Sleep monitoring control interaction system and method based on artificial intelligence and data fusion

Abstract

The invention relates to the technical field of smart home and health monitoring, and particularly discloses a sleep monitoring control interaction system and method based on artificial intelligence and data fusion, and the system comprises a sensing data acquisition module, a sleep environment recognition module, a mattress adjustment judgment module and a control strategy feedback module. A high-density sensor array network is embedded in a mattress body and is used for detecting pressure distribution data and temperature distribution data at a current sleep time point in real time; the sleep environment at the current sleep time point is further judged, the sleep environment is divided into a single-person sleep environment and a multi-person sleep environment, and the multi-person sleep environment is further divided into an adhesive multi-person sleep environment and a non-adhesive multi-person sleep environment; and finally executing dynamic adjustment strategy analysis. By monitoring and analyzing the pressure distribution in real time, the system can identify the part where a human body is in contact with the mattress and the pressure change, so that the supporting structure of the mattress is dynamically adjusted, and the sleep quality is improved.

Description

Technical Field

[0001] The present invention belongs to the technical fields of smart home and health monitoring, and relates to a sleep monitoring control interaction system and method based on artificial intelligence and data fusion. Background Art

[0002] With the increasing attention of people to sleep quality, traditional mattresses have gradually been replaced by smart mattresses. Existing smart mattresses mainly improve the user's sleep experience through simple pressure sensors or temperature regulation functions, but these technologies have the following problems: Existing smart mattresses mainly rely on overall sensor data for unified adjustment, which means that the mattress will perform a single adjustment strategy based on the overall pressure and temperature data. This method may work well in a single-person sleep environment, but it is insufficient in a multi-person sleep environment, especially in a sticky multi-person environment. Since it is impossible to independently monitor and adjust different areas, the mattress may not be able to meet the comfort needs of all users at the same time. For example, one user may need firmer support, while another user may need a softer mattress, and such needs are difficult to achieve under unified adjustment. In addition, temperature regulation also faces the same problem and cannot perform precise local adjustment according to the body temperature changes of different users; The defect of this unified adjustment not only affects the user's sleep experience, but may also lead to long-term health problems. For users with special needs, such as pregnant women, the elderly, or people with specific health problems, the inability to provide personalized support and temperature regulation may cause discomfort or even exacerbate health problems. In addition, the unified adjustment strategy lacks flexibility and cannot adapt to the changing needs of users in different seasons or different health states, which limits the wide application and acceptance of smart mattresses. Summary of the Invention

[0003] In view of the above problems existing in the prior art, the present invention provides a sleep monitoring control interaction system and method based on artificial intelligence and data fusion to solve the above technical problems.

[0004] In order to achieve the above and other purposes, the technical solutions adopted by the present invention are as follows: On the one hand, the present invention provides a sleep monitoring control interaction system based on artificial intelligence and data fusion. The system includes a perception data acquisition module, a sleep environment recognition module, a mattress adjustment judgment module, and a control strategy generation module. The above-mentioned modules are connected by wired and / or wireless connection methods to realize data transmission between the modules; Perception data acquisition module: Embed a high-density sensor array network in the mattress body to be used for real-time detection of pressure distribution data and temperature distribution data at the current sleep time point; Sleep environment recognition module: Extract features from the pressure distribution data at the current sleep time point, and then judge the sleep environment at the current sleep time point. The sleep environment is divided into a single-person sleep environment and a multi-person sleep environment, and the multi-person sleep environment is further subdivided into an adhesive multi-person sleep environment and a non-adhesive multi-person sleep environment; Mattress adjustment judgment module: Based on the sleep environment classification result output by the sleep environment recognition module, combined with the pressure distribution data and temperature distribution data at the current sleep time point provided by the perception data acquisition module, perform dynamic adjustment strategy analysis; Control strategy feedback module: Feed back the dynamic adjustment strategy of the current mattress to the mattress data adjustment terminal, and the mattress data adjustment terminal makes corresponding adjustments to the mattress.

[0005] The logic for judging the sleep environment at the current sleep time point is as follows: Based on the pressure distribution data at the current sleep time point, the pressure values of each two-dimensional coordinate in the corresponding sensor array at the current sleep time point are collected in real time, and a two-dimensional pressure matrix of the mattress corresponding to the current sleep time point is generated; and the two-dimensional pressure matrix is input into a pre-trained deep learning model to extract the coordinates of each pressure center and the gradient change rate of the mattress corresponding to the current sleep time point; And based on each detected pressure center of the mattress corresponding to the current sleep time point, each pressure coverage area is divided, denoted as where represents the pressure coverage area centered on , i is the number of each pressure center, i = 1, 2... n, represents the coordinates of the i-th pressure center; Thus, the dynamic correlation coefficient between each pressure center is calculated, and the value range is [0, 1], which is used to measure and 's dynamic correlation degree; j is the number of each other pressure center, j = 1, 2... n - 1, and i is not equal to j; respectively represent the weight factor of the pressure gradient change rate and the area overlap degree between each pressure center; Obtain the number n of pressure centers of the mattress corresponding to the current sleep time point; When it is detected that the number n of pressure centers of the mattress corresponding to the current sleep time point is 1, it is determined that the sleep environment at the current sleep time point is a single-person sleep environment; When it is detected that the number n of pressure centers of the mattress corresponding to the current sleep time point is greater than or equal to 2, if the dynamic correlation coefficients between all pressure centers are lower than the set threshold S, it is determined as a non-adhesive multi-person sleep environment; if at least one of the dynamic correlation coefficients is greater than or equal to the set threshold S, it is determined as an adhesive multi-person sleep environment.

[0006] The specific calculation formulas for the weight factors of the pressure gradient change rates between pressure centers and the regional overlap are shown below: ; ; In the above calculation formula, exp(·) is an exponential function with the natural constant e as the base. They represent the gradient change rate of the i-th pressure center and the j-th pressure center of the mattress at the current sleep time point, respectively. σ is the regularization parameter of the gradient change, which is used to smooth the gradient difference. Indicates The pressure coverage area is centered. represents the coordinates of the jth additional pressure center; express and The overlapping area, express and The combined area.

[0007] When the sleeping environment at the current sleeping time point is identified as a single person sleeping environment, the operation process of the mattress adjustment judgment module includes: Based on the pressure distribution data at the current sleep time point, the pressure value of each two-dimensional coordinate in the sensor array corresponding to the current sleep time point is extracted, and it is mapped with the position of each pressure coverage area to obtain the pressure value of each pressure coverage area of ​​the sleeping user corresponding to the current sleep time point. , further obtain the weight value W of the user sleeping on the mattress at the current sleeping time point; c is the number of each pressure coverage area, c=1,2,...n, n is the total number of pressure coverage areas; And collect the pressure area of ​​the sleeping user in each pressure coverage area corresponding to the current sleeping time point ; Thus, when the sleeping environment at the current sleeping time point is a single-person sleeping environment, the support strength compensation value corresponding to each pressure coverage area of ​​the sleeping user at the current sleeping time point is calculated. , t is the current sleep time point, T is the duration corresponding to the set standard sleep cycle; α1, α2 and α3 represent the predefined mattress material elastic coefficient, pressure sensitivity weight and dynamic adjustment factor respectively; Synchronously calculate the temperature compensation value of each pressure coverage area of ​​the sleeping user at the current sleeping time point when the sleeping environment at the current sleeping time point is a single sleeping environment , The average body temperature of the sleeping user on the mattress is obtained by weighted summing up the temperatures at each historical monitoring time point. It represents the environmental temperature at the current sleep time point, τ is the thermal inertia time constant, α4 and α5 respectively represent the thermal conductivity of the set mattress material and the weight of the temperature compensation value decaying with time; e is the natural constant.

[0008] When it is recognized that the sleep environment at the current sleep time point is a sticky multi-person sleep environment, the operation process of the mattress adjustment judgment module includes: Based on the pressure distribution data at the current sleep time point, extract the pressure values of each two-dimensional coordinate in the corresponding sensor array at the current sleep time point, and further obtain the total number M of sleep users corresponding to the current sleep time point in the sticky multi-person sleep environment; Map it to the positions of each pressure coverage area to obtain the pressure values of each pressure coverage area of the sleep users corresponding to the current sleep time point, and further obtain the weight values of each sleep user on the mattress at the current sleep time point ; d is the number of each sleep user in the sticky multi-person sleep environment, d = 1, 2,... M; Similarly, collect the compressed area of each sleep user corresponding to the current sleep time point in each pressure coverage area , and thus calculate the support strength compensation value of each pressure coverage area of the sleep users corresponding to the current sleep time point when the sleep environment at the current sleep time point is a sticky multi-person sleep environment , β1 and β2 respectively represent the sticky correction coefficient and the junction area weight coefficient, is the contact distance between the d-th sleep user and the f-th user's sleep, and f also represents the number of each sleep user in the sticky multi-person sleep environment, and d ≠ f; is the pressure gradient value of the junction area between the d-th sleep user and the f-th user's sleep; Synchronously calculate the temperature compensation value of each pressure coverage area of the sleep users corresponding to the current sleep time point when the sleep environment at the current sleep time point is a sticky multi-person sleep environment , It represents the environmental temperature at the current sleep time point, β3 and β4 respectively represent the set group coupling coefficient and the heat diffusion rate, represents the average body temperature obtained by weighted summing the temperatures of the d-th sleep user on the mattress corresponding to each historical monitoring time point; represents the second derivative of heat conduction.

[0009] The logic for obtaining the total number M of sleep users corresponding to the current sleep time point in the sticky multi-person sleep environment is as follows: Use the clustering algorithm to identify the main pressure coverage areas from the pressure values of each two-dimensional coordinate in the corresponding sensor array at the current sleep time point; Obtain the pressure value of each main pressure coverage area; Set the pressure threshold of the user; Sum the pressure values of each main pressure coverage area, and use the summation result as the denominator of a fraction, and use the user's pressure threshold as the numerator of the fraction. Perform a fractional operation on them, and round up the resulting fractional result. Thus, calculate the total number M of sleeping users corresponding to the current sleep time point in a multi-person sleeping environment with adhesions.

[0010] When it is recognized that the sleep environment at the current sleep time point is a non-adhesive multi-person sleep environment, the operation process of the mattress adjustment judgment module includes: Based on the pressure distribution data at the current sleep time point, extract the pressure values of each two-dimensional coordinate in the sensor array corresponding to the current sleep time point, and further obtain the total number N of sleeping users corresponding to the current sleep time point in a non-adhesive multi-person sleep environment; Map it to the positions of each pressure coverage area to obtain the pressure values of each pressure coverage area of the sleeping users corresponding to the current sleep time point, and further obtain the weight values of each sleeping user on the mattress at the current sleep time point ; b is the number of each sleeping user in a non-adhesive multi-person sleep environment, b = 1, 2,... N; Similarly, collect the pressure-receiving areas of each sleeping user corresponding to the current sleep time point in each pressure coverage area , and thus calculate the support strength compensation value of each pressure coverage area of each sleeping user corresponding to the current sleep time point when the sleep environment at the current sleep time point is a non-adhesive multi-person sleep environment , where γ1 and γ2 respectively represent the set independent partition coefficient and load balancing factor, respectively represent the weight value of the jth sleeping user on the mattress at the current sleep time point and the maximum load-bearing threshold of the mattress. j also represents the number of each sleeping user in a non-adhesive multi-person sleep environment, and b ≠ f; Synchronously calculate the temperature compensation value of each pressure coverage area of each sleeping user corresponding to the current sleep time point when the sleep environment at the current sleep time point is a non-adhesive multi-person sleep environment , represents the average body temperature obtained by weighted summation of the temperatures of the bth and fth sleeping users on the mattress corresponding to each historical monitoring time point. γ3 and γ4 respectively represent the set independent temperature control compensation coefficient and environmental homogenization weight.

[0011] On the other hand, the present invention provides a sleep monitoring control interaction method based on artificial intelligence and data fusion. The method includes the following steps: Step S1, embed a high-density sensor array network in the mattress body for real-time detection of the pressure distribution data and temperature distribution data at the current sleep time point; Step S2: Extract features from the pressure distribution data at the current sleep time point, and then determine the sleep environment at the current sleep time point. The sleep environment is divided into a single-person sleep environment and a multi-person sleep environment, and the multi-person sleep environment is further subdivided into an adhesive multi-person sleep environment and a non-adhesive multi-person sleep environment; Step S3: Based on the sleep environment classification result, the pressure distribution data, and the temperature distribution data at the current sleep time point, perform dynamic adjustment strategy analysis and feedback it to the mattress data adjustment terminal, and the mattress data adjustment terminal makes corresponding adjustments to the mattress.

[0012] As described above, the sleep monitoring control interaction system and method based on artificial intelligence and data fusion provided by the present invention have at least the following beneficial effects: The sleep monitoring control interaction system and method based on artificial intelligence and data fusion provided by the present invention embed a high-density sensor array network into the mattress for real-time detection of pressure and temperature distribution data, which is an important technical means to improve sleep quality and health monitoring. First, this sensor array can provide high-resolution pressure and temperature data, enabling the system to accurately identify the sleep environment, including single-person and multi-person sleep environments. For the multi-person sleep environment, it is further subdivided into adhesive multi-person and non-adhesive multi-person sleep environments, which can help analyze and adjust the support and temperature of the mattress more precisely to meet the personalized needs of different users; By real-time monitoring and analyzing the pressure distribution, the system can identify the parts of the body in contact with the mattress and the pressure changes, thereby dynamically adjusting the support structure of the mattress. This adjustment can effectively relieve the pressure on certain parts of the body, promote blood circulation, reduce the number of turns, and thus improve sleep quality. At the same time, the temperature sensor can monitor the temperature distribution on the surface of the mattress and automatically adjust the temperature of the mattress in combination with the user's body temperature change to keep it within a suitable range and avoid overheating or overcooling; In addition, in a multi-person sleep environment, the sensor system can distinguish between adhesive and non-adhesive sleep states. This recognition ability is crucial for the implementation of the dynamic adjustment strategy. In the adhesive multi-person environment, the system needs to coordinate the needs of different users, avoid interference, and provide balanced support and temperature adjustment. In the non-adhesive environment, the system can independently adjust the sleep area of each user to provide more personalized services; The intelligent mattress can not only respond to user needs more precisely and provide a higher-quality sleep experience, but also demonstrate strong necessity and potential in health monitoring and home intelligence Brief Description of the Drawings To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0013] Figure 1 It is a schematic diagram of the connection of each module of the system of the present invention.

[0014] Figure 2 It is a schematic diagram of the connection of each step of the method of the present invention. Specific embodiments

[0015] The following will combine the embodiments of the present invention. The above content is only an example and illustration of the concept of the present invention. Those skilled in the art of this technology can make various modifications or supplements to the described specific embodiments or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by this claim book, they should all belong to the protection scope of the present invention.

[0016] Embodiment 1 Please refer to Figure 1 As shown, a sleep monitoring control interaction system based on artificial intelligence and data fusion. The system includes a perception data acquisition module, a sleep environment recognition module, a mattress adjustment judgment module, and a control strategy feedback module. The above-mentioned modules are connected by wired and / or wireless connection methods to realize data transmission between modules; Perception data acquisition module: Embed a high-density sensor array network in the mattress body to be used for real-time detection of pressure distribution data and temperature distribution data at the current sleep time point; The above-mentioned pressure distribution data at the current sleep time point includes the pressure values of each two-dimensional coordinate in the sensor array; The temperature distribution data at the current sleep time point includes the temperature values of each two-dimensional coordinate in the sensor array.

[0017] Sleep environment recognition module: Extract features from the pressure distribution data at the current sleep time point, and then judge the sleep environment at the current sleep time point. The sleep environment is divided into a single-person sleep environment and a multi-person sleep environment, and the multi-person sleep environment is further divided into an adhesive multi-person sleep environment and a non-adhesive multi-person sleep environment; The logic for judging the sleep environment at the current sleep time point is: Based on the pressure distribution data at the current sleep time point, the pressure values of each two-dimensional coordinate in the corresponding sensor array at the current sleep time point are collected in real time, and a two-dimensional pressure matrix of the mattress corresponding to the current sleep time point is generated therefrom; and the two-dimensional pressure matrix is input into a pre-trained deep learning model to extract the coordinates of each pressure center and the gradient change rate of the mattress corresponding to the current sleep time point. And based on each pressure center of the mattress corresponding to the detected current sleep time point, each pressure coverage area is divided, denoted as , where represents the pressure coverage area centered on , i is the number of each pressure center, i = 1, 2... n, represents the coordinates of the i-th pressure center; Thereby calculating the dynamic correlation coefficient between each pressure center, with a value range of [0, 1], used to measure and of the dynamic correlation degree; j is the number of each other pressure center, j = 1, 2... n - 1, and i is not equal to j; respectively represent the weight factor and the area overlap degree of the pressure gradient change rate between each pressure center; Obtain the number n of pressure centers of the mattress corresponding to the current sleep time point; When it is detected that the number n of pressure centers of the mattress corresponding to the current sleep time point is 1, it is determined that the sleep environment at the current sleep time point is a single-person sleep environment; When it is detected that the number n of pressure centers of the mattress corresponding to the current sleep time point is greater than or equal to 2, if the dynamic correlation coefficients between all pressure centers are lower than the set threshold S, it is determined as a non-adhesive multi-person sleep environment; if at least one of the dynamic correlation coefficients is greater than or equal to the set threshold S, it is determined as an adhesive multi-person sleep environment.

[0018] The specific calculation formulas of the weight factor and the area overlap degree of the pressure gradient change rate between each pressure center are shown as follows: ; ; In the above calculation formulas, exp(·) is the exponential function with the natural constant e as the base, respectively represent the gradient change rate of the i-th pressure center and the gradient change rate of the j-th other pressure center of the mattress corresponding to the current sleep time point, σ is the regularization parameter of the gradient change, used to smooth the gradient difference, represents the pressure coverage area centered on , represents the coordinates of the j-th other pressure center; represents and The overlapping area, express and The combined area.

[0019] Mattress adjustment judgment module: Based on the sleep environment classification results output by the sleep environment recognition module, combined with the pressure distribution data and temperature distribution data of the current sleep time point provided by the perception data acquisition module, dynamic adjustment strategy analysis is performed; When the sleeping environment at the current sleeping time point is identified as a single person sleeping environment, the operation process of the mattress adjustment judgment module includes: Based on the pressure distribution data at the current sleep time point, the pressure value of each two-dimensional coordinate in the sensor array corresponding to the current sleep time point is extracted, and it is mapped with the position of each pressure coverage area to obtain the pressure value of each pressure coverage area of ​​the sleeping user corresponding to the current sleep time point. , further obtain the weight value W of the user sleeping on the mattress at the current sleeping time point; c is the number of each pressure coverage area, c=1,2,...n, n is the total number of pressure coverage areas; And collect the pressure area of ​​the sleeping user in each pressure coverage area corresponding to the current sleeping time point ; Thus, when the sleeping environment at the current sleeping time point is a single-person sleeping environment, the support strength compensation value corresponding to each pressure coverage area of ​​the sleeping user at the current sleeping time point is calculated. , t is the current sleep time point, T is the duration corresponding to the set standard sleep cycle; α1, α2 and α3 represent the predefined mattress material elastic coefficient, pressure sensitivity weight and dynamic adjustment factor respectively; Synchronously calculate the temperature compensation value of each pressure coverage area of ​​the sleeping user at the current sleeping time point when the sleeping environment at the current sleeping time point is a single sleeping environment , The average body temperature of the sleeping user on the mattress is obtained by weighted summing up the temperatures at each historical monitoring time point. represents the ambient temperature at the current sleeping time, τ is the thermal inertia time constant, α4 and α5 represent the thermal conductivity of the set mattress material and the weight of the temperature compensation value decaying over time, respectively; e is a natural constant.

[0020] When the sleep environment for calculating the current sleep time point is a single-person sleep environment, the first term in the formula for the support strength compensation value of each pressure coverage area of the sleep user corresponding to the current sleep time point reflects the basic influence of the user's weight and contact area on the mattress support strength. The greater the weight, the higher the pressure per unit area, and the mattress needs to provide greater support force to maintain the user's comfort and body alignment. The larger the contact area, the pressure is dispersed, and the support requirement is relatively reduced. The coefficient α1 represents the elastic characteristics of the mattress material (such as hardness and deformation ability) and is used to adjust the formula to adapt to mattresses of different materials; Ergonomic research shows that the support force of the mattress needs to match the user's weight and contact area to avoid excessive local pressure leading to blood circulation obstruction or muscle fatigue. The relevant data comes from pressure distribution experiments (such as using a pressure sensor array), which can accurately measure the contact pressure between the human body and the mattress; The second term uses real-time pressure sensor data to capture the dynamic pressure changes of the user during sleep and make compensation. For example, when the user turns over or adjusts the posture, the mattress needs to respond dynamically to redistribute the support force. The coefficient α2 adjusts the influence of the sensor sensitivity on the support strength; Sleep research shows that the human body frequently adjusts its posture during sleep (usually 20 - 30 times per night), and these movements change the pressure distribution. Real-time pressure monitoring can capture these changes and, by dynamically adjusting the support strength, avoid affecting sleep quality due to insufficient or excessive local support; The third term introduces a periodic dynamic adjustment factor, which simulates the change in support demand during the sleep cycle through a sine function. The sleep cycle T is usually set to 90 minutes, and the support demand of the human body at the current sleep time point t (such as light sleep, deep sleep, rapid eye movement period) will change. The coefficient α3 is used to adjust the amplitude of this change. Sleep science research finds that during the deep sleep period, the human muscles are completely relaxed, and the support demand for the mattress is greater, while it is relatively smaller during the light sleep period and the rapid eye movement period. By introducing periodic adjustment, it can better meet the physiological needs of the user and improve sleep comfort; This formula comprehensively considers the user's weight, contact area, real-time pressure changes, and the dynamic demands of the sleep cycle, making the calculation of the support strength more accurate and personalized. A single static support force cannot adapt to the dynamic changes of the human body during sleep, while dynamic adjustment can effectively avoid discomfort caused by excessive local pressure or insufficient support, thereby improving sleep quality.

[0021] When the sleep environment for calculating the current sleep time point is a single-person sleep environment, the first term in the formula for the temperature compensation value of each pressure coverage area of the sleep user corresponding to the current sleep time point reflects the direct influence of the temperature difference between the human body surface temperature and the ambient temperature on the temperature compensation. The human body dissipates heat during sleep, and the mattress needs to adjust its heat dissipation or heat preservation performance according to the temperature difference to maintain a comfortable temperature environment. The coefficient α4 represents the thermal conductivity of the mattress material. The ideal sleep surface temperature of the human body is usually between 31-35°C, and the fluctuation of the ambient temperature will affect this balance. Research shows that too large a temperature difference may lead to sleep interruption or discomfort (such as overheating or overcooling). By adjusting the temperature compensation in real time, the thermal comfort of the user can be better maintained; The second term introduces a time decay factor to simulate the thermal inertia effect of the mattress in temperature regulation. As the current sleep time point t progresses, the effect of the initial temperature compensation will gradually weaken and eventually tend to be stable. The thermal inertia time constant τ describes the thermal response time of the mattress material. The heat conduction and heat dissipation performance of the mattress are not instantaneous but a gradual process. Through the exponential decay model, this process can be more realistically simulated and over-adjustment or under-adjustment of temperature can be avoided; The above temperature compensation value formula combines the temperature difference between the human body and the environment and the thermal inertia effect of the mattress, and can dynamically adjust the temperature compensation value to maintain the thermal comfort of the user. The human body is extremely sensitive to temperature changes during sleep, and too high or too low a temperature will lead to a decline in sleep quality. By adjusting the temperature compensation in real time, sleep interruption caused by ambient temperature fluctuations or human body heat dissipation can be avoided.

[0022] When it is identified that the sleep environment at the current sleep time point is a sticky multi-person sleep environment, the operation process of the mattress adjustment judgment module includes: Based on the pressure distribution data at the current sleep time point, extract the pressure values of each two-dimensional coordinate in the sensor array corresponding to the current sleep time point, and further obtain the total number of sleep users M corresponding to the current sleep time point in the sticky multi-person sleep environment; Map it to the positions of each pressure coverage area to obtain the pressure values of each pressure coverage area of the sleep user corresponding to the current sleep time point, and further obtain the weight values of each sleep user on the mattress at the current sleep time point ; d is the number of each sleep user in the sticky multi-person sleep environment, d = 1, 2,... M; Similarly, collect the compressed areas of each sleep user corresponding to the current sleep time point in each pressure coverage area , and thus calculate the support strength compensation value of each pressure coverage area of the sleep user corresponding to the current sleep time point when the sleep environment at the current sleep time point is a sticky multi-person sleep environment , where β1 and β2 respectively represent the sticky correction coefficient and the junction area weight coefficient, The contact distance between the d-th sleeping user and the f-th sleeping user, where f also represents the number of each sleeping user in the multi-person adhesion sleeping environment, and d ≠ f; The pressure gradient value at the boundary region between the d-th sleeping user and the f-th sleeping user; When synchronously calculating that the sleeping environment at the current sleep time point is a multi-person adhesion sleeping environment, calculate the temperature compensation value of each pressure coverage area of the sleeping user corresponding to the current sleep time point , represents the ambient temperature at the current sleep time point, and β3 and β4 respectively represent the set group coupling coefficient and thermal diffusivity, represents the average body temperature obtained by weighted summation of the temperatures corresponding to the d-th sleeping user on the mattress at each historical monitoring time point; represents the second derivative of heat conduction.

[0023] In the above formula for calculating the support strength compensation value of each pressure coverage area of the sleeping user corresponding to the current sleep time point when the sleeping environment at the current sleep time point is a multi-person adhesion sleeping environment, the first term describes the contribution of each sleeping user to the support strength, and at the same time considers the influence of the distance between each sleeping user on the supporting force. The weight and pressure-receiving area of each sleeping user determine its pressure distribution on the mattress, and the distance factor is used to describe the superposition effect of the pressure between users. The closer the distance, the more significant the pressure superposition effect; the farther the distance, the weaker the influence. The coefficient β1 represents the elastic correction ability of the mattress material under the multi-person pressure distribution. When multiple people sleep, the contact between users will cause a significant change in the pressure distribution on the mattress. For example, when two people are very close, the pressure in a local area of the mattress will increase significantly, while the pressure in the area far from the contact point is smaller. This phenomenon can be measured by the pressure sensor array, and experimental data show that the pressure influence between users has a non-linear relationship with the distance; The second term introduces the influence of the pressure gradient in the boundary region on the support strength. The boundary region refers to the area where two users are in contact or close to each other, and the pressure change rate in this area reflects the mechanical interaction between users. The coefficient β2 is used to adjust the weight of the pressure in the boundary region on the overall support. In real life, the contact area between users is often the area where pressure is concentrated (such as when the shoulders or hips are close), and if the support strength is insufficient, it is easy to cause local discomfort or collapse. By monitoring the pressure gradient in the boundary region and making compensation, the pressure can be effectively dispersed to avoid the mattress losing stability.

[0024] When the sleep environment for calculating the current sleep time point is an overlapping multi-person sleep environment, the first term in the formula for the temperature compensation value of each pressure coverage area of the sleep users corresponding to the current sleep time point describes the influence of the weighted average of the temperatures of each sleep user in the total pressure area on the temperature compensation. The body surface temperature of each sleep user contributes differently to the total heat field according to the proportion of its pressure area. The coefficient β3 represents the thermal coupling performance of the mattress material. When multiple people sleep, the body temperature of the users will conduct to the mattress through the contact area and form a superimposed heat field inside the mattress. Research shows that the distribution of the heat field is positively correlated with the contact area and body temperature of the users. The heat field distribution can be more accurately described through weighted average; The second term introduces the second derivative of heat conduction , which is used to describe the diffusion of heat in the mattress. The second derivative reflects the uniformity of the heat field, and the coefficient β4 represents the heat diffusion performance of the mattress material; By comprehensively considering the user's body temperature, contact area ratio, environmental temperature, and heat diffusion effect, the above calculation formula can dynamically adjust the temperature compensation value of the mattress, so as to adapt to the complex heat interaction requirements of the multi-person sleep environment. In the overlapping multi-person sleep environment, the body temperatures of the users will affect each other, and the superimposed effect of the heat field may lead to local overheating or overall temperature imbalance. By introducing weighted average and heat diffusion models, more uniform temperature regulation can be achieved, avoiding discomfort caused by uneven temperature.

[0025] The logic for obtaining the total number M of sleep users corresponding to the current sleep time point in the overlapping multi-person sleep environment is as follows: Use the clustering algorithm to identify the main pressure coverage areas from the pressure values of each two-dimensional coordinate in the sensor array corresponding to the current sleep time point; Obtain the pressure value of each main pressure coverage area; Set the pressure threshold of the user; Sum the pressure values of each main pressure coverage area, use the sum result as the denominator of the fraction, use the pressure threshold of the user as the numerator of the fraction, perform fractional operation on them, and round up the fractional result obtained, thereby calculating the total number M of sleep users corresponding to the current sleep time point in the overlapping multi-person sleep environment.

[0026] When it is identified that the sleep environment at the current sleep time point is a non-overlapping multi-person sleep environment, the operation process of the mattress adjustment judgment module includes: Based on the pressure distribution data at the current sleep time point, extract the pressure values of each two-dimensional coordinate in the sensor array corresponding to the current sleep time point, and further obtain the total number N of sleep users corresponding to the current sleep time point in the non-overlapping multi-person sleep environment; Map it to the positions of each pressure coverage area to obtain the pressure values of each pressure coverage area of the sleeping user corresponding to the current sleep time point, and further obtain the weight values of each sleeping user on the mattress at the current sleep time point ; b is the number of each sleeping user in a non-adhesive multi-person sleep environment, b = 1, 2,... N; Similarly, collect the pressure-bearing areas of each sleeping user corresponding to each pressure coverage area at the current sleep time point , and calculate the support strength compensation value of each pressure coverage area of each sleeping user corresponding to the current sleep time point when the sleep environment at the current sleep time point is a non-adhesive multi-person sleep environment , where γ1 and γ2 respectively represent the set independent partition coefficient and load balancing factor, respectively represent the weight value of the j-th sleeping user on the mattress at the current sleep time point and the maximum load-bearing threshold of the mattress. j also represents the number of each sleeping user in a non-adhesive multi-person sleep environment, and b ≠ f; Synchronously calculate the temperature compensation value of each pressure coverage area of each sleeping user corresponding to the current sleep time point when the sleep environment at the current sleep time point is a non-adhesive multi-person sleep environment , represents the average body temperature obtained by weighted summation of the temperatures of the b-th and f-th sleeping users on the mattress corresponding to each historical monitoring time point. γ3 and γ4 respectively represent the set independent temperature control compensation coefficient and environmental homogenization weight.

[0027] In the formula for calculating the support strength compensation value of each pressure coverage area of each sleeping user corresponding to the current sleep time point when the sleep environment at the current sleep time point is a non-adhesive multi-person sleep environment, the first term reflects the basic influence of the weight and pressure-bearing area of the b-th user on the support strength of the mattress. The greater the weight, the higher the pressure per unit area, and the mattress needs to provide greater support force to maintain the comfort and body alignment of the user. The larger the contact area, the pressure is dispersed, and the support requirement is relatively reduced. The coefficient γ1 represents the elastic characteristics (such as hardness and deformation ability) of the mattress material, which is used to adjust the formula to adapt to mattresses of different materials; The second term adjusts the support strength of the b-th sleeping user by considering the proportion of the total weight of other sleeping users to the maximum load-bearing threshold of the mattress. As the total weight of other sleeping users increases, the overall load of the mattress increases, and the support strength of the b-th sleeping user needs to be adjusted accordingly. The coefficient γ2 is used to adjust the influence of this load balancing. In a multi-person sleep environment, the mattress needs to support the weights of multiple users simultaneously. When the total weight borne by the mattress approaches its maximum load-bearing threshold, its support performance will change. By considering the total weight of other users, the support strength can be more accurately distributed to avoid local overload or insufficient support.

[0028] When the sleep environment for calculating the current sleep time point is a non - adhered multi - person sleep environment, the first term in the calculation formula of the temperature compensation value for each pressure coverage area of each sleeping user corresponding to the current sleep time point reflects the direct impact of the temperature difference between the temperature of the b - th sleeping user and the ambient temperature on temperature compensation. The human body dissipates heat during sleep, and the mattress needs to adjust its heat dissipation or heat preservation performance according to the temperature difference to maintain a comfortable temperature environment. The coefficient γ3 represents the thermal conductivity of the mattress material; The second term adjusts the temperature compensation value of the b - th sleeping user by considering the average body surface temperature of all sleeping users. This can balance the overall temperature distribution of the mattress and avoid local overheating or over - cooling. The coefficient γ4 is used to adjust the impact of this temperature balance. The temperature compensation formula maintains the thermal comfort of each user through the dynamic adjustment of the temperature difference and the average body surface temperature.

[0029] Control strategy feedback module: Feed back the dynamic adjustment strategy of the current mattress to the mattress data adjustment terminal, and the mattress data adjustment terminal makes corresponding adjustments to the mattress.

[0030] Embodiment 2 Please refer to Figure 2 As shown, a sleep monitoring and control interaction method based on artificial intelligence and data fusion, the method includes the following steps: Step S1: Embed a high - density sensor array network in the mattress body to be used for real - time detection of the pressure distribution data and temperature distribution data at the current sleep time point; Step S2: Extract features from the pressure distribution data at the current sleep time point, and then judge the sleep environment at the current sleep time point, where the sleep environment is divided into a single - person sleep environment and a multi - person sleep environment, and the multi - person sleep environment is further divided into an adhered multi - person sleep environment and a non - adhered multi - person sleep environment; Step S3: Based on the sleep environment classification result, as well as the pressure distribution data and temperature distribution data at the current sleep time point, perform dynamic adjustment strategy analysis, and feed it back to the mattress data adjustment terminal, and the mattress data adjustment terminal makes corresponding adjustments to the mattress.

[0031] It should be understood that in various embodiments of the present application, the magnitudes of the serial numbers of the above - mentioned processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0032] It should be understood that determining B according to A does not mean determining B only according to A, and B can also be determined according to A and / or other information.

[0033] Some of the data in the above formula are calculated by removing the dimension and taking their numerical values. The formula is the one that is closest to the actual situation obtained through software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0034] Finally, the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A sleep monitoring control interaction system based on artificial intelligence and data fusion, characterized in that It includes the following modules: Perceived data acquisition module: Embed a high-density sensor array network in the mattress body to detect the pressure distribution data and temperature distribution data at the current sleep time point in real time; Sleep environment recognition module: Extract features from the pressure distribution data at the current sleep time point, and then judge the sleep environment at the current sleep time point. The sleep environment is divided into a single-person sleep environment and a multi-person sleep environment, and the multi-person sleep environment is further divided into an adhesive multi-person sleep environment and a non-adhesive multi-person sleep environment; Mattress adjustment judgment module: Based on the sleep environment classification result output by the sleep environment recognition module, combined with the pressure distribution data and temperature distribution data at the current sleep time point provided by the perceived data acquisition module, perform dynamic adjustment strategy analysis; Control strategy feedback module: Feed back the dynamic adjustment strategy of the current mattress to the mattress data adjustment terminal, and the mattress data adjustment terminal makes corresponding adjustments to the mattress.

2. The sleep monitoring control interaction system based on artificial intelligence and data fusion according to claim 1, wherein The logic for judging the sleep environment at the current sleep time point is as follows: Based on the pressure distribution data at the current sleep time point, the pressure values of each two-dimensional coordinate in the corresponding sensor array at the current sleep time point are collected in real time, and a two-dimensional pressure matrix of the mattress corresponding to the current sleep time point is generated; and the two-dimensional pressure matrix is input into a pre-trained deep learning model to extract the coordinates and gradient change rates of the pressure centers of the mattress corresponding to the current sleep time point; And based on the pressure centers of the mattress corresponding to the detected current sleep time points, each pressure coverage area is divided and denoted as , where represents the pressure coverage area centered on , i is the number of each pressure center, i = 1, 2... n, represents the coordinates of the i-th pressure center; Calculate the dynamic correlation coefficient between each pressure center accordingly , with a value range of [0, 1], used to measure and 's degree of dynamic correlation; j is the number of each other pressure center, j = 1, 2... n - 1, and i is not equal to j; respectively represent the weight factor of the pressure gradient change rate between each pressure center and the regional overlap degree; Obtain the number n of pressure centers of the mattress corresponding to the current sleep time point; When it is detected that the number n of pressure centers of the mattress corresponding to the current sleep time point is n = 1, it is determined that the sleep environment at the current sleep time point is a single-person sleep environment; When it is detected that the number n of pressure centers of the mattress corresponding to the current sleep time point is greater than or equal to 2, if the dynamic correlation coefficients between all pressure centers are lower than the set threshold S, it is determined as a non-adhesive multi-person sleep environment; if at least one of the dynamic correlation coefficients is greater than or equal to the set threshold S, it is determined as an adhesive multi-person sleep environment.

3. The sleep monitoring control interaction system based on artificial intelligence and data fusion according to claim 2, wherein The specific calculation formulas for the weight factor of the pressure gradient change rate and the regional overlap degree between each pressure center are shown as follows: ； ； In the above calculation formula, exp(·) is the exponential function with the natural constant e as the base. respectively represent the gradient change rate of the i-th pressure center and the gradient change rate of the j-th other pressure center of the mattress corresponding to the current sleep time point. σ is the regularization parameter of the gradient change, which is used to smooth the gradient difference. It means is the pressure coverage area centered on represents the coordinates of the j-th other pressure center; It means and is the overlapping area of It means and is the combined area of 4. The sleep monitoring control interaction system based on artificial intelligence and data fusion according to claim 1, characterized in that, When the sleep environment at the current sleep time point is identified as a single-person sleep environment, the operation process of the mattress adjustment judgment module includes: Based on the pressure distribution data at the current sleep time point, extract the pressure values of each two-dimensional coordinate in the corresponding sensor array at the current sleep time point, map them to the positions of each pressure coverage area, and obtain the pressure values of each pressure coverage area of the sleeping user corresponding to the current sleep time point , and further obtain the weight value W of the sleeping user on the mattress at the current sleep time point; c is the number of each pressure coverage area, c = 1, 2,... n, and n is the total number of pressure coverage areas; And collect the pressure-bearing areas of the sleeping user corresponding to the current sleep time point in each pressure coverage area ; When calculating that the sleep environment at the current sleep time point is a single-person sleep environment, the support intensity compensation value of each pressure coverage area of the sleep user corresponding to the current sleep time point , where \(t\) is the current sleep time point and \(T\) is the duration corresponding to the set standard sleep cycle; \(\alpha_1\), \(\alpha_2\), and \(\alpha_3\) respectively represent the predefined elastic coefficient of the mattress material, the pressure sensitivity weight, and the dynamic adjustment factor; When the sleep environment at the current sleep time point is calculated synchronously as a single-person sleep environment, the temperature compensation values of each pressure coverage area of the sleep user corresponding to the current sleep time point , is the average body temperature obtained by weighted summation of the temperatures of the sleep user corresponding to each historical monitoring time point on the mattress, represents the ambient temperature at the current sleep time point, τ is the thermal inertia time constant, α4 and α5 respectively represent the thermal conductivity of the set mattress material and the weight of the decay of the temperature compensation value over time; e is the natural constant.

5. The sleep monitoring control interaction system based on artificial intelligence and data fusion according to claim 1, wherein When the sleep environment at the current sleep time point is identified as an adhesive multi-person sleep environment, the operation process of the mattress adjustment judgment module includes: Based on the pressure distribution data at the current sleep time point, extract the pressure values of each two-dimensional coordinate in the corresponding sensor array at the current sleep time point, and further obtain the total number M of sleep users corresponding to the current sleep time point in the adhesive multi-person sleep environment; Map it to the positions of each pressure coverage area to obtain the pressure values of each pressure coverage area of the sleeping user corresponding to the current sleep time point, and further obtain the weight values of each sleeping user on the mattress at the current sleep time point ; d is the number of each sleeping user in the multi-person sleeping environment with adhesion, d = 1, 2,... M; Similarly, collect the pressure-bearing areas of each sleeping user corresponding to the current sleep time point in each pressure coverage area , and calculate the support strength compensation value of each pressure coverage area of the sleeping user corresponding to the current sleep time point when the sleep environment at the current sleep time point is an adhesive multi-person sleep environment , where β1 and β2 respectively represent the adhesion correction coefficient and the boundary area weight coefficient is the contact distance between the d-th sleeping user and the f-th sleeping user, and f also represents the number of each sleeping user in the adhesive multi-person sleep environment, and d≠f; is the pressure gradient value of the boundary area between the d-th sleeping user and the f-th sleeping user When synchronously calculating that the sleep environment at the current sleep time point is an adhesive multi-person sleep environment, the temperature compensation values of each pressure coverage area of the sleep users corresponding to the current sleep time point , represents the ambient temperature at the current sleep time point, and β3 and β4 respectively represent the set group coupling coefficient and heat diffusion rate, represents the average body temperature obtained by weighted summation of the temperatures of the d-th sleep user on the mattress corresponding to each historical monitoring time point; represents the second derivative of heat conduction.

6. The sleep monitoring control interaction system based on artificial intelligence and data fusion according to claim 5, characterized in that, The logic for obtaining the total number M of sleep users corresponding to the current sleep time point in the adhesive multi-person sleep environment is as follows: Use a clustering algorithm to identify the main pressure coverage areas from the pressure values of each two-dimensional coordinate in the corresponding sensor array at the current sleep time point; Obtain the pressure values of each main pressure coverage area; Set the pressure threshold of the user; Sum up the pressure values of each main pressure coverage area, and use the summation result as the denominator of a fraction. Use the user's pressure threshold as the numerator of the fraction, perform fractional arithmetic on it, and round up the resulting fractional result. Thus, calculate the total number M of sleeping users corresponding to the current sleep time point in the sticky multi-person sleep environment.

7. The sleep monitoring control interaction system based on artificial intelligence and data fusion according to claim 1, characterized in that, When it is recognized that the sleep environment at the current sleep time point is a non-sticky multi-person sleep environment, the operation process of the mattress adjustment judgment module includes: Based on the pressure distribution data at the current sleep time point, extract the pressure values of each two-dimensional coordinate in the corresponding sensor array at the current sleep time point, and further obtain the total number N of sleeping users corresponding to the current sleep time point in the non-sticky multi-person sleep environment; Map it to the positions of each pressure coverage area to obtain the pressure values of each pressure coverage area of the sleeping user corresponding to the current sleep time point, and further obtain the weight values of each sleeping user on the mattress at the current sleep time point ; b is the number of each sleeping user in the non-adhesive multi-person sleep environment, b = 1, 2,... N; Similarly, collect the pressure-bearing area of each sleeping user corresponding to the current sleep time point in each pressure coverage area , and calculate the support strength compensation value of each pressure coverage area of each sleeping user corresponding to the current sleep time point when the sleep environment at the current sleep time point is a non-adhesive multi-person sleep environment , where γ1 and γ2 respectively represent the set independent partition coefficient and load balancing factor respectively represent the weight value of the j-th sleeping user on the mattress at the current sleep time point and the maximum load-bearing threshold of the mattress. j also represents the number of each sleeping user in the non-adhesive multi-person sleep environment, and b≠f When the sleep environment at the current sleep time point is calculated synchronously as a non - sticky multi - person sleep environment, the temperature compensation values for each pressure coverage area of each sleep user corresponding to the current sleep time point , represents the average body temperature obtained by weighted summation of the temperatures of the b - th and f - th sleep users on the mattress corresponding to each historical monitoring time point. γ3 and γ4 respectively represent the set independent temperature control compensation coefficient and the environmental homogenization weight.

8. A sleep monitoring control interaction method based on artificial intelligence and data fusion, characterized in that It is implemented based on the sleep monitoring control interaction system based on artificial intelligence and data fusion described in any one of claims 1-7, and includes the following steps: Step S1: Embed a high-density sensor array network in the mattress body to be used for real-time detection of the pressure distribution data and temperature distribution data at the current sleep time point; Step S2: Extract features from the pressure distribution data at the current sleep time point, and then judge the sleep environment at the current sleep time point. The sleep environment is divided into a single-person sleep environment and a multi-person sleep environment, and the multi-person sleep environment is further divided into a sticky multi-person sleep environment and a non-sticky multi-person sleep environment; Step S3: Based on the sleep environment classification result and the pressure distribution data and temperature distribution data at the current sleep time point, perform dynamic adjustment strategy analysis and feedback it to the mattress data adjustment terminal, and the mattress data adjustment terminal makes corresponding adjustments to the mattress.

Images