that psychologically important short-term stress is calculated not only by the entire period in which long-term stress is to be estimated, but also by calculating the characteristic amount for stress estimation in units of time shorter than that. This is because it is possible to incorporate information on the presence and frequency of experience into the estimation of long-term stress.

FIG. 11 is a block diagram showing the main part of the stress estimation apparatus. As shown in FIG. 11, the stress estimation device 10 generates biosignals for all periods by connecting biosignal data collected from a stress estimation target person over all periods for which stress is to be estimated. The biological signal constituting unit 12 (in the embodiment, realized by the biological signal constituting unit 102) is configured to connect the signal data over a plurality of short periods shorter than the entire period to generate a plurality of short term biological signals. ) And a stress feature quantity calculating unit 15 (in the embodiment, calculating a stress feature quantity from the whole period biosignal to obtain a whole period feature quantity and calculating a stress feature quantity from the short period biosignal to obtain a short period feature quantity. And a stress score estimation unit 16 (in the embodiment, stress score estimation) that estimates a stress score from the whole-period feature quantity and the short-term feature quantity. It is realized in parts 106.) And a.

FIG. 12 is a block diagram showing a main part of another aspect of the stress estimation apparatus. The stress estimation apparatus 10 illustrated in FIG. 12 further includes a full-period biosignal storage unit 13 (in the embodiment, realized by the full-period biosignal storage unit 103) that stores the generated full-period biosignal. Short-term biological signal storage means 14 (in the embodiment, realized by the short-term biological signal storage unit 104) for storing the short-term biological signal, and the stress feature amount calculating means 15 includes the whole-period biological signal. The whole-period feature value is calculated using the whole-period biological signal stored in the storage unit, and the short-term feature value is calculated using the short-term biological signal stored in the short-term biological signal storage unit.

FIG. 13 is a block diagram showing a main part of another aspect of the stress estimation apparatus. The stress estimation apparatus 10 shown in FIG. 13 further calculates the difference between the short-term feature values before and after the biological signal data non-acquisition period, and outputs the calculation result to the stress score estimation means 16 as an additional short-term feature value. Short-term feature amount supplementing means 18 (implemented by the short-term feature amount difference calculation unit 108 in the embodiment) is provided.

FIG. 14 is a block diagram showing a main part of another aspect of the stress estimation apparatus. The stress estimation apparatus 10 shown in FIG. 14 further calculates a feature value from the short-term feature value calculated from the short-term biological signal or the short-term feature value output by the short-term feature-value supplementing unit 18, Short-term feature value calculation means 19 (implemented by the feature value calculation unit 109 in the embodiment) that outputs to the stress score estimation means 16 as an additional short-term feature value is provided.

Some or all of the above embodiments can be described as the following supplementary notes, but the configuration of the present invention is not limited to the following configurations.

(Supplementary note 1) Biosignal data collected from the subject of stress estimation is connected over the entire period in which stress is to be estimated to generate a biosignal for the entire period, and the biosignal data is obtained from the entire period. A plurality of short-term biological signals connected to each other over a plurality of short short-term biological signals,
A stress feature quantity calculating means for calculating a stress feature quantity from the whole-period biosignal to obtain a full-period feature quantity, and calculating a stress feature quantity from the short-term biosignal to make a short-term feature quantity;
A stress estimation apparatus comprising: a stress score estimation unit that estimates a stress score from the whole period feature quantity and the short period feature quantity.

(Supplementary note 2) Whole-period biosignal storage means for storing the generated whole-period biosignal;
Short-term biosignal storage means for storing the generated short-term biosignal,
The stress feature quantity calculation means calculates the whole period feature quantity using the whole period biosignal stored in the whole period biosignal storage means, and is stored in the short period biosignal storage means. The stress estimation apparatus according to claim 1, wherein the short-term feature value is calculated using the short-term biological signal.

(Supplementary note 3) Short-term feature quantity supplementing means for calculating a difference between the short-term feature quantities before and after the biosignal data non-acquisition period and outputting the calculation result to the stress score estimating means as an additional short-term feature quantity The stress estimation apparatus according to Supplementary Note 1 or Supplementary Note 2.

(Supplementary Note 4) A feature value is calculated from the short-term feature value calculated from the short-term biological signal or the short-term feature value output by the short-term feature value supplementing means, and the calculation result is added to the additional short-term feature. The stress estimation apparatus according to supplementary note 3, further comprising short-term feature value calculation means for outputting to the stress score estimation means as a quantity.

(Supplementary Note 5) The biological signal data is a part or all of signals of skin electrical conductivity, skin surface temperature, pulse wave, heartbeat, voice, and acceleration. The stress estimation device according to any one of Supplementary notes 1 to 4 .

(Appendix 6) Biosignal data collected from the subject of stress estimation is linked over the entire period for which stress is to be estimated to generate a biosignal for the entire period, and the biosignal data is obtained from the entire period. Are connected over each of a plurality of short periods to generate a plurality of short-term biological signals,
A stress feature value is calculated from the whole-period biosignal to obtain a full-time feature value, a stress feature value is calculated from the short-term biosignal to a short-term feature value,
A stress estimation method for estimating a stress score from the whole-period feature value and the short-term feature value.

(Supplementary Note 7) The generated whole-period biological signal is stored in the whole-period biological signal storage means,
The generated short-term biosignal is stored in a short-term biosignal storage means,
Using the whole-period biosignal stored in the whole-period biosignal storage means to calculate the whole-period feature amount, and using the short-term biosignal stored in the short-term biosignal storage means The stress estimation method according to appendix 6, wherein the short-term feature amount is calculated.

(Additional remark 8) The difference of the said short-term feature-value before and behind biosignal data non-acquisition period is calculated, and let a calculation result be the additional short-term feature-value used for the said stress score estimation. Stress estimation method.

(Supplementary note 9) An additional short-term feature value is calculated for calculating a stress score by calculating a feature value from the short-term feature value calculated from the short-term biological signal or the short-term feature value based on the difference. The stress estimation method according to appendix 8, which is used as an inter-feature value.

(Appendix 10)
The biological signal data collected from the subject of stress estimation is connected over the entire period in which stress is to be estimated to generate a whole period biological signal, and the biological signal data is a plurality of shorter than the whole period. A process of generating a plurality of short-term biosignals connected over each of a short period;
Calculating a stress feature quantity from the whole-period biosignal to obtain a full-period feature quantity, and calculating a stress feature quantity from the short-term biosignal to make a short-term feature quantity;
The stress estimation program for performing the process which estimates a stress score from the said whole period feature-value and the said short-term feature-value.

(Supplementary note 11)
Processing for storing the generated all-period biosignal in the all-period biosignal storage means;
A process of storing the generated short-term biosignal in the short-term biosignal storage means;
Using the whole-period biosignal stored in the whole-period biosignal storage means to calculate the whole-period feature amount, and using the short-term biosignal stored in the short-term biosignal storage means The stress estimation program according to appendix 10, wherein the processing for calculating the short-term feature amount is executed.

(Supplementary note 12)
The difference between the short-term feature values before and after the biosignal data non-acquisition period is calculated, and the calculation result is executed as an additional short-term feature value used for the stress score estimation. Stress estimation program.

(Supplementary note 13)
A feature value is calculated from the short-term feature value calculated from the short-term biological signal or the short-term feature value based on the difference, and the calculated result is used as an additional short-term feature value used for stress score estimation; The stress estimation program according to appendix 12, which causes the processing to be executed.

(Supplementary note 14) A non-temporary recording medium storing a stress estimation program, and when the stress estimation program is executed by a processor,
The biological signal data collected from the subject of stress estimation is connected over the entire period in which stress is to be estimated to generate a whole period biological signal, and the biological signal data is a plurality of shorter than the whole period. Connect over each of the short periods to generate multiple short-term biosignals,
A stress feature value is calculated from the whole-period biosignal to obtain a full-time feature value, a stress feature value is calculated from the short-term biosignal to a short-term feature value,
A stress score is estimated from the all-period feature value and the short-term feature value.

(Supplementary Note 15) When the stress estimation program is executed by the processor,
The generated whole period biological signal is stored in the whole period biological signal storage means,
The generated short-term biosignal is stored in a short-term biosignal storage means,
Using the whole-period biosignal stored in the whole-period biosignal storage means to calculate the whole-period feature amount, and using the short-term biosignal stored in the short-term biosignal storage means To calculate the short-term feature value.
The recording medium of appendix 14.

(Supplementary Note 16) When the stress estimation program is executed by the processor,
The recording medium according to supplementary note 14 or supplementary note 15, wherein a difference between the short-term feature values before and after the biosignal data non-acquisition period is calculated, and the calculation result is an additional short-term feature value used for the stress score estimation.

(Supplementary Note 17) When the stress estimation program is executed by the processor,
A feature value is calculated from the short-term feature value calculated from the short-term biological signal or the short-term feature value based on the difference, and the calculated result is used as an additional short-term feature value used for stress score estimation; The recording medium according to appendix 16, wherein the processing is executed.

DESCRIPTION OF SYMBOLS 10 Stress estimation apparatus 12 Biological signal structure means 13 Whole period biological signal storage means 14 Short period biological signal storage means 15 Stress feature-value calculation means 16 Stress score estimation means 18 Short-term feature-value supplement means 100 Information processing server 101 Biosignal storage part DESCRIPTION OF SYMBOLS 102 Biosignal structure part 103 Whole period biosignal storage part 104 Short period biosignal storage part 105 Stress feature-value calculation part 106 Stress score estimation part 107 Stress score output part 108 Short-term feature-value difference calculation part 109 Feature-value calculation part 110 Stress Score storage unit 111 Communication interface 400 Analysis server 410a, 410b, 410c, 410d Communication means 420, 420A, 420B Biological signal sensor 430, 430A, 430B Information display device 500 Wearable terminal 510 Analysis Location

Claims (10)

1. The biological signal data collected from the subject of stress estimation is connected over the entire period in which stress is to be estimated to generate a whole period biological signal, and the biological signal data is a plurality of shorter than the whole period. A biological signal composing means that generates a plurality of short-term biological signals connected over a short period of time;
   A stress feature quantity calculating means for calculating a stress feature quantity from the whole-period biosignal to obtain a full-period feature quantity, and calculating a stress feature quantity from the short-term biosignal to make a short-term feature quantity;
   A stress estimation apparatus comprising: a stress score estimation unit that estimates a stress score from the whole period feature quantity and the short period feature quantity.
2. A lifetime biosignal storage means for storing the generated lifetime biosignal;
   Short-term biosignal storage means for storing the generated short-term biosignal,
   The stress feature quantity calculation means calculates the whole period feature quantity using the whole period biosignal stored in the whole period biosignal storage means, and is stored in the short period biosignal storage means. The stress estimation apparatus according to claim 1, wherein the short-term feature value is calculated using the short-term biological signal.
3. The short-term feature quantity supplementing means for calculating a difference between the short-term feature quantities before and after the biosignal data non-acquisition period and outputting the calculation result to the stress score estimating means as an additional short-term feature quantity. The stress estimation apparatus according to claim 1 or 2.
4. A feature value is calculated from the short-term feature value calculated from the short-term biological signal or the short-term feature value output by the short-term feature value supplementing unit, and the calculation result is used as an additional short-term feature value to calculate the stress. The stress estimation apparatus according to claim 3, further comprising a short-term feature amount calculation means for outputting to the score estimation means.
5. The stress according to any one of claims 1 to 4, wherein the biological signal data is part or all of skin electrical conductivity, skin surface temperature, pulse wave, heartbeat, voice, and acceleration signals. Estimating device.
6. The biological signal data collected from the subject of stress estimation is connected over the entire period in which stress is to be estimated to generate a whole period biological signal, and the biological signal data is a plurality of shorter than the whole period. Connect over each of the short periods to generate multiple short-term biosignals,
   A stress feature value is calculated from the whole-period biosignal to obtain a full-time feature value, a stress feature value is calculated from the short-term biosignal to a short-term feature value,
   A stress estimation method for estimating a stress score from the whole-period feature value and the short-term feature value.
7. The generated whole period biological signal is stored in the whole period biological signal storage means,
   The generated short-term biosignal is stored in a short-term biosignal storage means,
   Using the whole-period biosignal stored in the whole-period biosignal storage means to calculate the whole-period feature amount, and using the short-term biosignal stored in the short-term biosignal storage means The stress estimation method according to claim 6, wherein the short-term feature amount is calculated.
8. The difference of the said short-term feature-value before and behind biosignal data non-acquisition period is calculated, and a calculation result is made into the additional short-term feature-value used for the said stress score estimation. Stress estimation method.
9. A feature value is calculated from the short-term feature value calculated from the short-term biological signal or the short-term feature value based on the difference, and the calculated result is used as an additional short-term feature value used for stress score estimation; The stress estimation method according to claim 8.
10. On the computer,
    The biological signal data collected from the subject of stress estimation is connected over the entire period in which stress is to be estimated to generate a whole period biological signal, and the biological signal data is a plurality of shorter than the whole period. A process of generating a plurality of short-term biosignals connected over each of a short period;
    Calculating a stress feature quantity from the whole-period biosignal to obtain a full-period feature quantity, and calculating a stress feature quantity from the short-term biosignal to make a short-term feature quantity;
    The stress estimation program for performing the process which estimates a stress score from the said whole period feature-value and the said short-term feature-value.

Images