Multi-area pressure monitoring and feedback method for sleep monitoring mattress

Abstract

The invention discloses a multi-region pressure monitoring and feedback method for a sleep monitoring mattress, which comprises the following steps: step 1, arranging a monitoring feedback system to collect pressure data and biological characteristic signals, step 2, dynamically planning a human body contact region during sleep, step 3, identifying a sleep stage and an abnormal sleep event, and step 4, determining the sleep stage and the abnormal sleep event. 4, dynamic feedback is output based on cooperation of multiple actuators; by means of dynamic area reconstruction and high-density sensor layout, millisecond-level pressure changes of key parts of a human body can be captured, the detection precision is greatly improved, active intervention of user health is achieved through multi-actuator collaborative dynamic adjustment, meanwhile, the method is combined with user body shape parameters, sleep monitoring of the whole crowd can be covered, and the method is suitable for being popularized and applied. The applicability of the mattress is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of smart mattress technology, and in particular to a method for multi-zone pressure monitoring and feedback of a sleep monitoring mattress. Background Technology

[0002] With increasing awareness of sleep health, smart mattresses, as the core device for home sleep monitoring, are experiencing rapid market demand growth. In existing technologies, smart mattresses mainly collect pressure distribution data between the human body and the mattress through pressure sensor arrays to monitor basic states such as sleeping posture and turning frequency.

[0003] However, such monitoring methods mostly remain at the data display level, lacking proactive intervention capabilities and unable to dynamically adjust mattress firmness based on the user's real-time sleep stage, posture changes, or individual physiological characteristics. This results in delayed and limited monitoring feedback. Furthermore, users of different body types have different pressure distribution characteristics, and current monitoring mattress sensors and systems cannot cover the monitoring differences of the entire population, leading to limited detection accuracy. Moreover, existing monitoring mattress sensors mostly use a uniform distribution method, but the pressure distribution in different parts of the human body also varies. Uniformly distributed sensors cannot capture local pressure changes in real time, resulting in significant errors in monitoring feedback. Therefore, this invention proposes a multi-zone pressure monitoring and feedback method for sleep monitoring mattresses to solve the problems existing in the prior art. Summary of the Invention

[0004] To address the aforementioned issues, the present invention aims to propose a multi-zone pressure monitoring and feedback method for a sleep monitoring mattress. This method achieves precise monitoring of sleep status and proactive health intervention through refined zone division, multimodal data fusion, and personalized feedback strategies, thereby improving the user's sleep quality.

[0005] To achieve the objectives of this invention, the invention is implemented through the following technical solution: a method for multi-zone pressure monitoring and feedback of a sleep monitoring mattress, comprising the following steps:

[0006] Step 1: Deploy a monitoring and feedback system on the smart mattress, including a multi-density distribution array of pressure sensors, a photoplethysmography pulse sensor, a triaxial accelerometer, and a microphone array, to collect pressure data and biometric signals, including heart rate, blood oxygen saturation, breathing sounds and frequencies, and body movement amplitude and frequency, in the contact area between the human body and the mattress.

[0007] Step 2: Dynamically plan the human contact areas, including the head, shoulders, chest, waist, hips, and legs, during sleep based on human anatomy models and clustering algorithms;

[0008] Step 3: Integrate stress data and biometric signals to identify sleep stages and abnormal sleep events using machine learning models;

[0009] Step 4: Based on the user's personalized profile and the identification results of sleep stages and abnormal events, dynamic feedback is output through multi-actuator collaboration.

[0010] Further improvements are made in the following aspects: Step two, which involves dynamically planning the contact area, specifically includes initialization calibration and real-time dynamic correction. Upon first use, the user inputs body shape parameters such as height, weight, and waist-to-hip ratio. Combined with the baseline pressure distribution under unloaded mattress conditions, an initial area division model is generated, defining the basic coordinate range of each contact area. During use, after the user lies down, real-time pressure data is collected through a pressure sensor array. A clustering algorithm is used to identify the effective contact area, and the coordinate range of each contact area is adjusted in conjunction with a standard human anatomical contour database. The anatomical contour database is constructed based on the Chinese adult human body size standard (GB / T 10000-1988).

[0011] A further improvement is made in that: the sleep stages and abnormal sleep events in step three specifically include:

[0012] S1. Feature extraction: Extract time series features of pressure peak from pressure data, extract spectral features of heart rate variability, frequency and amplitude features of respiratory signals, and frequency and intensity features of body movements from biosignal signals.

[0013] S2. Sleep stage classification: The extracted features are input into the long short-term memory network model, and the sleep stage classification results are output, including the sleep onset period, light sleep period, deep sleep period and REM sleep period.

[0014] The network model was trained based on sleep data from 5000+ users of different body types, with a classification accuracy of ≥92%.

[0015] S3. Abnormal event detection: When the sustained high pressure value in a certain area of ​​the pressure distribution exceeds twice the standard deviation of the user's historical average, and is accompanied by abnormal respiratory rate or heart rate, it is determined to be a sleep apnea or cardiovascular abnormal risk event.

[0016] The abnormal respiratory rate threshold is defined as >25 breaths / min or <10 breaths / min, the abnormal heart rate threshold is defined as >100 beats / min or <40 beats / min, and the sustained hypertension threshold is defined as >5 min.

[0017] A further improvement is made in that the dynamic feedback strategy in step four includes:

[0018] Postural correction feedback: When the system detects that the user is in a supine position and snoring, it controls the built-in air pumps in the lumbar and shoulder areas to inflate the area, increasing the firmness of the area by 20%-30%, prompting the user to turn to a side-lying position; when the system detects that the user is in a prone position, it reduces the firmness of the chest area and deflates the air pumps in the edge areas to raise the head area and reduce cervical spine pressure; for the lumbar and shoulder areas, the firmness is increased by 5%-10% each time, gradually adjusted in 2-3 times;

[0019] Sleep stage adaptation feedback: During deep sleep, reduce the overall mattress firmness and adjust the temperature of each area to 22-24℃; during light sleep, adjust the temperature of each area to 24-25℃. If frequent turning over is detected, locally increase the firmness of the shoulder and hip areas to stabilize the position; during REM sleep, maintain the temperature of each area at 23-24℃.

[0020] Health risk warning feedback: When persistent high pressure is detected in the lumbar region and accompanied by abnormal breathing, a warning message will be pushed through the mattress's accompanying APP, and the user will be advised to adjust their sleeping position or contact a medical institution.

[0021] Further improvements are made in the following aspects: In the posture correction feedback, snoring is determined by the spectrum analysis of respiratory sound signals, and the prone position is determined by the abnormal increase in the peak pressure in the chest area and the forward shift of the pressure distribution in the shoulder area; in the sleep stage adaptation feedback, a heart rate of 45-60 beats / min and a body movement frequency of <0.5 times / 10 minutes indicate deep sleep, while a heart rate of 50-65 beats / min and a body movement frequency of 0.5-2 times / 10 minutes indicate light sleep.

[0022] Further improvements include: the monitoring and feedback system comprises a pressure sensing module, a biometric acquisition module, a data processing and decision-making module, a feedback execution module, and an adaptive learning module. The pressure sensing module acquires real-time pressure data of the mattress contact area based on a multi-regional pressure sensor array. The biometric acquisition module acquires biometric signals based on a photoplethysmography (PPG) sensor, a triaxial accelerometer, and a microphone array integrated within the mattress. The data processing and decision-making module processes monitoring data and generates feedback decisions based on an edge computing chip with a built-in adaptive region segmentation algorithm, a multimodal sleep state recognition model, and a personalized feedback strategy library. The feedback execution module executes feedback strategies based on the mattress's built-in zoned air pump, electro-variable pressure pad, and temperature control module. The adaptive learning module is used for data acquisition and recording, optimizing model parameters, and updating personalized profiles. The adaptive learning module also includes a built-in user feedback learning unit that optimizes feedback parameters based on the "feedback valid / invalid" information marked by the user via the APP.

[0023] A further improvement is that the multi-regional pressure sensor array is composed of flexible thin-film pressure sensors, with a sensor density of ≥25 sensors / dm² on the human torso and buttocks. 2Head and leg sensor density ≥9 units / dm 2 .

[0024] Further improvements include: the photoplethysmography sensor is used to acquire heart rate and blood oxygen saturation data and fluctuations, the triaxial accelerometer is used to acquire body movement frequency and amplitude data, and the microphone array is used to acquire respiratory sound signals.

[0025] Further improvements include: the zoned air pumps are distributed in the torso, hip, head, and leg areas of the mattress; the electro-variable mattress pad is embedded in the mattress surface and adjusts the local support stiffness through voltage; the temperature control module is distributed in each contact area of ​​the human body and adjusts the local temperature through Pellets, with a temperature adjustment range of 20-30℃.

[0026] Further improvements include: the monitoring and feedback system also includes a communication module, which connects to the accompanying APP via Wi-Fi / Bluetooth, for setting sleep preferences and receiving early warning information. The communication module has a built-in data sharing unit that supports synchronizing sleep data with the medical and health APP.

[0027] The beneficial effects of this invention are as follows: Through dynamic region reconstruction and high-density sensor layout, this invention can capture millisecond-level pressure changes in key parts of the human body, greatly improving detection accuracy. Through multi-actuator collaborative dynamic adjustment, it can achieve proactive intervention in user health. At the same time, this method combines user body shape parameters, which can cover sleep monitoring of the entire population and effectively improve the applicability of mattresses. Attached Figure Description

[0028] Figure 1 This is a flowchart of the monitoring and feedback method of the present invention.

[0029] Figure 2 This is a diagram of the monitoring and feedback system architecture of the present invention. Detailed Implementation

[0030] To enhance understanding of the present invention, the present invention will be further described in detail below with reference to embodiments. These embodiments are only used to explain the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0031] according to Figure 1 and Figure 2 As shown, this embodiment provides a multi-zone pressure monitoring and feedback method for a sleep monitoring mattress, including the following steps:

[0032] Step 1: Deploy a monitoring and feedback system on the smart mattress, including a multi-density distribution array of pressure sensors, a photoplethysmography pulse sensor, a triaxial accelerometer, and a microphone array, to collect pressure data and biometric signals, including heart rate, blood oxygen saturation, breathing sounds and frequencies, and body movement amplitude and frequency, in the contact area between the human body and the mattress.

[0033] Step 2: Dynamically plan the human contact areas, including the head, shoulders, chest, waist, hips, and legs, during sleep based on human anatomy models and clustering algorithms;

[0034] The dynamic planning of the contact area specifically includes initial calibration and real-time dynamic correction:

[0035] Initial calibration: When using it for the first time, the user inputs body shape parameters such as height, weight and waist-to-hip ratio. Combined with the baseline pressure distribution under the unloaded state of the mattress, an initial area division model is generated, and the basic coordinate range of each contact area is defined.

[0036] Real-time dynamic correction: When the user lies down, real-time pressure data is collected through a pressure sensor array. Clustering algorithm is used to identify the effective contact area, eliminate edge interference that is not human contact, and adjust the coordinate range of each contact area in combination with a standard human anatomical contour database. The anatomical contour database is constructed based on the Chinese adult human body size standard (GB / T 10000-1988).

[0037] Step 3: Integrate stress data and biometric signals to identify sleep stages and abnormal sleep events using machine learning models;

[0038] Sleep stages and abnormal sleep events specifically include:

[0039] S1. Feature extraction: Extract time-series features of pressure peaks from pressure data, such as peak intervals and amplitude fluctuations; extract spectral features of heart rate variability, frequency and amplitude features of respiratory signals, and frequency and intensity features of body movements from biosignal signals.

[0040] S2. Sleep stage classification: The extracted features are input into the Long Short-Term Memory (LSTM) network model, and the sleep stage classification results are output, including the sleep onset period, light sleep period, deep sleep period and rapid eye movement (REM) period.

[0041] The network model was trained based on sleep data from 5000+ users of different body types, with a classification accuracy of ≥92%.

[0042] S3. Abnormal event detection: When the sustained high pressure value in a certain area of ​​the pressure distribution exceeds twice the standard deviation of the user's historical average, and is accompanied by abnormal respiratory rate or heart rate, it is determined to be a sleep apnea or cardiovascular abnormal risk event.

[0043] The abnormal respiratory rate threshold is defined as >25 breaths / min or <10 breaths / min, the abnormal heart rate threshold is defined as >100 beats / min or <40 beats / min, and the sustained high blood pressure threshold is defined as >5 min, to prevent misjudging short-term pressure peaks as abnormal.

[0044] Step 4: Based on the user's personalized profile and the identification results of sleep stages and abnormal events, dynamic feedback is output through multi-actuator collaboration; the dynamic feedback strategy includes:

[0045] Postural correction feedback: When the system detects that the user is in a supine position and snoring, it controls the in-situ air pumps in the lumbar and shoulder areas to inflate these areas, increasing their firmness by 20%-30%, prompting the user to turn to a side-lying position. When the system detects that the user is in a prone position, it reduces the firmness of the chest area and deflates the air pumps in the edge areas to raise the head area and reduce cervical spine pressure. For the lumbar and shoulder areas, the firmness is increased by 5%-10% each time, gradually adjusted in 2-3 times. Snoring is determined by spectral analysis of respiratory sound signals, and the prone position is determined by analysis of abnormally high pressure peaks in the chest area and forward shift of pressure distribution in the shoulder area.

[0046] Sleep stage adaptation feedback: Deep sleep is characterized by a heart rate of 45-60 beats / min and a body movement frequency of <0.5 times / 10 minutes. The LSTM model identifies delta waves as dominant, reducing overall mattress firmness by 15%-20% and adjusting the temperature of each area to 22-24℃. Light sleep is characterized by a heart rate of 50-65 beats / min and a body movement frequency of 0.5-2 times / 10 minutes. The temperature of each area is adjusted to 24-25℃. The LSTM model identifies theta waves as dominant. If frequent turning over is detected (>2 times / 10 minutes), the firmness of the shoulder and hip areas is locally increased to stabilize position, increasing by 10%-15%. During REM sleep, the temperature of each area is maintained at 23-24℃.

[0047] Health risk warning feedback: When persistent high pressure is detected in the lumbar region and accompanied by abnormal breathing, a warning message will be pushed through the mattress's accompanying APP, and the user will be advised to adjust their sleeping position or contact a medical institution.

[0048] The monitoring and feedback system includes a pressure sensing module, a biometric acquisition module, a data processing and decision-making module, a feedback execution module, and an adaptive learning module. The pressure sensing module acquires real-time pressure data of the mattress contact area based on a multi-regional pressure sensor array. The biometric acquisition module acquires biometric signals based on a photoplethysmography (PPG) sensor, a triaxial accelerometer, and a microphone array integrated inside the mattress. The data processing and decision-making module processes monitoring data and generates feedback decisions based on an edge computing chip with a built-in adaptive region partitioning algorithm (based on DBSCAN clustering), a multimodal sleep state recognition model (based on an LSTM network), and a personalized feedback strategy library. The personalized feedback strategy library stores feedback parameters for users with different body types and medical histories. The feedback execution module executes feedback strategies based on the mattress's built-in zoned air pump, electro-variable pressure pad, and temperature control module. The adaptive learning module is used for data acquisition and recording, optimizing model parameters, and updating personalized profiles. The adaptive learning module has a built-in user feedback learning unit that optimizes feedback parameters based on the "valid / invalid" information marked by the user through the APP. If the feedback is invalid, the firmness adjustment range is reduced.

[0049] The multi-region pressure sensor array consists of flexible thin-film pressure sensors, preferably PVDF piezoelectric thin-film or capacitive pressure sensors, with a sensor density of ≥25 sensors / dm² on the human torso (shoulder, chest, and waist) and hips. 2 Head and leg sensor density ≥9 units / dm 2 It also supports dynamic networking to adapt to changes in the contact area of ​​users of different body sizes.

[0050] The photoplethysmography (PPG) sensor is used to acquire heart rate and blood oxygen saturation data and fluctuations, the triaxial accelerometer is used to acquire body movement frequency and amplitude data, and the microphone array is used to acquire respiratory sound signals.

[0051] The zoned air pumps are distributed in the torso, hip, head, and leg areas of the mattress. They are connected to the air bladder assembly via air tubes to control local air pressure to adjust firmness. Each zone corresponds to one air pump. The zoned air pumps are miniature brushless DC air pumps with a maximum output pressure of 0.5MPa per pump, a response time of ≤1 second, and support independent control.

[0052] Electro-variable mattress pads are embedded in the mattress surface and adjust local support stiffness through voltage. The response time is ≤0.5 seconds and the hardness variation range is 0.1-50HA.

[0053] Temperature control modules are distributed in each area that comes into contact with the human body, and local temperature is adjusted via Peltier, with a temperature adjustment range of 20-30℃.

[0054] The adaptive learning module includes user behavior data collection, model parameter optimization, and personalized profile updates. User behavior data collection records users' weekly sleep duration, preferred sleeping positions (e.g., percentage of left lateral decubitus and supine decubitus), and feedback acceptance (e.g., whether a feedback strategy is actively turned off). Model parameter optimization targets users who consistently sleep on their side (e.g., left lateral decubitus >80%), reducing the sensor spacing between the corresponding shoulder and hip areas (from 2cm to 1.5cm) to improve local pressure sampling accuracy. For users who frequently trigger feedback (e.g., adjusting firmness >5 times per week), the module learns their "comfortable firmness threshold" and updates the feedback strategy library (e.g., reducing the lumbar region firmness adjustment step size from 10% to 5%). Personalized profile updates are performed every 30 days based on the latest collected data, updating user body parameters (e.g., waist-to-hip ratio changes) and health risk models (e.g., sleep apnea threshold adjustments).

[0055] The monitoring and feedback system also includes a communication module that connects to a companion app via Wi-Fi / Bluetooth. This module is used to set sleep preferences and receive alerts, such as target deep sleep duration and sleep firmness preferences. The communication module also has a built-in data sharing unit that supports synchronizing sleep data with a medical and health app.

[0056] Example

[0057] This embodiment focuses on sleep apnea intervention in obese users.

[0058] User BMI=32, height 175cm, weight 95kg, first time using this mattress:

[0059] Initial calibration: After the user inputs body shape parameters, the system generates an initial region division model. The horizontal range of the waist area is 25-35cm, and the vertical range is 10-20cm.

[0060] Dynamic Region Reconstruction: After the user lies down, the pressure sensor array detects that the contact area in the lumbar region is 15% larger than that of the standard model. The system expands the lateral range of the lumbar region to 30-40cm and the longitudinal range to 12-22cm through DBSCAN clustering. Sleep State Analysis: At 2:15 AM, the system detects that the user is in a supine position, with a peak pressure of 4.2kPa in the lumbar region (historical average 1.8kPa). At the same time, the respiratory sound signal spectrum shows a low-frequency component of 12Hz, which is characteristic of snoring. The LSTM model determines that the user is at "high risk of supine sleep apnea".

[0061] Feedback execution: The system controls the air pumps in the lumbar and shoulder areas to inflate, increasing the hardness of the area from "medium" to "firm" within 30 seconds. Specifically, the hardness count value increases from 30HA to 36HA, with each adjustment being 6HA, and every 15 seconds. The user's position changes to lying on their right side, the snoring disappears, and the peak pressure in the lumbar area drops to 2.5 kPa, which is within the normal range.

[0062] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for multi-zone pressure monitoring and feedback of a sleep monitoring mattress, characterized in that, Includes the following steps: Step 1: Deploy a monitoring and feedback system on the smart mattress, including a multi-density distribution array of pressure sensors, a photoplethysmography pulse sensor, a triaxial accelerometer, and a microphone array, to collect pressure data and biometric signals, including heart rate, blood oxygen saturation, breathing sounds and frequencies, and body movement amplitude and frequency, in the contact area between the human body and the mattress. Step 2: Dynamically plan the human contact areas, including the head, shoulders, chest, waist, hips, and legs, during sleep based on human anatomy models and clustering algorithms; Step 3: Integrate stress data and biometric signals to identify sleep stages and abnormal sleep events using machine learning models; Step 4: Based on the user's personalized profile and the identification results of sleep stages and abnormal events, dynamic feedback is output through multi-actuator collaboration.

2. The method for multi-zone pressure monitoring and feedback of a sleep monitoring mattress according to claim 1, characterized in that: Step two, which involves dynamically planning the contact area, specifically includes initial calibration and real-time dynamic correction. Upon first use, the user inputs body shape parameters such as height, weight, and waist-to-hip ratio. Combined with the baseline pressure distribution under unloaded mattress conditions, an initial area division model is generated, defining the basic coordinate range of each contact area. During use, after the user lies down, real-time pressure data is collected through a pressure sensor array. A clustering algorithm is used to identify effective contact areas, and the coordinate range of each contact area is adjusted in conjunction with a standard human anatomical contour database. The anatomical contour database is constructed based on the Chinese adult human body size standard (GB / T 10000-1988).

3. The method for multi-zone pressure monitoring and feedback of a sleep monitoring mattress according to claim 1, characterized in that: The sleep stages and abnormal sleep events in step three specifically include: S1. Feature extraction: Extract time series features of pressure peak from pressure data, extract spectral features of heart rate variability, frequency and amplitude features of respiratory signals, and frequency and intensity features of body movements from biosignal signals. S2. Sleep stage classification: The extracted features are input into the long short-term memory network model, and the sleep stage classification results are output, including the sleep onset period, light sleep period, deep sleep period and REM sleep period. The network model was trained based on sleep data from 5000+ users of different body types, with a classification accuracy of ≥92%. S3. Abnormal event detection: When the sustained high pressure value in a certain area of ​​the pressure distribution exceeds twice the standard deviation of the user's historical average, and is accompanied by abnormal respiratory rate or heart rate, it is determined to be a sleep apnea or cardiovascular abnormal risk event. The abnormal respiratory rate threshold is defined as >25 breaths / min or <10 breaths / min, the abnormal heart rate threshold is defined as >100 beats / min or <40 beats / min, and the sustained hypertension threshold is defined as >5 min.

4. The method for multi-zone pressure monitoring and feedback of a sleep monitoring mattress according to claim 1, characterized in that: The dynamic feedback strategy in step four includes: Postural correction feedback: When the system detects that the user is in a supine position and snoring, it controls the built-in air pumps in the lumbar and shoulder areas to inflate the area, increasing the firmness of the area by 20%-30%, prompting the user to turn to a side-lying position; when the system detects that the user is in a prone position, it reduces the firmness of the chest area and deflates the air pumps in the edge areas to raise the head area and reduce cervical spine pressure; for the lumbar and shoulder areas, the firmness is increased by 5%-10% each time, gradually adjusted in 2-3 times; Sleep stage adaptation feedback: During deep sleep, reduce the overall mattress firmness and adjust the temperature of each area to 22-24℃; during light sleep, adjust the temperature of each area to 24-25℃. If frequent turning over is detected, locally increase the firmness of the shoulder and hip areas to stabilize the position; during REM sleep, maintain the temperature of each area at 23-24℃. Health risk warning feedback: When persistent high pressure is detected in the lumbar region and accompanied by abnormal breathing, a warning message will be pushed through the mattress's accompanying APP, and the user will be advised to adjust their sleeping position or contact a medical institution.

5. The method for multi-zone pressure monitoring and feedback of a sleep monitoring mattress according to claim 4, characterized in that: In the postural correction feedback, snoring is determined by the spectral analysis of respiratory sound signals, and the prone position is determined by the abnormal increase in the peak pressure in the chest area and the forward shift of the pressure distribution in the shoulder area. In the sleep stage adaptation feedback, a heart rate of 45-60 beats / min and a body movement frequency of <0.5 times / 10 minutes indicate deep sleep, while a heart rate of 50-65 beats / min and a body movement frequency of 0.5-2 times / 10 minutes indicate light sleep.

6. The method for multi-zone pressure monitoring and feedback of a sleep monitoring mattress according to claim 1, characterized in that: The monitoring and feedback system includes a pressure sensing module, a biometric acquisition module, a data processing and decision-making module, a feedback execution module, and an adaptive learning module. The pressure sensing module acquires real-time pressure data of the mattress contact area based on a multi-regional pressure sensor array. The biometric acquisition module acquires biometric signals based on a photoplethysmography (PPG) sensor, a triaxial accelerometer, and a microphone array integrated within the mattress. The data processing and decision-making module processes monitoring data and generates feedback decisions based on an edge computing chip with a built-in adaptive region segmentation algorithm, a multimodal sleep state recognition model, and a personalized feedback strategy library. The feedback execution module executes feedback strategies based on the mattress's built-in zoned air pump, electro-variable pressure pad, and temperature control module. The adaptive learning module is used for data acquisition and recording, optimizing model parameters, and updating personalized profiles. The adaptive learning module also includes a built-in user feedback learning unit that optimizes feedback parameters based on the "valid / invalid feedback" information marked by the user via the app.

7. The method for multi-zone pressure monitoring and feedback of a sleep monitoring mattress according to claim 6, characterized in that: The multi-regional pressure sensor array consists of flexible thin-film pressure sensors, with a sensor density of ≥25 sensors / dm² on the human torso and buttocks. 2 Head and leg sensor density ≥9 units / dm 2 .

8. The method for multi-zone pressure monitoring and feedback of a sleep monitoring mattress according to claim 6, characterized in that: The photoplethysmography (PPG) sensor is used to acquire heart rate and blood oxygen saturation data and fluctuations, the triaxial accelerometer is used to acquire body movement frequency and amplitude data, and the microphone array is used to acquire respiratory sound signals.

9. The method for multi-zone pressure monitoring and feedback of a sleep monitoring mattress according to claim 6, characterized in that: The zoned air pumps are distributed in the torso, hip, head, and leg areas of the mattress; the electro-variable mattress pad is embedded in the surface of the mattress and adjusts the local support stiffness through voltage; the temperature control module is distributed in each contact area of ​​the human body and adjusts the local temperature through Pellets, with a temperature adjustment range of 20-30℃.

10. The method for multi-zone pressure monitoring and feedback of a sleep monitoring mattress according to claim 6, characterized in that: The monitoring and feedback system also includes a communication module that connects to a companion app via Wi-Fi / Bluetooth. This module is used to set sleep preferences and receive alerts. The communication module also has a built-in data sharing unit that supports synchronizing sleep data with a medical and health app.

Images