WO2016119400A1

WO2016119400A1 - Method and system for detecting human physiological status transition

Info

Publication number: WO2016119400A1
Application number: PCT/CN2015/083293
Authority: WO
Inventors: 姚健欣
Original assignee: 姚健欣
Priority date: 2015-01-29
Filing date: 2015-07-03
Publication date: 2016-08-04
Also published as: CN104545911B; CN104545911A; SG11201600052WA

Abstract

Disclosed are a method and a system for detecting a human physiological status transition. By means of an adaptive algorithm, different physiological status transition analysis parameters are set up according to different physique conditions of users, and accordingly, the problem of individual differences can be overcome to reflect the physiological status transition of a subject more accurately.

Description

Method and system for detecting human physiological state transition

The present application claims priority to Chinese Patent Application No. 201510045365.0, entitled "A Method and System for Detecting the Transformation of Human Physiological State", filed on Jan. 29, 2015, the entire contents of which is incorporated herein by reference. In the application.

Technical field

The invention relates to the field of biomedical engineering, in particular to a method and a system for detecting a physiological state transition of a human body.

Background technique

When the human body is subjected to external stimuli or internal mood fluctuations, the sympathetic nervous system changes, and the sympathetic nerves cause changes in the activity of the sweat glands, eventually leading to changes in skin resistance. In general, when external stimuli or internal mood fluctuations are greater, the sympathetic nerves are more active, the sweat glands are secreted, the skin's electrical conductivity is enhanced, and the human skin resistance is reduced. Conversely, in a calm state, the sympathetic nerves are in a state of inhibition, sweat gland activity. When it is weakened, the electrical conductivity of the skin is weakened, and the skin resistance of the human body becomes large. Using this principle, we can judge the changes of human body's different states by analyzing the changes in human skin resistance.

Based on this principle, the Chinese patent application CN201410128494.1 (named: an emotion detection method based on human skin resistance change) discloses a method and system for detecting human emotions, using the magnitude, slope and state duration of the resistance. Reflect emotions and quantify the level of excitement through data. The method is mainly applied to the accurate grading of the excitability of the tester when the tester is in a normal living and working state. This method cannot detect whether the tester has transitioned from a normal living state to another physiological state.

In some cases, we need to judge the change in the physiological state of the tester. The physiological state includes a normal living working state, a micro sleep state, a sleep state, and the like. The so-called micro-sleep state means that some brain cells will be in a little while when people are tired. Scientists believe that this phenomenon can explain why our brains often "short-circuit" or "go away" when we are tired. The more scientific micro-sleep state is explained by the fact that 4-7 Hz brainwave activity replaces 8-13 Hz brain waves in the awake state within 3-14 seconds. Magnetic resonance imaging analysis showed that the activity of the thalamus, posterior occipital and occipital cortex decreased at the moment of micro-sleep, while the activity of the frontal lobe, posterior parietal lobe and hippocampus increased. A change in physiological state also causes a change in skin resistance. The change in skin resistance in the micro-sleep state or the sleep state has its particularity with respect to changes in skin resistance under normal living conditions. At the same time, the general application has higher accuracy requirements for detecting the micro-sleep state or the sleep state, for example, detecting the driver's micro-sleep state. In the prior art, the changes in skin resistance have different characteristics under different physiological conditions. Principle, a method for accurately detecting and judging the change of the physiological state of the tester.

Summary of the invention

The technical problem to be solved by the present invention is that, in view of the deficiencies of the prior art, a method and system for detecting a physiological state transition of a human body are proposed, which can accurately reflect the physiological state transition of the tester.

In a first aspect, a method for detecting a physiological state transition of a human body is provided, comprising the steps of: initializing parameters, collecting body resistance data, analyzing body resistance data, and outputting a physiological state transition state;

Analysis of human body resistance data includes the following steps:

Pre-processing the collected human body resistance data; taking the natural logarithm of the pre-processed human body resistance data to obtain the resistance logarithm value; averaging the resistance pair values in the previous window at the current time to obtain the current window in the previous window The resistance logarithmic mean MeanSum(i), where the window size is WinLength; the difference between the resistance log ln(data(i)) at the current time and the logarithmic mean value MeanSum(i) in the previous window Diff(i); takes the absolute value of Diff(i) to obtain the absolute difference of the log AbsDiff(i); inputs the logarithmic absolute difference AbsDiff(i) into the convergence function to obtain the convergence value MeanDiff; and judges the physiology according to the convergence value MeanDiff State transition situation.

Further, the convergence function is:

MeanDiff=α×MeanDiff+(1-α)×AbsDiff(i)

α is a convergence parameter, 0 < α < 1.

This function is a convergence function and is only interested in the current value of MeanDiff, so no history values are stored during iteration. 0<α<1 is a convergence parameter, and the magnitude of the value determines the speed of convergence. The closer the α is to 1, the faster the convergence speed.

Further, the determining the physiological state transition according to the convergence value MeanDiff, the determining method is:

If MeanDiff>MeanDiff _microsleep , the tester enters the normal living and working state;

If MeanDiff _microsleep ≥MeanDiff>MeanDiff _sleep , the tester enters the micro-sleep state;

If MeanDiff ≤ MeanDiff _sleep , the tester enters a sleep state; MeanDiff _microsleep is used to judge the tester's transition threshold from the normal living working state to the micro-sleep state, and MeanDiff _sleep is used to judge the tester to go to sleep from the micro-sleep state. The threshold for the transition of the state.

Further, different values of MeanDiff _microsleep and MeanDiff _sleep are set for different physical conditions of the tester;

The tester's physical condition is judged based on the following two indicators:

mSlope _up =max{slope=(data(i+StepSize-1)-data(i))/StepSize|i=1,2...}

mSlope _down = max{slope=(data(i)-data(i+StepSize-1))/StepSize|i=1,2...}

Where slope is the slope, indicating the average value of the resistance, according to the resistance curve at the ends of a window Resistance and window size calculation; data(i) is the resistance value at time i, data(i+StepSize-1) is the resistance value at time i+StepSize-1; StepSize is the window size;

mSlope _up is the maximum value of the calculated i-up slopes. The upward slope is the last resistance value data(i+StepSize-1) in a certain window minus the first resistance value data by the resistance curve. i) the obtained slope value; mSlope _down is the maximum value of the calculated i-down slopes, and the down-slope is the first resistance value data(i) in the window with the resistance curve calculated by subtracting the last one The slope value obtained by the resistance value data(i+StepSize-1);

If mSlope _up > δ, and mSlope _down > δ, the tester is judged to be relatively sensitive;

If mSlope _up <δ, or mSlope _down <δ, it is judged that the tester is insensitive; wherein δ is the judgment threshold, and its size is related to the window size StepSize.

Further, in the parameter initialization, setting WinLength to 100, the resistance sampling frequency is 50 Hz; the initial value of MeanDiff is 0.03, and α is 0.999;

Set the window size StepSize to 10 and δ to 15;

For testers with relatively sensitive physique, the thresholds MeanDiff _microsleep and MeanDiff _sleep were set to 0.0006 and 0.0002, respectively; for testers with insensitive physique, the thresholds MeanDiff _microsleep and MeanDiff _sleep were set to 0.001 and 0.0005, respectively.

Further, the output physiological state transition state of the human body is a state of transition of the physiological state of the human body by sound, light, vibration or odor.

Further, the collecting body resistance data is obtained by collecting body conductance data, and calculating a resistance value according to a reciprocal relationship between the resistance and the conductance.

In a second aspect, a system for detecting a physiological state transition of a human body is provided, comprising a medical pole piece sequentially connected, a bridge resistance/conductance measuring circuit, an amplifying circuit, an A/D conversion circuit, a CPU, and a human-computer interaction interface;

The bridge resistance/conductance measuring circuit is used for collecting human body resistance/conductance data, and the CPU uses the above method to detect a physiological state transition state of the human body; the human-machine interaction interface outputs a physiological state transition state to the user.

Further, the CPU is a single chip microcomputer, a mobile communication device, a mobile computer device or a desktop computer device.

Further, the human-machine interaction interface includes a voice module, a display module, a vibration module, or an odor generation module.

In the method of the present invention, window processing is employed in two places. The window sizes used are WinLength and StepSize respectively; WinLength is used when comparing the difference between the current time and the previous resistance, and StepSize is used when judging different physical conditions.

The invention is based on the particularity of different physiological states, uses the change of the skin resistance of the human body to detect the change of the physiological state of the human body, and judges that the human body changes from the normal living working state to the micro sleep state, changes from the micro sleep state to the sleep state, or from The micro-sleep state or the sleep state transitions to a normal living and working state. The detection of a physiological state transition has a strong use value. Taking the transition from normal living conditions to micro-sleep states as an example, the detection of changes in the physiological state of the human body can be used to prevent safety accidents of repetitive work and high accident rates such as motorists and heavy machinery operators. The detection of changes in the physiological state of the human body can also be used to judge the concentration of attention of students and listeners, thereby adjusting the teaching methods and attracting their attention.

The invention adopts an adaptive algorithm, and sets different physiological state transition analysis parameters according to different physical conditions of the user, thereby overcoming the individual difference problem, and can more accurately reflect the physiological state transition of the tester.

DRAWINGS

Figure 1 is a diagram showing the physiological state transition of the human body according to the present invention;

2 is a flow chart of data analysis of the present invention;

3 is a schematic diagram of window division according to the present invention;

4 is a schematic diagram of data analysis of the present invention, FIG. 4(a) is resistance data in a test example, 4(b) is logarithmic data in a test example, and FIG. 4(c) is logarithm in a test example. Absolute difference data;

Figure 5 is a schematic diagram of the system of the present invention;

Figure 6 shows the variation of the MeanDiff value in a test case;

Figure 7 is a view showing the detection of the micro-sleep state in a test case;

Fig. 8 is a view showing the detection of the sleep state in one test example.

detailed description

The present invention will be further specifically described below in conjunction with the drawings and specific embodiments.

Example 1:

1 is a transition state of a human physiological state according to the present invention, including a transition from a normal living state to a micro-sleep state, a transition from a micro-sleep state to a sleep state, and a transition from a micro-sleep state or a sleep state to a normal living state. .

A method for detecting a physiological state transition of a human body comprises the following steps: parameter initialization, collecting body resistance data, analyzing body resistance data, and outputting a physiological state transition state.

2 is a flow chart of analyzing human body resistance data in the method of the present invention, including preprocessing, logarithm, window processing, logarithmic difference comparison, calculation of convergence value, and analysis of physiological state transition. When testing, first read the tester's skin resistance data, perform pre-processing, and then the data goes through a logarithmic process to get the current time. The resistance logarithm value. The window processing is to average the logarithm of the resistance in the window. The logarithmic difference comparison is the difference between the logarithm of the current time and the logarithm of the resistance of the previous window. The difference is input to a convergence function, and after a period of time, a convergence value is obtained, and the convergence value is used as a basis for determining the physiological state transition. Each step is described in detail below.

1. Pretreatment:

Since the skin resistance collection instrument has a specified range, the resistance value collected after the range is out of range, or because of poor contact, the default resistance value will be broken, so the acquired resistance data needs to be preprocessed. The value of the short-lived resistance in the open state is replaced with the resistance value at the previous moment. Secondly, when the finger just touches the instrument, an unstable glitch signal is generated, and filtering is also required. The resistance data in a test example is shown in Figure 4(a).

2, take the logarithm:

In order to better reflect the change in the intensity of the skin resistance, a logarithmic operation is performed on the pre-processed resistance data. Data(i) is the resistance value at the current time after preprocessing, and the natural logarithm is obtained to obtain ln(data(i)). Shown in Figure 4(b) is the logarithmic data in a test case.

3, window processing

The resistance is analyzed to determine the state change of the tester. If a single data is used in the analysis, the instantaneous data change will be treated as a state change, so that the noise has a strong interference with the analysis. When comparing the differences later, if the resistance at the previous moment includes a relatively strong noise, the result will be inaccurate. Therefore, it is necessary to select a certain window for averaging to reduce fluctuations and facilitate analysis. Figure 3 is a schematic diagram of window division.

The resistance data of the entire window is logarithm, the length of the window is WinLength, and the data after the logarithm is ln(data(i-WinLength)), ln(data(i-WinLength+1)),... ....ln(data(i-1)), ln(data(i)). The logarithm of the resistance in one window before the current resistance value is averaged to obtain the log mean MeanSum(i).

(Formula 1)

The window length WinLength is set to 100, and the resistance sampling frequency is 50 Hz, that is, a window covers resistance data within 2 seconds.

4, logarithmic difference comparison

The reason we use logarithmic operation is that after logarithmic processing, the difference is reflected in the ratio of the skin resistance. The logarithmic value of the current time and the logarithmic mean of the previous window are used to obtain a logarithmic difference Diff(i).

Diff(i)=ln(data(i))-MeanSum(i) (Equation 2)

Taking the absolute value gives the logarithmic absolute difference AbsDiff(i).

AbsDiff(i)=abs(Diff(i)) (Equation 3)

Shown in Figure 4(c) is the logarithmic absolute difference data in a test case.

5, calculate the convergence value

Define a variable MeanDiff and a convergence function as follows,

MeanDiff=α×MeanDiff+(1-α)×AbsDiff(i) (Equation 4)

Under the action of the convergence function, the variable MeanDiff reflects the cumulative variation of the logarithmic absolute difference AbsDiff. The initial value of the variable MeanDiff is set to 0.03 and α is set to 0.999.

6. Analysis of physiological state transition

Under normal living and working conditions, any slight emotional change can cause the difference between the logarithm of the resistance at the current time and the logarithmic mean of the previous window. The absolute value of the logarithmic absolute difference AbsDiff is large, and MeanDiff cannot converge to a small Value. In the micro-sleep state, the nervous system is in a relative inhibition state, the skin sweat gland secretion is in a short-term stable state, the logarithmic absolute difference AbsDiff value decreases, and the MeanDiff converges to a small value. After the micro-sleep state, entering the sleep state, the logarithmic absolute difference AbsDiff value is nearly zero, and the MeanDiff converges to a smaller value. From the micro-sleep state or the sleep state to the normal living and working state, the nervous system recovers the excited state in an instant, the skin sweat gland is secreted, and the logarithmic absolute difference AbsDiff value increases instantaneously, and the MeanDiff value increases in an instant.

At the same time, different testers have different physical conditions, some people are relatively sensitive to the nervous system and sweat gland system, and are relatively easy to be excited, while others are relatively insensitive and difficult to reach the state of excitement. According to different physical conditions, different physiological state transition analysis parameters are set to overcome individual differences. In order to judge different physical conditions, the window processing method is used again, and a certain window StepSize is selected, and the physical condition is judged by analyzing the resistance data in the window. By sliding the window, you get the following two data:

mSlope _up = max{slope=(data(i+StepSize-1)-data(i))/StepSize|i=1,2...} (Equation 5)

mSlope _down = max{slope=(data(i)-data(i+StepSize-1))/StepSize|i=1,2...} (Equation 6)

The rules for judging different physical conditions are as follows:

体质状况Physical condition	判断标准Judging criteria
体质状况Physical condition	判断标准Judging criteria	相对敏感Relatively sensitive	mSlope_up>δAND mSlope_down>δmSlope _up >δAND mSlope _down >δ
不敏感Not sensitive	mSlope_up<δOR mSlope_down<δmSlope _up <δOR mSlope _down <δ	相对敏感Relatively sensitive

The window length StepSize is set to 10 and δ is set to 15.

The physiological state transition can be judged by setting the MeanDiff threshold for different physical conditions. For relatively sensitive physical conditions, the threshold MeanDiff _{sen, microsleep is} used to judge the tester's transition from the normal living state to the micro-sleep state. The threshold MeanDiff _{sen, sleep is} used to judge the tester's transition from the micro-sleep state to the sleep state. For insensitive physical conditions, the threshold MeanDiff _{insen, microsleep is} used to judge the tester's transition from the normal living state to the micro-sleep state. The threshold MeanDiff _{insen, sleep is} used to judge the tester's transition from the micro-sleep state to the sleep state. The specific rules of judgment are as follows:

The threshold MeanDiff _{sen, microsleep is} set to 0.0006, MeanDiff _{sen, sleep is} set to 0.0002, MeanDiff _{insen, microsleep is} set to 0.001, MeanDiff _{insen, and sleep is} set to 0.0005.

Since conductance is the reciprocal of electrical resistance, the above analytical method is equally applicable to the analysis of physiological physiological state transitions based on conductance changes.

The invention relates to an emotion detecting system based on human skin resistance change, which comprises a medical pole piece connected in series, a bridge resistance/conductance measuring circuit, an amplifying circuit, an A/D conversion circuit, a CPU and a human-computer interaction interface.

FIG. 5 is a schematic diagram of a system for detecting a physiological state transition of a human body, including a medical pole piece, a bridge resistance/conductance measuring circuit, an amplifying circuit, an A/D conversion circuit, a control device, and a human-machine interaction interface. The medical pole piece and the bridge type resistance/conductance measuring circuit are used for collecting human skin resistance (conductance) data, and then detecting human physiological state transition based on human skin resistance (conductance) change, and realizing the method of the present invention. The control device can select a device with data processing capability such as a single chip microcomputer, a mobile communication device, a mobile computer device or a desktop computer device.

Example 2:

Figure 6 shows the variation of the MeanDiff value in a test case. According to the law of physical condition judgment, the tester is judged to be relatively sensitive, so the thresholds MeanDiff _{sen, microsleep} and MeanDiff _{sen, sleep are used to} judge the change of the physiological state of the tester.

It is judged that the value of MeanDiff is less than or equal to MeanDiff _{sen, microsleep} , that is, the human body changes from a normal living state to a micro-sleep state. Figure 7 is a diagram showing the detection of the micro-sleep state in a test example, in which the abscissa indicates time; the ordinate indicates that the value of MeanDiff is less than or equal to 0.0006 (i.e. _, the value of MeanDiff _{sen, microsleep} ), that is, the micro-sleep state or Sleep state; the ordinate is 0 means that the MeanDiff value is greater than 0.0006 (ie _, the value of MeanDiff _{sen, microsleep} ), that is, the normal living state.

It is judged that the value of MeanDiff is less than or equal to MeanDiff _{sen, sleep} , that is, the human body transitions from the micro sleep state to the sleep state. Fig. 8 is a view showing the detection of the sleep state in a test example, in which the abscissa indicates time and the ordinate indicates that the value 1 indicates that the value of MeanDiff is less than or equal to 0.0002 (i.e. _, the value of MeanDiff _{sen, sleep} ), that is, the sleep state, vertical A coordinate of 0 indicates a case where the value of MeanDiff is greater than 0.0002 (i.e. _, the value of MeanDiff _{sen, sleep} ), that is, a micro sleep state or a normal living state.

Claims

A method for detecting a physiological state transition of a human body, comprising the steps of: initializing parameters, collecting body resistance data, analyzing body resistance data, and outputting a physiological state transition state;

Analysis of human body resistance data includes the following steps:

Pre-processing the collected human body resistance data; taking the natural logarithm of the pre-processed human body resistance data to obtain the resistance logarithm value; averaging the resistance pair values in the previous window at the current time to obtain the current window in the previous window The resistance logarithmic mean MeanSum(i), where the window size is WinLength; the difference between the resistance log ln(data(i)) at the current time and the logarithmic mean value MeanSum(i) in the previous window Diff(i); takes the absolute value of Diff(i) to obtain the absolute difference of the log AbsDiff(i); inputs the logarithmic absolute difference AbsDiff(i) into the convergence function to obtain the convergence value MeanDiff; and judges the physiology according to the convergence value MeanDiff State transition situation.
A method of detecting a physiological state transition of a human body according to claim 1, wherein

The convergence function is:

MeanDiff=α×MeanDiff+(1-α)×AbsDiff(i)

α is a convergence parameter, 0 < α < 1.
The method for detecting a physiological state transition of a human body according to claim 2, wherein the determining a physiological state transition according to a convergence value MeanDiff is as follows:

If MeanDiff>MeanDiff microsleep , the tester enters the normal living and working state;

If MeanDiff microsleep ≥MeanDiff>MeanDiff sleep , the tester enters the micro-sleep state;

If MeanDiff ≤ MeanDiff sleep , the tester enters a sleep state; MeanDiff microsleep is used to judge the tester's transition threshold from the normal living working state to the micro-sleep state, and MeanDiff sleep is used to judge the tester to go to sleep from the micro-sleep state. The threshold for the transition of the state.
The method for detecting a physiological state transition of a human body according to claim 3, wherein different values of MeanDiff microsleep and MeanDiff sleep are set for different physical conditions of the tester;

The tester's physical condition is judged based on the following two indicators:

mSlope up =max{slope=(data(i+StepSize-1)-data(i))/StepSize|i=1,2...}

mSlope down = max{slope=(data(i)-data(i+StepSize-1))/StepSize|i=1,2...}

Where slope is the slope, indicating the average change value of the resistance, calculated according to the resistance value and the window size of the resistance curve at both ends of the window; data(i) is the resistance value at time i, and data(i+StepSize-1) is i+ The resistance value at StepSize-1; StepSize is the window size;

mSlope up is the maximum value of the calculated i-up slopes. The upward slope is the last resistance value data(i+StepSize-1) in a certain window minus the first resistance value data by the resistance curve. i) the obtained slope value; mSlope down is the maximum value of the calculated i-down slopes, and the down-slope is the first resistance value data(i) in the window with the resistance curve calculated by subtracting the last one The slope value obtained by the resistance value data(i+StepSize-1);

If mSlope up > δ, and mSlope down > δ, the tester is judged to be relatively sensitive;

If mSlope up <δ, or mSlope down <δ, it is judged that the tester is insensitive to the constitution; wherein δ is the judgment threshold.
The method for detecting a physiological state transition of a human body according to claim 4, wherein in the parameter initialization, WinLength is set to 100, the resistance sampling frequency is 50 Hz; the initial value of MeanDiff is 0.03, and α is 0.999;

Set the window size StepSize to 10 and δ to 15;

For testers with relatively sensitive physique, the thresholds MeanDiff microsleep and MeanDiff sleep were set to 0.0006 and 0.0002, respectively; for testers with insensitive physique, the thresholds MeanDiff microsleep and MeanDiff sleep were set to 0.001 and 0.0005, respectively.
The method for detecting a physiological state transition of a human body according to any one of claims 1 to 5, wherein the output physiological state transition state is a state in which a physiological state transition is reported by sound, light, vibration or odor.
The method for detecting a physiological state transition of a human body according to any one of claims 1 to 5, wherein the collecting the body resistance data is performed by collecting human body conductance data, and calculating a resistance value according to a reciprocal relationship between the resistance and the conductance.
A system for detecting a physiological state transition of a human body, comprising: a medical pole piece connected in series, a bridge resistance/conductance measuring circuit, an amplifying circuit, an A/D conversion circuit, a control device, and a human-machine interaction interface;

The bridge resistor/conductance measuring circuit is configured to collect body resistance/conductance data, and the control device is configured to perform the method of any one of claims 1 to 5 to detect a physiological state transition of the human body; The human-computer interaction interface outputs the physiological state transition of the human body to the user.
The system for detecting a physiological state transition of a human body according to claim 8, wherein the control device is a single chip microcomputer, a mobile communication device, a mobile computer device, or a desktop computer device.
The system for detecting a physiological state transition of a human body according to claim 8 or 9, wherein the human-computer interaction interface comprises a voice module, a display module, a vibration module or an odor generating module.