KR20200048027A - System of mental stress analysis and operating method thereof - Google Patents

Abstract

The present invention relates to a stress analysis system for measuring the stress of a user in a mobile environment more conveniently and accurately. The stress analysis system comprises: a server including a stress analysis model calculation part and a stress analysis part; a portable sensor detecting a biosignal to transmit the biosignal to the server; and a mobile terminal receiving emotional information to transmit the emotional information to the server, wherein the stress analysis model calculation part determines a biosignal in a static state, and the stress analysis part can analyze stress based on at least one of the biosignal in the static state and the emotional information.

Description

Stress analysis system and its operating method

The present invention relates to a stress analysis system and its operation method, and more particularly, to a stress analysis system and its operation method for calculating a stress analysis model capable of analyzing a user's stress.

In general, stress is a feeling of internal tension, which is a physical or mental tension that causes great discomfort in the stability of the mind or staying with others, or causing such tension. In other words, whether it is pleasant or unpleasant, it means that an uncharacteristic body reaction occurs due to the pressure on the body organs. Temporary reactions to temporary stress may be a natural phenomenon, but if the physical response to stress persists for a long time, it may be damaged mentally or physically. Therefore, a method of measuring stress and alleviating stress is needed to prevent damage caused by stress.

The stress is typically measured by measuring a biological change due to stress using a device that measures a biosignal such as a heartbeat, analyzing a questionnaire that can measure stress, or interviewing a specialist doctor.

Furthermore, since stress may occur temporarily due to factors generated from the outside, it is preferable to measure in a state in which these factors are reflected. In addition, in order to analyze chronic stress, it is desirable to continuously measure the stress and analyze the change trend.

Accordingly, a portable sensor and a mobile terminal that measure a bio-signal can be used to analyze a user's stress, and such an example is described in Korean Patent Publication No. 10-1435680.

Registered Patent Publication No. 10-1435680 (announced on September 20, 2014)

The present invention intends to more accurately analyze a user's stress by using only a biosignal that changes according to a user's brain activity, except for a biosignal according to the user's physical activity.

The present invention is to analyze the stress of the user more accurately by reflecting the characteristics of the biosignal sensed by the portable sensor when the portable sensor is mounted on the user to detect the biosignal during stress analysis.

The stress analysis system according to an embodiment of the present invention includes a stress analysis model calculation unit, a server including a stress analysis unit, a portable sensor that detects and transmits bio signals, and a mobile terminal that receives and transmits emotion information to the server. Including, the stress analysis model calculator determines a biosignal in the static state, and the stress analyzer analyzes the stress based on at least one of the biosignal in the static state and emotion information.

The portable sensor is provided with an accelerometer sensor, and the stress analysis model calculation unit may classify a biosignal detected by the portable sensor into a biosignal in a static state and a biosignal in a dynamic state using measurement information of the accelerometer sensor.

The stress analysis model calculator builds a stress analysis model by calculating at least one of a minimum number of analysis days, a minimum number of measurement time zones, and a minimum measurement time for detecting a biosignal, and the stress analysis unit uses a stress analysis model to detect a biosignal in a static state. And emotional information to analyze the user's stress.

The stress analysis model calculator may determine a minimum number of analysis days by calculating a root mean square (RMS) error for each of a plurality of biosignals measured on different days.

The stress analysis model calculation unit may determine the number of minimum measurement time zones based on the number of input variables for cosine curve fitting.

The stress analysis model calculation unit may determine the minimum measurement time by calculating accuracy for each of the plurality of measurement times while fixing the determined number of minimum measurement time zones.

A method of operating a stress analysis system according to an embodiment of the present invention includes receiving a biosignal, receiving emotion information, determining a biosignal in a static state among biosignals, and a biosignal in a static state. And analyzing the stress based on at least one of the emotion information.

The step of determining the biosignal in the static state is to divide the biosignal detected by the portable sensor into a biosignal in a static state and a biosignal in a dynamic state using measurement information of an accelerometer sensor provided in the portable sensor. It can contain.

The method may further include calculating at least one of a minimum number of analysis days, a minimum number of measurement time zones, and a minimum measurement time for detecting a biosignal.

The step of analyzing the stress may include analyzing the biosignal and emotion information in the static state through the stress analysis model to analyze the user's stress.

According to an exemplary embodiment of the present invention, a change in autonomic nervous system activity when a user does not move using a biosignal in a static state can be measured, so that the stress of the user can be analyzed more accurately. That is, it is possible to exclude various variables affecting the stress analysis without having to provide a separate laboratory, which has the advantage of improving the accuracy of the stress analysis.

According to an embodiment of the present invention, when detecting a bio-signal through a portable sensor such as a wearable device, a minimum for detecting the bio-signal in consideration of the battery usage time of the wearable device and the convenience of a user wearing the wearable device There is an advantage that a stress analysis model can be constructed by calculating at least one of the number of analysis days, the number of minimum measurement time periods, and the minimum measurement time.

1 is a block diagram of a stress analysis system according to an embodiment of the present invention.
2 is a flowchart illustrating a method of operating a stress analysis system according to an embodiment of the present invention.
FIG. 3 is a flowchart illustrating the construction of the stress analysis model of FIG. 2 (S30).
4 is an exemplary diagram illustrating a change in RMS error between measurement dates according to the number of measurement days according to an embodiment of the present invention.
5 is an exemplary diagram showing a circadian rhythm according to an embodiment of the present invention.
6 is an exemplary view showing an output state of a stress analysis result according to an embodiment of the present invention.
7 is an exemplary diagram illustrating an emotion input screen according to an embodiment of the present invention.

Hereinafter, a specific embodiment of the present invention will be described in detail with the accompanying drawings.

1 is a block diagram of a stress analysis system according to an embodiment of the present invention.

As shown in FIG. 1, the stress analysis system 1 includes at least one portable sensor 2, at least one mobile terminal 3, and a server 100, and the server 100 analyzes stress It may include a model calculation unit 10, a stress analysis unit 20, a communication unit 30, and a memory 40.

Each of the portable sensor 2 and the mobile terminal 3 can transmit and receive signals to / from the server 100 through a wired / wireless communication network. In addition, the portable sensor 2 and the mobile terminal 3 may transmit and receive signals between each other through a wired / wireless communication network.

The portable sensor 2 may detect a user's biological signal and transmit the detected biological signal to the server 100.

Here, the biosignal may be body reaction data that is converted according to the degree of stress that the user receives. For example, the biosignal may be an electrocardiogram, a pulse wave wave, or the like.

Electrocardiogram (ECG) represents an electrical phenomenon that occurs during cardiac activity. On the electrocardiogram, contractions of the atrium with the most active activity are most prominent, and this is called a QRS peak or R peak. After measuring the time domain location of the R peak, the time interval between adjacent R peaks is called RRI (RR peak interval). The heart rate per minute (BPM) can be calculated by multiplying the reciprocal of RRI by 1,000 and dividing by 60.

Photoplethysmogram (PPG) shows the change in blood vessel volume at the distal end of the body. When the heart contracts, the pressure of the blood flow is transmitted along the blood flow of the arteries, resulting in a pulse phenomenon on the wrist, etc. At this time, expressing the pulse using the phenomenon that the amount of scattering of the incident light decreases and increases when the capillaries contract and expand This is a pulse wave. The interval between peaks of the pulse wave is called a beat to beat interval (BBI), and is closely related to RRI. Therefore, the heart rate per minute (BPM) may be calculated by multiplying the inverse of the BBI by 1,000 and dividing by 60.

The mobile terminal 3 may receive emotion for grasping the user's stress level. The mobile terminal 3 includes a display module that displays at least one emotion (eg, fear, anger, sadness, joy, excitement, emotion, etc.) and an input module that receives scores for each emotion displayed on the display module. It can contain.

The mobile terminal 3 may randomly output an alarm at least once a day to receive the user's emotion. Specifically, the mobile terminal 3 may randomly output an alarm in the morning, afternoon, and evening hours and then output an emotion input screen as shown in FIG. 7 to receive emotion information from a user.

The emotion information can be input by logging the emotions that the user feels when the alarm is output in the input window. The emotion information may include at least one kind of emotion and a degree for each kind of emotion.

The mobile terminal 3 may transmit emotion information to the server 100.

For example, the portable sensor 2 is a wearable device, such as a heart rate sensor, a wrist band, and a smart watch, and the mobile terminal 3 is a smart phone. Can be.

Meanwhile, although the portable sensor 2 and the mobile terminal 3 are shown in FIG. 1 as being divided into separate devices, this is merely an example for convenience of description, and the portable sensor 2 and the mobile terminal ( 3) can be integrated. For example, when the smartphone senses biometric information and acquires emotion information, the portable sensor 2 and the mobile terminal 3 may be formed as one device.

The server 100 may include a stress analysis model calculation unit 10, a stress analysis unit 20, a communication unit 30, and a memory 40.

The stress analysis model calculating unit 10 may build a stress analysis model applied to stress analysis, and the stress analysis unit 20 may analyze stress using the built-in stress analysis model.

Depending on the embodiment, at least one of the construction of the stress analysis model and the stress analysis may be performed in the mobile terminal 3, but in the following, the stress analysis model calculation unit 10 constructs a stress analysis model and the stress analysis unit ( It is assumed that 20) analyzes stress.

The communication unit 30 may transmit and receive signals and information with at least one of the portable sensor 2 and the mobile terminal 3.

The memory 40 may store a stress analysis model and data necessary for stress analysis. The memory 40 may construct a database necessary for calculating a stress analysis model and a database necessary for stress analysis.

2 is a flowchart illustrating a method of operating a stress analysis system according to an embodiment of the present invention, and FIG. 3 is a flowchart illustrating the construction of the stress analysis model of FIG. 2 (S30).

The communication unit 30 may receive a bio signal (S10).

The communication unit 30 may receive a bio-signal from the portable sensor 2.

The biosignal may be a signal measuring the heart rate.

Since various psychiatric actions appear as dynamic changes in heart rate intervals through the regulation of the autonomic nervous system, it is possible to analyze stress levels by analyzing heart rate variability (HRV). Particularly, in the group under chronic stress, because of increased hormones, the autonomic nervous system regulation action is hindered, and thus HRV is slower than in normal persons, and circadian rhythm of hormone secretion is also slowed down. Therefore, SD based on the deviation of the neighboring heartbeat interval among the heartbeat microvariate variables, RMSSD based on the deviation of the neighboring heartbeat interval, and the LF (which can be obtained through frequency band changes) It is possible to calculate the stress level by statistically analyzing the long frequency (HF) and high frequency (HF) factors and their ratios, and the official values obtained from the height and width of the histogram of heart rate variability.

Therefore, it is possible to receive a signal measuring the heart rate in order to analyze the stress of the user.

The communication unit 30 may receive a bio-signal every predetermined period.

The communication unit 30 may receive emotion information (S20).

The communication unit 30 may receive emotion information from the mobile terminal 3.

The emotion information includes a plurality of emotion types for grasping the user's stress level, and score information for each of the emotion types.

Stress not only causes physical changes to the user, but can also cause mental pain and problems. However, the mental pain or problem felt by the user is limited in being measured through a bio-signal, and thus can be measured through a user's selected option among a plurality of options after providing a question and a plurality of options to the user.

The stress analysis model calculating unit 10 may construct a stress analysis model (S30).

The stress may be analyzed through at least one of biological signals and emotional information. At this time, a stress analysis model serving as a standard for stress analysis is required.

The stress analysis model may mean a criterion for stress analysis. The stress analysis model may include a reception cycle / reception number / reception time zone, analysis method, and the like of each of bio signals and emotion information for stress analysis.

The stress analysis model calculator 10 may build a stress analysis model to improve the accuracy of stress analysis.

Next, a method for constructing a stress analysis model will be described with reference to FIG. 3.

The stress analysis model calculator 10 may determine a biosignal in a static state among the biosignals (S31).

The biosignal may be detected and received from the portable sensor 2 worn by the user, and the biosignal may include both the biosignal measured when the user is dynamic and the biosignal measured when the user is static. Can be.

The biosignal in the dynamic state may mean a biosignal measured when the user moves beyond the reference size, and the biosignal in the static state may mean a biosignal measured when the user moves below the reference size.

On the other hand, the bio-signal in the dynamic state is likely to include noise due to the movement of the user's muscles, and the likelihood of an error in the heart rate is high. Conventionally, the signal in a dynamic state or a signal in a static state is not distinguished, so in a daily life, a mobile experiment is a user's movement state (whether running or walking), drugs (coffee or cigarette, alcohol), emotion (angry, pleasure) ) Since the autonomic nervous system analyzes stress with a biosignal that includes a heart rate variability, which is affected by the degree, the stress analysis results vary depending on physical and mental factors independent of the degree of stress.

Accordingly, the present invention is intended to exclude signals generated in a number of at least dynamic states.

Accordingly, the present invention seeks to provide a method of analyzing a user's movement and determining only a biosignal in the static state, and then analyzing the stress using only the biosignal in the static state.

Specifically, the portable sensor 2 according to an embodiment of the present invention may include an accelerometer sensor for detecting a user's movement. The portable sensor 2 may acquire motion information together through an accelerometer sensor when detecting a biosignal. The portable sensor 2 may transmit bio signals and motion information to the server 100 together.

According to an embodiment, after the portable sensor 2 determines a static state through motion information to obtain a biosignal, only the biosignal in the static state may be transmitted to the server 100. Hereinafter, it is assumed that the stress analysis model calculator 10 determines a static state through motion information.

The stress analysis model calculator 10 may obtain a biosignal in the static state after determining the static state of the user based on motion information sensed by the accelerometer sensor.

The method for determining the static state of the user by the stress analysis model calculator 10 or the portable sensor 2 may be as follows.

The accelerometer sensor is divided into stationary and mobile, and calculates the maximum and minimum values for a certain time (for example, 3 seconds) of the acceleration signal for each of the x-axis, y-axis, and z-axis, and calculates them from a total of 18 variables. One standard deviation (STDEV), the root mean square (RMS) obtained through the root operator on the sum of the squares of the data values on the x, y, and z axes, and the maximum and minimum values of the RMS value for a predetermined time (3 seconds). The difference (DIFF_RMS) can be calculated. 2) It is possible to calculate a feature calculation that expresses a sudden change of a total of 18 accelerometer signals such as a slope, zero crossing, and an accelerometer signal having a signal difference over a certain value. 3) Machine Learning Package Weka can select Logistic among classification functions, select test option as 10 fold cross validation, and adjust ridge value to 0.01 to avoid overfitting. A dynamic state classification model can be calculated using logistic regression, one of the machine learning techniques. The classification accuracy of the model can be tested using N fold cross validation, and if the number of data is small compared to the number of variables, overfitting should be avoided and outliers can be eliminated.

Of the variables 1) and 2), a body movement detection model can be obtained by selecting and inputting only variables smaller than 1.0 based on the Odd Ratio. The logit function used in the model may be as follows.

Z = -0.007 STDEV -0.0456 RMS -0.1023 DIFF_RMS +50.1113

The final discriminant model is a logistic regression model that comes out in sigmoid form and can be as follows.

Y = 1.0 / (1 + exp [Z])

The stress analysis model calculating unit 10 may determine that the Y value obtained by inputting three variables is a dynamic state if it is 0.5 or more, and a static state when it is less than that.

According to the above model, the probability of determining a dynamic state as a dynamic state (True positive) is about 93.9%, the probability of determining a dynamic state as a static state (False Positive) is about 5.9%, and the static state is determined as a static state The False Negative was calculated to be about 94%, and the overall accuracy was calculated to be about 94%. The ROC Area may be 0.987. By applying this body movement detection model to the minimum time-measurement stress analysis model, the judgment error of the stress analysis model obtained from static state data (up to about 36.1%) is the error of the judgment of the stress analysis model obtained from dynamic state data (up to about 40.3%). There is an advantage to be improved.

The stress analysis model calculating unit 10 may determine the static state of the user through the above-described body movement detection model, and then determine the biological signal in the static state among the biological signals.

The stress analysis model calculating unit 10 may determine the minimum number of analysis days (S33).

Recently, with the development of wearable devices and batteries, it is possible to acquire bio signals for a long time. Accordingly, the present invention requires a method of analyzing stress with a biosignal detected for a long time (eg, 3 days or more) with a wearable device without discharging the battery.

At this time, a minimum number of analysis days of the biosignal is required to improve the accuracy of the stress analysis. Here, the minimum number of days of analysis means the minimum number of days required to improve the accuracy of stress analysis among the days of continuous detection of a biosignal. That is, the minimum number of analysis days may indicate the number of days for which detection of a biosignal is required for stress analysis. The minimum number of days of analysis may indicate the minimum amount of biosignal data.

For example, the stress analysis model calculating unit 10 analyzes the heart rate data of a randomly selected measurement day to calculate a model using the circadian rhythm and emotion logging data, and the accuracy of which is a predetermined criterion (for example, 70%. The minimum number of days to be abnormal can be determined as the minimum number of analysis days.

On the other hand, if the number of measurement time zones is set to 12 and the measurement time is set to 2 hours, the accuracy of the model using data from 1, 2, and 3 days changes to a large value (see Table 1 below), so the minimum number of analysis days There may be a problem that is difficult to determine.

Change of model accuracy according to the number of measurement days Measurement days Number of time zones / hour Model input variables True Positive False Positive Precision Recall One 12/2 Fear, touching, sd1 amplitude 69.2 30.3 69.7.2 69.3 2 12/2 Excitation, sd1 amplitude 56.9 48.8 55.7 56.9 3 12/2 Fear, sd1 amplitude 68.2 38.0 67.7 68.2

In addition, if the number of measurement days is increased, the number of subjects in the sample group may decrease, resulting in a problem of inaccurate accuracy. Accordingly, the present invention reduces the number of measurement days by one day, while the emotion logging of measurement days adjacent to each other and the root of RMS (RMS between circadian rhythm variables) We want to determine the minimum number of days of analysis through the trend of decreasing mean squares.

4 is an exemplary diagram illustrating a change in RMS error between measurement dates according to the number of measurement days according to an embodiment of the present invention.

4, the RMS error of the data averaged over 2 days is close to about 0.4, but the error is 0.1 or less in the data averaged over 3 days, and the error is measured to be 0.1 or less in the data averaged over 4 days. . Therefore, in this case, the stress analysis model calculating unit 10 may determine the minimum number of analysis days as 3 days.

As such, the stress analysis model calculator 10 may calculate a root mean square (RMS) error for each of a plurality of bio signals measured on different days. The stress analysis model calculating unit 10 calculates RMS errors for each of a plurality of measurement days when a biosignal is measured, acquires RMS changes between measurement dates, and acquires the number of days when the RMS change decreases by a predetermined value or more. The number of days of analysis can be determined.

The stress analysis model calculator 10 may determine the minimum number of measurement time periods (S35).

Conventionally, it is simply five predetermined sections (for example, the first section from 9:00 to 12:00, the second section from 12:00 to 15:00, the third section from 15:00 to 18:00, and the section from 18:00 to 21:00) Biological signals were detected in the 4th section and the 5th section from 21 to 24 hours). However, as the accuracy of the stress analysis varies according to the number of measurement time zones as described below, it may be required to determine the minimum number of measurement time zones.

Specifically, 144 time zones of 10 minutes long, 48 time zones of 30 minutes long, 24 time zones of 1 hour long, 12 hours of 2 hours long, 8 hours of 3 hours long, 6 hours of 4 hours long, 6 hours of 6 hours long After dividing the bio-signal into 4 hours of length and analyzing, cosine curve fitting, which will be described later, calculated differently as shown in Table 2 below.

Changes in model accuracy according to the number of time zones and the length of the heart rate Number of time zones Measurement time in time zone Model input variables True Positive False Positive Precision Recall 4 6 Fear, excitement, heart rate, sd2 frequency 65.1 36.0 65.5 65.1 6 4 Excitation, heart rate, pnn50 frequency 67.5 36.6 67.2 67.5 8 3 Fear, sadness, excitement, heart rate frequency, heart rate phase, heart rate average, pnn50 frequency, sd1 amplitude, sd2 frequency, sd2 average 73.0 36.7 72.2 73.0 12 2 Fear, excitement, emotion, sd1 amplitude, sd2 average 69.4 34.9 69.1 69.4 24 One Pnn50 phase 64.6 53.5 71.3 64.6 48 30 minutes Fear, excitement, sd1 amplitude 64.1 60.3 58.4 64.1 144 10 minutes Fear, angry, excited 61.9 46.1 60.5 61.9

At this time, if the number of time zones is changed, the number of model input variables for cosine curve fitting varies. As the number increases, the accuracy of curve fitting increases. However, at this time, if the number of time zones is increased, the length of the time zone for obtaining the heart rate average value is reduced, so that the reliability of the heart rate average may decrease. Therefore, it can be confirmed that the accuracy of the model has the maximum value when cosine curve fitting is performed with 8 time zone values and the heart rate is averaged over a length of about 3 hours. You can see that the model's accuracy is high. Therefore, in this case, the stress analysis model calculating unit 10 may determine the number of the minimum measurement time periods as one of 8 to 12 (pieces) and the length of the measurement time as one of 2 to 3 (times). Similarly, the stress analysis model calculation unit 10 may determine the number of minimum measurement time zones based on the number of model input variables for cosine curve fitting. The stress analysis model calculation unit 10 determines the minimum measurement time at a specific time zone. It can be determined (S37).

The stress analysis model calculator 10 may determine a minimum measurement time in a time zone for detecting a biosignal. Here, the minimum measurement time may mean a minimum detection time of a biosignal required for stress analysis.

The stress analysis model calculating unit 10 changes the measurement time while the number of time zones is fixed, calculates the accuracy of the model for a plurality of measurement times, and acquires or calculates the measurement time with the highest accuracy among the calculated accuracy. At least one measurement time having a reference accuracy or higher among the obtained accuracy may be acquired and determined as a minimum measurement time.

Change of model accuracy according to the length of heart rate measurement time at a given measurement time Number of time zones Measurement time Model input variables True Positive False Positive Precision Recall 8 3 Fear, sadness, excitement, heart rate frequency, heart rate phase, heart rate average, pnn50 frequency, sd1 amplitude, sd2 frequency, sd2 average 73.0 36.7 72.2 73.0 8 2 Fear, excitement, sd1 amplitude, sd2 average 70.5 29.8 70.9 70.5 8 One Fear, excitement, emotion, sd2 frequency 69.6 35.0 69.6 69.6 8 30 minutes Fear, excitement, sd1 amplitude 70.4 38.1 69.6 70.4

Referring to Table 3 above, the number of time zones is fixed to 8, and the accuracy is calculated differently by calculating the accuracy by varying the measurement time. In this case, the accuracy of the stress analysis model calculation unit 10 is The measurement time 3 or 2 (time) that is greater than or equal to a reference accuracy (eg, 70%) may be determined as the minimum measurement time. As described above, the stress analysis model calculation unit 10 may fix a plurality of time zones after fixing The minimum measurement time may be determined based on the accuracy calculated for each measurement time. The stress analysis model calculation unit 10 determines a biosignal in a static state, the minimum number of analysis days, the minimum number of measurement time zones, and the minimum measurement By determining the time, you can build a stress analysis model.

Again, FIG. 2 will be described.

The stress analysis unit 20 may analyze a user's stress (S40).

The stress analysis unit 20 may receive a bio-signal according to the built-in stress analysis model, or acquire necessary data among the bio-signals according to the built-in stress analysis model. For example, the stress analysis unit 20 has eight time zones (for example, the first section from 0 to 2 o'clock, the second section from 3 to 5 o'clock, and the time from 6 to 8 o'clock for at least 3 hours). Section 3, section 4 from 9:00 to 11:00, section 5 from 12:00 to 14:00, section 6 from 15:00 to 17:00, section 7 from 18-20, section 8 from 21-24 ), It is possible to analyze the stress by acquiring at least 3 days of detected bio signals.

The stress analysis unit 20 defines a time point having a valid value by fitting a cosine curve to a biosignal, and compares the reference with the input variation to determine whether the stress is high stress or low stress. .

Meanwhile, the stress analysis unit 20 may analyze the stress by reflecting the emotion information in the biological signal. For example, the stress analysis unit 20 may calculate a score according to the emotion information, and determine whether the user is high stress or low stress based on the calculated score and the cosine curve fitting result.

The method of analyzing stress through cosine curve fitting is as follows.

[Equation 1] Y = A · cos (ω · t ± h) + K

The stress analysis unit 20 may fit a cosine curve through Equation 1 as described above for a heart rate graph according to a biosignal.

[Equation 2] Y = ax ² + bx + c

On the other hand, conventionally, when considering curve fitting with a quadratic equation such as Equation 2 above, it is possible to accurately analyze stress through a biosignal that is sensed continuously or for a long time.

In addition, since the cosine curve fitting result is similar to the circadian rhythm as shown in FIG. 5 below, it has an easy advantage in stress analysis.

5 is an exemplary diagram showing a circadian rhythm according to an embodiment of the present invention.

The circadian rhythm refers to a phenomenon in which physiological activity is synchronized (synchronization) with changes in day and night due to global rotation from a single cell to an individual. When the autonomic nervous system is abnormally activated due to a specific physical disease (including brain disease), physiological repetition over time is disturbed, resulting in a pattern different from the normal circadian rhythm pattern (in FIG. 5, the red dotted line is the average work of people) Periodic rhythm).

Therefore, stress analysis is possible with a circadian rhythm, and stress can be more easily and accurately analyzed through cosine curve fitting output similar to the circadian rhythm.

The display module (not shown) provided in the mobile terminal 3 or the server 100 may output a stress analysis result (S50).

When receiving the stress analysis result from the server 100, the mobile terminal 3 may output the stress analysis result.

6 is an exemplary view showing an output state of a stress analysis result according to an embodiment of the present invention.

As illustrated in FIG. 6, the mobile terminal 3 may display the stress level as one of high and low according to the cosine curve fitting result and the questionnaire result.

The above description is merely illustrative of the technical idea of the present invention, and those of ordinary skill in the art to which the present invention pertains may make various modifications and variations without departing from the essential characteristics of the present invention.

Therefore, the embodiments disclosed in the present invention are not intended to limit the technical spirit of the present invention, but to explain, and the scope of the technical spirit of the present invention is not limited by these embodiments.

The scope of protection of the present invention should be interpreted by the claims below, and all technical spirits within the equivalent range should be interpreted as being included in the scope of the present invention.

1: Stress analysis system
2: Handheld sensor
3: Mobile terminal
10: stress analysis model calculation unit
20: stress analysis unit
30: communication department
40: memory
100: server

Claims (10)

In the stress analysis system,
A server including a stress analysis model calculation unit and a stress analysis unit;
A portable sensor that detects a bio-signal and transmits it to the server; And
And a mobile terminal that receives emotion information and transmits it to the server,
The stress analysis model calculator determines a biosignal in a static state,
The stress analysis unit stress analysis system for analyzing the stress based on at least one of the bio-signal in the static state and the emotion information. According to claim 1,
The portable sensor is provided with an accelerometer sensor,
The stress analysis model calculation unit uses the measurement information of the accelerometer sensor to analyze the biosignal detected by the portable sensor into a biosignal in the static state and a biosignal in the dynamic state. According to claim 1,
The stress analysis model calculator builds a stress analysis model by calculating at least one of the minimum number of analysis days, the minimum number of measurement time zones, and the minimum measurement time for detecting the biosignal,
The stress analysis unit analyzes the biosignal and the emotion information in the static state through the stress analysis model to analyze a user's stress. According to claim 1,
The stress analysis model calculator calculates a root mean square (RMS) error for each of a plurality of biosignals measured on different days to determine the minimum number of analysis days. According to claim 1,
The stress analysis model calculation unit determines a minimum number of measurement time zones based on the number of input variables for cosine curve fitting. The method of claim 5,
The stress analysis model calculating unit determines a minimum measurement time by calculating accuracy for each of the plurality of measurement times while fixing the determined number of minimum measurement time zones. Receiving a biosignal;
Receiving emotion information;
Determining a biosignal in a static state among the biosignals; And
And analyzing stress based on at least one of the bio-signal in the static state and the emotion information. The method of claim 7,
Determining the bio-signal in the static state is
Method of operating a stress analysis system comprising the step of dividing the bio-signal detected by the portable sensor into a bio-signal in a static state and a bio-signal in a dynamic state using measurement information of an accelerometer sensor provided in the portable sensor. The method of claim 7,
And calculating at least one of a minimum number of analysis days, a minimum number of measurement time zones, and a minimum measurement time for detecting the bio-signals. The method of claim 9,
The step of analyzing the stress
And analyzing a user's stress by analyzing the biological signal and the emotion information in the static state through the stress analysis model.

Images