Sleeping pillow comfort evaluation method based on multi-source data monitoring

Abstract

The present invention discloses a pillow comfort evaluation method based on multi-source data monitoring, which relates to the technical field of pillow evaluation, obtains the influence of sensor data and physiological data on comfort, generates corresponding weights for each sensor data and physiological data, integrates the sensor data and physiological data with corresponding weights to obtain an initial comfort score, uses an MCMC algorithm to perform multiple sampling in the probability distribution of each sensor data and physiological data, regenerates the weight of each sensor data and physiological data according to the sampling result after each sampling, recalculates the comfort score based on the newly generated weight, and judges whether to accept the newly generated weight based on the acceptance probability criterion of the MCMC algorithm. Through multi-source data and objective feedback, the multi-source data weight parameters are dynamically adjusted to generate a personalized comfort evaluation for each user, provide personalized pillow design suggestions, and significantly improve the user's comfort.

Description

Technical Field

[0001] The present invention relates to the technical field of pillow evaluation, and in particular to a pillow comfort evaluation method based on multi-source data monitoring. Background Art

[0002] A pillow is a common sleeping tool used to support the head and neck. It is designed to improve sleep quality, provide comfort, and prevent or relieve neck pain. Its design has a long history. The earliest pillow can be traced back to ancient civilizations. Over time, the shape, material, and function of the pillow have gradually evolved. The pillow comfort evaluation system is usually designed based on ergonomics, material science, and sensing technology to provide users with a more comfortable sleeping experience.

[0003] The prior art has the following defects:

[0004] Traditional pillow designs are often based on universal standards and lack response to the personalized needs of different users (for example, the differences in sensitivity to pressure, temperature, and humidity among different users cannot be fully considered), which leads to poor experience for many users in actual use. They also usually rely on a single data source or limited physiological indicators (such as only detecting head pressure), and fail to fully integrate the user's various sleep data, resulting in inaccurate evaluation results.

[0005] Based on this, the present invention proposes a sleeping pillow comfort evaluation method based on multi-source data monitoring. Through multi-source data and objective feedback, the multi-source data weight parameters are dynamically adjusted to generate a personalized comfort evaluation for each user, provide personalized pillow design suggestions, and significantly improve the user's comfort. Summary of the invention

[0006] The purpose of the present invention is to provide a pillow comfort evaluation method based on multi-source data monitoring to address the deficiencies in the background technology.

[0007] In order to achieve the above object, the present invention provides the following technical solution: a pillow comfort evaluation method based on multi-source data monitoring, the evaluation method comprising the following steps:

[0008] The evaluation system collects sensor data from several pillows during the test process, monitors the physiological data of users when testing the pillows through monitoring equipment, analyzes historical data based on the probability distribution model, obtains the impact of sensor data and physiological data on comfort, and generates corresponding weights for each sensor data and physiological data. After integrating the sensor data and physiological data with the corresponding weights, the initial comfort score is obtained;

[0009] Use the MCMC algorithm to perform multiple sampling in the probability distribution of each sensor data and physiological data. After each sampling, regenerate the weight of each sensor data and physiological data according to the sampling result, recalculate the comfort score based on the newly generated weight, and use the newly acquired comfort score based on the acceptance probability criterion of the MCMC algorithm to determine whether to accept the newly generated weight;

[0010] When any iteration satisfies the convergence regulation, the comfort score obtained in this iteration and the weight corresponding to each sensor data and physiological data are output, and the comfort of the pillow is evaluated based on the final comfort score.

[0011] In a preferred embodiment, the evaluation system collects sensor data of several pillows during the testing process, and monitors the physiological data of the user when testing the pillows through monitoring equipment. The sensor data includes the pressure fluctuation accumulation coefficient of multiple sensing points and the thermal conductivity coefficient of multiple sensing points, and the physiological data includes the frequency of body position changes and the skin electrical response factor.

[0012] In a preferred embodiment, after analyzing historical data based on a probability distribution model, the influence of sensor data and physiological data on comfort is obtained, and corresponding weights are generated for each sensor data and physiological data, including the following steps:

[0013] A certain number of pillows are selected from a batch of pillows to establish a pillow set, and the multi-sensing point pressure fluctuation accumulation coefficient, multi-sensing point heat conductivity coefficient, body position change frequency and skin electrical response factor of each pillow in the pillow set are recorded;

[0014] The probability distribution model performs normal distribution on the cumulative coefficients of pressure fluctuations at multiple sensing points and the thermal conductivity coefficients at multiple sensing points of multiple pillows. The probability distribution of the cumulative coefficients of pressure fluctuations at multiple sensing points is N y (A avg ,A Q ), A avg is the average value of the cumulative coefficient of pressure fluctuation at multiple sensing points, A Q is the standard deviation of the cumulative coefficient of pressure fluctuation at multiple sensing points, and the probability distribution of the heat transfer coefficient at multiple sensing points is N r (B avg ,B Q ), B avg is the average value of the thermal conductivity coefficient of multiple sensing points, B Q is the standard deviation of the heat transfer coefficient at multiple sensing points. The Poisson distribution of the body position change frequency of multiple pillows is used. The probability distribution of the body position change frequency is P(C avg ), C avg is the mean frequency of body position change, and the skin galvanic response factors of multiple pillows are exponentially distributed. The probability distribution of the skin galvanic response factor is E(D avg ), Davg is the mean value of the galvanic skin response factor;

[0015] Calculate the Pearson correlation coefficient between each sensor data and physiological data and the user's deep sleep duration. The expression is: In the formula, r is the Pearson correlation coefficient, X i is the current data of the i-th pillow, Y i is the deep sleep duration of the user of the i-th pillow, is the current data mean, The average deep sleep duration of the user;

[0016] The total value of the Pearson correlation coefficient was obtained by summing up the cumulative coefficient of pressure fluctuation at the sensing point, the thermal conductivity coefficient at multiple sensing points, the frequency of body position change, and the skin electrical response factor. The weight was obtained by dividing the Pearson correlation coefficient by the total value of the Pearson correlation coefficient.

[0017] In a preferred embodiment, the sensor data and the physiological data are combined with corresponding weights to obtain an initial comfort score, which includes the following steps:

[0018] The cumulative coefficient of pressure fluctuation at multiple sensing points, the thermal conductivity coefficient at multiple sensing points, the frequency of body position change and the galvanic skin response factor of the current pillow are obtained, and the cumulative coefficient of pressure fluctuation at multiple sensing points, the thermal conductivity coefficient at multiple sensing points, the frequency of body position change and the galvanic skin response factor are normalized so that their value ranges are mapped to [0,1]. The weights of the cumulative coefficient of pressure fluctuation at multiple sensing points, the weights of the thermal conductivity coefficient at multiple sensing points, the weights of the frequency of body position change and the weights of the galvanic skin response factor are obtained through the probability distribution model, and then the initial comfort score is calculated.

[0019] In a preferred embodiment, the MCMC algorithm is used to perform multiple sampling in the probability distribution of each sensor data and physiological data, and the weight of each sensor data and physiological data is regenerated according to the sampling result after each sampling, including the following steps:

[0020] Randomly select an initial value from the probability distribution of each data as the sampling starting point, marked as state;

[0021] Using the Markov Chain Monte Carlo algorithm, starting from the current state of each parameter, multiple samplings are performed according to its probability distribution to generate a series of states;

[0022] According to the objective function, whether the new sampled state should be accepted is determined. The objective function expression is: In the formula, CFT(state new ) is the comfort score calculated under the new state, CFT(state old) is the comfort score calculated under the current state, min(*) means selecting the minimum value, J(accepts state new ) represents the probability of accepting the new state;

[0023] If J(accepts state new )≥80%, judge to accept the new state of the current data;

[0024] Calculate the comfort score difference of the current data and the new state, the expression is: C = CFT (state new )-CFT(state old ), where C is the comfort score difference of the new state, CFT(state new ) is the comfort score calculated under the new state, CFT(state old ) is the comfort score calculated in the current state;

[0025] The weight adjustment coefficient is obtained by summing up the difference of all new state comfort scores of the current data. The expression is: In the formula, C z is the weight adjustment coefficient, m is the number of times the new state of the current data is accepted, C i is the comfort score difference of the i-th new state of the current data;

[0026] If the weight adjustment coefficient of the current data is greater than or equal to the preset adjustment threshold, the weight of the current data is increased; if the weight adjustment coefficient of the current data is less than the preset adjustment threshold, the weight of the current data is decreased.

[0027] In a preferred embodiment, if the weight adjustment coefficient of the current data is greater than or equal to the preset adjustment threshold, the weight of the current data is increased; if the weight adjustment coefficient of the current data is less than the preset adjustment threshold, the weight of the current data is decreased. The adjustment algorithm is:

[0028] In the formula, ω new is the weight adjusted by the current data, ω old is the weight before current data adjustment, C z is the weight adjustment coefficient, and F is the coefficient threshold.

[0029] In a preferred embodiment, the logic for obtaining the pressure fluctuation accumulation coefficient of multiple sensing points is as follows: pressure strain gauges are set at multiple monitoring points of the pillow. During the test, a constant pressure is applied to multiple monitoring points by a pressure testing device. After recording the actual pressure of each monitoring point, the pressure standard deviation is calculated based on the actual pressure of each monitoring point. The pressure standard deviation is integrated and accumulated over a period of time to obtain the pressure fluctuation accumulation coefficient of multiple sensing points. The expression is: Where ps is the cumulative coefficient of pressure fluctuations at multiple sensing points, yb(t) is the standard deviation of pressure at time t, and T is the monitoring duration.

[0030] The logic for obtaining the multi-sensing point heat transfer coefficient is as follows: temperature sensors are set at multiple monitoring points of the pillow, and after heating one side of the pillow, the heat flux density at multiple monitoring points is calculated. The expression is: In the formula, q is the heat flux density, k is the thermal conductivity of the pillow material, ΔT is the temperature difference between the pillow surface and the heated side, and Δx is the thickness of the pillow. The heat flux density of multiple monitoring points is summed to obtain the multi-sensing point thermal conductivity coefficient.

[0031] In a preferred embodiment, the logic for obtaining the frequency of body position changes is: 2 minutes after the user enters deep sleep, the polysomnography monitor is used to record the number of body position changes of the user within a period of time, and the frequency of body position changes is obtained by dividing the number of body position changes by the recorded time.

[0032] The logic for obtaining the galvanic skin response factor is as follows: 2 minutes after the user enters deep sleep, the polysomnography monitor is used to detect that the user has placed electrodes on the fingers, palms, or feet, and a voltage of 0.5-5 volts is applied between the electrodes to measure the conductivity of the skin. The change in skin conductivity is continuously measured to generate a time series signal, and the skin conductivity at multiple time points is obtained. The skin conductivity at multiple time points is subtracted and the absolute value is taken as the galvanic skin response factor.

[0033] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0034] After analyzing historical data through a probability distribution model, the present invention obtains the influence of sensor data and physiological data on comfort, generates corresponding weights for each sensor data and physiological data, integrates the sensor data and physiological data with the corresponding weights to obtain an initial comfort score, uses an MCMC algorithm to perform multiple samplings in the probability distribution of each sensor data and physiological data, regenerates the weight of each sensor data and physiological data according to the sampling results after each sampling, recalculates the comfort score based on the newly generated weight, and determines whether to accept the newly generated weight based on the acceptance probability criterion of the MCMC algorithm. Through multi-source data and objective feedback, the multi-source data weight parameters are dynamically adjusted to generate a personalized comfort evaluation for each user, provide personalized pillow design suggestions, and significantly improve the user's comfort. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0036] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] Example 1: Please refer to Figure 1 As shown, the pillow comfort evaluation method based on multi-source data monitoring described in this embodiment includes the following steps:

[0039] The evaluation system collects sensor data from several pillows during the test process, monitors the physiological data of users when testing the pillows through monitoring equipment, obtains the impact of sensor data and physiological data on comfort after analyzing historical data based on the probability distribution model, and generates corresponding weights for each sensor data and physiological data. The sensor data and physiological data are combined with the corresponding weights to obtain the initial comfort score, and the MCMC algorithm is used to perform multiple sampling in the probability distribution of each sensor data and physiological data. After each sampling, the weight of each sensor data and physiological data is regenerated according to the sampling results, and the comfort score is recalculated based on the newly generated weight. The newly acquired comfort score is used to determine whether to accept the newly generated weight based on the acceptance probability criterion of the MCMC algorithm. When any iteration meets the convergence adjustment, the comfort score obtained in this iteration and the corresponding weight of each sensor data and physiological data are output, and the comfort of the pillow is evaluated based on the finally obtained comfort score.

[0040] After analyzing historical data through a probability distribution model, this application obtains the impact of sensor data and physiological data on comfort, generates corresponding weights for each sensor data and physiological data, integrates the sensor data and physiological data with the corresponding weights to obtain an initial comfort score, uses the MCMC algorithm to perform multiple sampling in the probability distribution of each sensor data and physiological data, regenerates the weight of each sensor data and physiological data based on the sampling results after each sampling, recalculates the comfort score based on the newly generated weight, and uses the newly acquired comfort score based on the acceptance probability criterion of the MCMC algorithm to determine whether to accept the newly generated weight. Through multi-source data and objective feedback, the multi-source data weight parameters are dynamically adjusted to generate a personalized comfort evaluation for each user, provide personalized pillow design suggestions, and significantly improve the user's comfort.

[0041] The assessment system includes:

[0042] Data collection module: collects sensor data (including pressure, temperature, humidity, posture and other sensor data) of several pillows during the test process, monitors the physiological data of users when testing the pillows through monitoring equipment, and sends the sensor data and physiological data to the initial score generation module;

[0043] Initial score generation module: After analyzing historical data based on the probability distribution model, the impact of sensor data and physiological data on comfort is obtained, and corresponding weights are generated for each sensor data and physiological data. The initial comfort score is obtained by combining the sensor data and physiological data with the corresponding weights. The initial comfort score and weight are sent to the iterative update module;

[0044] Iterative update module: Use the MCMC algorithm to perform multiple sampling in the probability distribution of each sensor data and physiological data. After each sampling, the weight of each sensor data and physiological data is regenerated according to the sampling result, and the comfort score is recalculated based on the newly generated weight. The newly acquired comfort score is used to determine whether to accept the newly generated weight based on the acceptance probability criterion of the MCMC algorithm. The comfort score and weight are sent to the comprehensive evaluation module;

[0045] Comprehensive evaluation module: When any iteration satisfies the convergence adjustment, the comfort score obtained in this iteration and the corresponding weight of each sensor data and physiological data are output, and the comfort of the pillow is evaluated based on the final comfort score.

[0046] Embodiment 2: The evaluation system collects sensor data of several pillows during the testing process, and monitors the physiological data of the user when testing the pillows through the monitoring device, including the following steps:

[0047] The sensing data include the pressure fluctuation accumulation coefficient of multiple sensing points and the thermal conductivity coefficient of multiple sensing points, and the physiological data include the frequency of body position changes and the skin electrical response factor;

[0048] The logic for obtaining the cumulative coefficient of pressure fluctuations at multiple sensing points is as follows: pressure strain gauges are set at multiple monitoring points of the pillow. During the test, constant pressure is applied to multiple monitoring points through the pressure testing equipment. After recording the actual pressure at each monitoring point, the pressure standard deviation is calculated based on the actual pressure at each monitoring point. The pressure standard deviation is integrated and accumulated over a period of time to obtain the cumulative coefficient of pressure fluctuations at multiple sensing points. The expression is: In the formula, ps is the cumulative coefficient of pressure fluctuations at multiple sensing points, yb(t) is the standard deviation of pressure at time t, and T is the monitoring duration. The larger the cumulative coefficient of pressure fluctuations at multiple sensing points, the worse the comfort of the pillow.

[0049] The human body wants to get even pressure distribution when sleeping. If the pressure distribution on the pillow surface is uneven, the pressure in some areas is too high or fluctuates frequently, which will cause discomfort to the user. The greater the cumulative coefficient of pressure fluctuation, the less support the pillow can provide, causing the user to constantly adjust their posture to relieve local pressure, which in turn affects the quality of sleep. High-frequency pressure fluctuations may cause poor local blood circulation and cause soreness in the long run, especially in the neck and head pressure-sensitive areas.

[0050] The logic for obtaining the multi-sensing point heat conduction coefficient is as follows: temperature sensors are set at multiple monitoring points of the pillow, and after heating one side of the pillow, the heat flux density at multiple monitoring points is calculated. The expression is: In the formula, q is the heat flux density, k is the thermal conductivity of the pillow material, ΔT is the temperature difference between the pillow surface and the heated side, and Δx is the thickness of the pillow. The thermal conductivity of the pillow material can be determined experimentally. Usually, pillow materials such as memory foam, latex, and down have their corresponding thermal conductivity coefficients. The heat flux density of multiple monitoring points is summed to obtain the multi-sensing point thermal conductivity coefficient. The smaller the multi-sensing point thermal conductivity coefficient, the worse the comfort of the pillow.

[0051] A low thermal conductivity means that the pillow cannot dissipate heat effectively. During sleep, the human body generates heat. If the thermal conductivity of the pillow material is poor, the heat cannot be quickly conducted and dissipated, resulting in local overheating and increased discomfort. High temperatures can cause users to feel stuffy and affect deep sleep. In addition, overheating may cause uncomfortable reactions such as sweating, which in turn affects comfort.

[0052] The logic for obtaining the frequency of body position changes is as follows: 2 minutes after the user enters deep sleep, the polysomnography monitor is used to record the number of body position changes within a period of time, and the frequency of body position changes is obtained by dividing the number of body position changes by the recorded duration. The greater the frequency of body position changes, the worse the comfort of the pillow.

[0053] A comfortable sleeping pillow should be able to stabilize the head and neck, allowing the user to maintain a comfortable state for a long time in the same position. If the user changes position frequently, it means that the sleeping pillow cannot provide enough support or there is local discomfort, causing the user to unconsciously adjust the posture to relieve the discomfort. Frequent changes in body position will affect the deep sleep stage, disrupt the user's sleep cycle, and reduce the overall sleep quality.

[0054] The logic of obtaining the skin electrical response factor is: 2 minutes after the user enters deep sleep through a polysomnography monitor, electrodes are placed on the fingers, palms, feet, etc. to obtain it. A very small voltage (usually harmless direct current or alternating current), usually 0.5-5 volts, is applied between the electrodes. This helps to form an electrical circuit on the surface of the skin to measure the conductivity of the skin, which changes with the increase in sweat gland activity. When the sympathetic nervous system is activated (such as stress, emotional fluctuations, external stimuli, etc.), sweat gland secretion increases, the conductivity of the skin is enhanced, and the skin conductivity increases. The sensor continuously measures the changes in skin conductivity and generates a time series signal. These data can reflect the dynamic changes of skin conductivity over time. Usually, the rapid changes of skin conductivity reflect the intensity of emotional fluctuations or stress reactions. The skin conductivity fluctuation state over a period of time is calculated to obtain the skin galvanic response factor, that is, the skin conductivity at multiple time points is obtained, the skin conductivity at multiple time points is subtracted and the absolute value is taken as the skin galvanic response factor. The larger the skin galvanic response factor, the worse the comfort of the pillow.

[0055] The galvanic skin response factor (GSR) reflects the activity of the sympathetic nervous system, which is usually related to stress or discomfort. When the user is in a state of discomfort or stress, the sympathetic nerve activity is enhanced, resulting in an increase in skin conductivity. If the user's galvanic skin response factor is high when using a sleeping pillow, it means that the user may feel uncomfortable with the pillow material, temperature or pressure, which in turn triggers a stress response. A high galvanic skin response factor indicates that the user may be in an emotionally or physically uncomfortable state, which directly affects the user's relaxation level and sleep quality.

[0056] After analyzing historical data based on the probability distribution model, the influence of sensor data and physiological data on comfort is obtained, and corresponding weights are generated for each sensor data and physiological data, including the following steps:

[0057] A certain number (e.g., 20) of pillows are selected from a batch of pillows to establish a pillow set, and the multi-sensing point pressure fluctuation accumulation coefficient, multi-sensing point heat conductivity coefficient, body position change frequency, and skin electrical response factor of each pillow in the pillow set are recorded;

[0058] The probability distribution model obtains the mean value of the cumulative coefficients of pressure fluctuations at multiple sensing points and the standard deviation of the cumulative coefficients of pressure fluctuations at multiple sensing points according to the cumulative coefficients of pressure fluctuations at multiple sensing points of multiple pillows, obtains the mean value of the thermal conductivity coefficients at multiple sensing points and the standard deviation of the thermal conductivity coefficients at multiple sensing points according to the thermal conductivity coefficients at multiple sensing points of multiple pillows, obtains the mean value of the body position change frequency according to the body position change frequency of multiple pillows, and obtains the mean value of the skin electrical response factor according to the skin electrical response factors of multiple pillows;

[0059] The pressure fluctuation accumulation coefficients at multiple sensing points and the heat conduction coefficients at multiple sensing points of multiple pillows are normally distributed. The probability distribution of the pressure fluctuation accumulation coefficients at multiple sensing points is N y (A avg ,A Q ), A avg is the average value of the cumulative coefficient of pressure fluctuation at multiple sensing points, A Q is the standard deviation of the cumulative coefficient of pressure fluctuation at multiple sensing points, and the probability distribution of the heat transfer coefficient at multiple sensing points is N r (B avg ,B Q ), B avg is the average value of the thermal conductivity coefficient of multiple sensing points, B Q is the standard deviation of the heat transfer coefficient at multiple sensing points. The Poisson distribution of the body position change frequency of multiple pillows is used. The probability distribution of the body position change frequency is P(C avg ), C avg is the mean frequency of body position change, and the skin galvanic response factors of multiple pillows are exponentially distributed. The probability distribution of the skin galvanic response factor is E(D avg ), D avg is the mean value of the galvanic skin response factor;

[0060] Calculate the Pearson correlation coefficient between each sensor data and physiological data and the user's deep sleep duration. The expression is: In the formula, r is the Pearson correlation coefficient, X i is the current data of the ith pillow (pressure fluctuation accumulation coefficient of the sensing point, thermal conductivity coefficient of multiple sensing points, frequency of body position change or skin electrical response factor), Y i is the deep sleep duration of the user of the i-th pillow, is the current data mean, is the mean deep sleep duration of the user. The larger the Pearson correlation coefficient is, the greater the correlation between the current data and the user's deep sleep duration is, that is, the greater the impact of the current data on the user's deep sleep duration is.

[0061] The Pearson correlation coefficients of the cumulative coefficient of pressure fluctuation at the sensing point, the thermal conductivity coefficient of multiple sensing points, the frequency of body position change and the skin electrical response factor were summed to obtain the total value of the Pearson correlation coefficient. The weight was obtained by dividing the Pearson correlation coefficient by the total value of the Pearson correlation coefficient. The larger the weight, the greater the impact.

[0062] The sensor data and physiological data are combined with corresponding weights to obtain the initial comfort score, including the following steps:

[0063] The multi-sensory point pressure fluctuation cumulative coefficient, multi-sensory point heat conduction coefficient, body position change frequency and skin galvanic response factor of the current pillow are obtained, and the multi-sensory point pressure fluctuation cumulative coefficient, multi-sensory point heat conduction coefficient, body position change frequency and skin galvanic response factor are normalized so that their value ranges are mapped to [0,1]. The weights of the multi-sensory point pressure fluctuation cumulative coefficient, the weights of the multi-sensory point heat conduction coefficient, the weights of the body position change frequency and the weights of the skin galvanic response factor are obtained through the probability distribution model, and the initial comfort score is calculated. The expression is: In the formula, CFT 初始 is the initial comfort score, q z , ps, tb, and pd are respectively the multi-sensing point heat conduction coefficient, the multi-sensing point pressure fluctuation accumulation coefficient, the body position change frequency, and the skin electrical response factor, ω 1 ,ω 2 ,ω 3 ,ω 4 They are the heat conduction coefficient of multiple sensing points, the cumulative coefficient of pressure fluctuations at multiple sensing points, the frequency of body position changes, and the weights of the skin electrical response factor.

[0064] The MCMC algorithm is used to perform multiple sampling in the probability distribution of each sensor data and physiological data. After each sampling, the weight of each sensor data and physiological data is regenerated according to the sampling result, and the comfort score is recalculated based on the newly generated weight. The newly acquired comfort score is used to determine whether to accept the newly generated weight based on the acceptance probability criterion of the MCMC algorithm, including the following steps:

[0065] Randomly select an initial value from the probability distribution of each data as the sampling starting point, marked as state;

[0066] Using Markov Chain Monte Carlo Using the Markov Chain Monte Carlo (MCMC) algorithm, starting from the current state of each parameter, multiple samplings are performed according to its probability distribution to generate a series of states;

[0067] According to the objective function, whether the new sampled state should be accepted is determined. The objective function expression is: In the formula, CFT(state new) is the comfort score calculated under the new state, CFT(state old ) is the comfort score calculated under the current state, min(*) means selecting the minimum value, J(accepts state new ) represents the probability of accepting the new state;

[0068] If J(accepts state new )≥80%, judge to accept the new state of the current data;

[0069] Calculate the comfort score difference of the current data and the new state, the expression is: C = CFT (state new )-CFT(state old ), where C is the comfort score difference of the new state, CFT(state new ) is the comfort score calculated under the new state, CFT(state old ) is the comfort score calculated in the current state;

[0070] The weight adjustment coefficient is obtained by summing up the difference of all new state comfort scores of the current data. The expression is: In the formula, C z is the weight adjustment coefficient, m is the number of times the new state of the current data is accepted, C i is the comfort score difference of the i-th new state of the current data;

[0071] If the weight adjustment coefficient of the current data is greater than or equal to the preset adjustment threshold, the weight of the current data is increased; if the weight adjustment coefficient of the current data is less than the preset adjustment threshold, the weight of the current data is decreased. The adjustment algorithm is:

[0072] In the formula, ω new is the weight adjusted by the current data, ω old is the weight before current data adjustment, C z is the weight adjustment coefficient, and F is the coefficient threshold.

[0073] When any iteration satisfies the convergence regulation, the comfort score obtained in this iteration and the weight corresponding to each sensor data and physiological data are output, and the comfort of the pillow is evaluated based on the finally obtained comfort score, including the following steps:

[0074] If the difference between the weight after current data adjustment and the weight before current data adjustment is less than the preset difference threshold, the difference threshold is preset to 0.1, it is judged that convergence adjustment is satisfied;

[0075] The comfort score obtained in this iteration and the weight corresponding to each sensor data and physiological data are output, and the comfort of the pillow is evaluated based on the final comfort score obtained.

[0076] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0077] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0078] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A pillow comfort evaluation method based on multi-source data monitoring, characterized in that: The evaluation method comprises the following steps: The evaluation system collects sensor data of several pillows during the test process. The sensor data includes the cumulative coefficient of pressure fluctuations at multiple sensing points and the thermal conductivity coefficient at multiple sensing points. The system monitors the physiological data of users when testing the pillows through monitoring equipment. After analyzing the historical data based on the probability distribution model, the system obtains the impact of sensor data and physiological data on comfort, and generates corresponding weights for each sensor data and physiological data. The probability distribution model performs normal distribution on the cumulative coefficients of pressure fluctuations at multiple sensing points and the thermal conductivity coefficients at multiple sensing points of multiple pillows, performs Poisson distribution on the frequency of body position changes of multiple pillows, performs exponential distribution on the skin electrical response factors of multiple pillows, calculates the Pearson correlation coefficient between each sensor data and physiological data and the user's deep sleep duration, sums the Pearson correlation coefficients of the cumulative coefficients of pressure fluctuations at sensing points, the thermal conductivity coefficients at multiple sensing points, the frequency of body position changes, and the skin electrical response factor to obtain the total value of the Pearson correlation coefficient, and obtains the weight by dividing the Pearson correlation coefficient by the total value of the Pearson correlation coefficient. The sensor data and physiological data are combined with corresponding weights to obtain the initial comfort score; Use the MCMC algorithm to perform multiple sampling in the probability distribution of each sensor data and physiological data. After each sampling, regenerate the weight of each sensor data and physiological data according to the sampling result, recalculate the comfort score based on the newly generated weight, and use the newly acquired comfort score based on the acceptance probability criterion of the MCMC algorithm to determine whether to accept the newly generated weight; When any iteration satisfies the convergence regulation, the comfort score obtained in this iteration and the weight corresponding to each sensor data and physiological data are output, and the comfort of the pillow is evaluated based on the final comfort score.

2. The pillow comfort evaluation method based on multi-source data monitoring according to claim 1 is characterized in that: The evaluation system uses monitoring equipment to monitor the user's physiological data when testing the sleeping pillow. The physiological data include the frequency of body position changes and skin electrical response factors.

3. The pillow comfort evaluation method based on multi-source data monitoring according to claim 2 is characterized in that: After analyzing historical data based on the probability distribution model, the sensor data and the impact of physiological data on comfort are obtained, including the following steps: A certain number of pillows are selected from a batch of pillows to establish a pillow set, and the multi-sensing point pressure fluctuation accumulation coefficient, multi-sensing point heat conductivity coefficient, body position change frequency and skin electrical response factor of each pillow in the pillow set are recorded; The probability distribution of the cumulative coefficient of pressure fluctuations at multiple sensing points is N y (A avg ,A Q ), A avg is the average value of the cumulative coefficient of pressure fluctuation at multiple sensing points, A Q is the standard deviation of the cumulative coefficient of pressure fluctuation at multiple sensing points, and the probability distribution of the heat transfer coefficient at multiple sensing points is N r (B avg ,B Q ), B avg is the average value of the thermal conductivity coefficient of multiple sensing points, B Q is the standard deviation of the thermal conductivity coefficient at multiple sensing points, and the probability distribution of the frequency of body position change is P(C avg ), C avg is the mean frequency of body position change, and the probability distribution of the skin electrical response factor is E(D avg ), D avg is the mean value of the galvanic skin response factor; Calculate the Pearson correlation coefficient between each sensor data and physiological data and the user's deep sleep duration. The expression is: In the formula, r is the Pearson correlation coefficient, X i is the current data of the i-th pillow, Y i is the deep sleep duration of the user of the i-th pillow, is the current data mean, It is the average deep sleep duration of the user.

4. The pillow comfort evaluation method based on multi-source data monitoring according to claim 3 is characterized in that: The sensor data and physiological data are combined with corresponding weights to obtain the initial comfort score, including the following steps: The cumulative coefficient of pressure fluctuation at multiple sensing points, the thermal conductivity coefficient at multiple sensing points, the frequency of body position change and the galvanic skin response factor of the current pillow are obtained, and the cumulative coefficient of pressure fluctuation at multiple sensing points, the thermal conductivity coefficient at multiple sensing points, the frequency of body position change and the galvanic skin response factor are normalized so that their value ranges are mapped to [0,1]. The weights of the cumulative coefficient of pressure fluctuation at multiple sensing points, the weights of the thermal conductivity coefficient at multiple sensing points, the weights of the frequency of body position change and the weights of the galvanic skin response factor are obtained through the probability distribution model, and then the initial comfort score is calculated.

5. The pillow comfort evaluation method based on multi-source data monitoring according to claim 4 is characterized in that: The MCMC algorithm is used to perform multiple sampling in the probability distribution of each sensor data and physiological data. After each sampling, the weight of each sensor data and physiological data is regenerated according to the sampling result, including the following steps: Randomly select an initial value from the probability distribution of each data as the sampling starting point, marked as state; Using the Markov Chain Monte Carlo algorithm, starting from the current state of each parameter, multiple samplings are performed according to its probability distribution to generate a series of states; According to the objective function, whether the new sampled state should be accepted is determined. The objective function expression is: In the formula, CFT(state new ) is the comfort score calculated under the new state, CFT(state old ) is the comfort score calculated under the current state, min(*) means selecting the minimum value, J(accepts state new ) represents the probability of accepting the new state; If J(accepts state new )≥80%, judge to accept the new state of the current data; Calculate the comfort score difference of the current data and the new state, the expression is: C = CFT (state new )-CFT(state old ), where C is the comfort score difference of the new state, CFT(state new ) is the comfort score calculated under the new state, CFT(state old ) is the comfort score calculated in the current state; The weight adjustment coefficient is obtained by summing up the difference of all new state comfort scores of the current data. The expression is: In the formula, C z is the weight adjustment coefficient, m is the number of times the new state of the current data is accepted, C i is the comfort score difference of the i-th new state of the current data; If the weight adjustment coefficient of the current data is greater than or equal to the preset adjustment threshold, the weight of the current data is increased; if the weight adjustment coefficient of the current data is less than the preset adjustment threshold, the weight of the current data is decreased.

6. The pillow comfort evaluation method based on multi-source data monitoring according to claim 5 is characterized in that: If the weight adjustment coefficient of the current data is greater than or equal to the preset adjustment threshold, the weight of the current data is increased; if the weight adjustment coefficient of the current data is less than the preset adjustment threshold, the weight of the current data is decreased. The adjustment algorithm is: In the formula, ω new is the weight adjusted by the current data, ω old is the weight before current data adjustment, C z is the weight adjustment coefficient, and F is the coefficient threshold.

7. The pillow comfort evaluation method based on multi-source data monitoring according to claim 2, characterized in that: The logic for obtaining the pressure fluctuation accumulation coefficient of multiple sensing points is as follows: pressure strain gauges are set at multiple monitoring points of the pillow. During the test, constant pressure is applied to multiple monitoring points by the pressure testing equipment. After recording the actual pressure of each monitoring point, the pressure standard deviation is calculated based on the actual pressure of each monitoring point. The pressure standard deviation is integrated and accumulated over a period of time to obtain the pressure fluctuation accumulation coefficient of multiple sensing points. The expression is: Where ps is the cumulative coefficient of pressure fluctuations at multiple sensing points, yb(t) is the standard deviation of pressure at time t, and T is the monitoring duration. The logic for obtaining the multi-sensing point heat transfer coefficient is as follows: temperature sensors are set at multiple monitoring points of the pillow, and after heating one side of the pillow, the heat flux density at multiple monitoring points is calculated. The expression is: In the formula, q is the heat flux density, k is the thermal conductivity of the pillow material, ΔT is the temperature difference between the pillow surface and the heated side, and Δx is the thickness of the pillow. The heat flux density of multiple monitoring points is summed to obtain the multi-sensing point thermal conductivity coefficient.

8. The pillow comfort evaluation method based on multi-source data monitoring according to claim 2, characterized in that: The logic for obtaining the body position change frequency is as follows: 2 minutes after the user enters deep sleep, the polysomnography monitor is used to record the number of body position changes of the user within a period of time, and the body position change frequency is obtained by dividing the number of body position changes by the recorded time. The logic for obtaining the galvanic skin response factor is as follows: 2 minutes after the user enters deep sleep, the polysomnography monitor is used to detect that the user has placed electrodes on the fingers, palms, or feet, and a voltage of 0.5-5 volts is applied between the electrodes to measure the conductivity of the skin. The change in skin conductivity is continuously measured to generate a time series signal, and the skin conductivity at multiple time points is obtained. The skin conductivity at multiple time points is subtracted and the absolute value is taken as the galvanic skin response factor.

Images