Intelligent bed body linkage control system and method based on sleep state perception

Abstract

The invention discloses an intelligent bed body linkage control system and method based on sleep state perception, and relates to the technical field of intelligent home furnishing.The system comprises the steps that snore data are continuously detected through a sleep state perception module, and bed body pressure data are obtained through a bed body perception module; the sleeping posture recognition module inputs the sleeping state sensing data and the bed body pressure data for feature extraction, and performs sleeping posture recognition to obtain a first sleeping posture label; the bed body angle control module performs strategy analysis according to the first sleeping posture label, outputs an angle control strategy and performs linkage adjustment by controlling a bed body pneumatic device. The technical problems that in the prior art, the bed body angle cannot be dynamically adjusted according to the change of the sleep state of the user, individuation and dynamic adaptability are lacked, and the sleep adjusting capacity is poor are solved, and the purposes that individualized bed body linkage control is conducted by monitoring the sleep state of the user and bed body pressure data in real time, and the user experience is improved are achieved. And the sleep regulation capability is improved.

Description

Technical Field

[0001] The present invention relates to the field of smart home technology, and in particular to a smart bed linkage control system and method based on sleep state perception. Background Art

[0002] With the development of smart home technology, traditional beds are unable to dynamically adjust to the user's sleep state and posture, resulting in insufficient support or discomfort, which in turn affects sleep quality. Existing smart bed systems mostly rely on fixed angle adjustment functions, unable to perceive changes in the user's sleep depth and sleeping position in real time, and unable to provide personalized adjustment solutions. Furthermore, most existing technologies only focus on a single data source and lack the ability to comprehensively consider the user's sleep state and bed pressure distribution. This makes it impossible to coordinate control with the bed and optimize the sleeping environment in real time. Summary of the Invention

[0003] The present application provides an intelligent bed linkage control system and method based on sleep state perception, which is used to solve the technical problems that the existing technology cannot dynamically adjust the bed angle according to changes in the user's sleep state, lacks personalization and dynamic adaptability, and has poor sleep regulation ability.

[0004] The first aspect of the present application provides an intelligent bed linkage control system based on sleep state perception, the system comprising: a sleep state perception module, which continuously detects sleep state perception data of a sleeping user according to the sleep state perception module, the sleep state perception data including snoring perception data; a bed perception module, which obtains bed pressure perception data according to the bed perception module; a sleeping posture recognition module, which inputs the sleep state perception data and the bed pressure perception data into the sleeping posture recognition module for feature extraction to obtain a state change feature vector and a pressure distribution feature vector, performs sleeping posture recognition according to the state change feature vector and the pressure distribution feature vector, and obtains a first sleeping posture label; a bed angle control module, which inputs the first sleeping posture label into the bed angle control module for strategy analysis to output an angle control strategy, and controls the bed pneumatic device for linkage adjustment according to the angle control strategy.

[0005] A second aspect of the present application provides an intelligent bed linkage control method based on sleep state perception, the method comprising: continuously detecting sleep state perception data of a sleeping user according to a sleep state perception module, the sleep state perception data including snoring perception data; obtaining bed pressure perception data according to the bed perception module; inputting the sleep state perception data and the bed pressure perception data into a sleeping posture recognition module for feature extraction to obtain a state change feature vector and a pressure distribution feature vector, performing sleeping posture recognition according to the state change feature vector and the pressure distribution feature vector to obtain a first sleeping posture label; inputting the first sleeping posture label into a bed angle control module for strategy analysis to output an angle control strategy, and controlling the bed pneumatic device for linkage adjustment according to the angle control strategy.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] The present application provides an intelligent bed linkage control system and method based on sleep state perception, which relates to the field of smart home technology. The sleep state perception module monitors snoring and bed pressure data, and combines the feature vectors extracted by the sleep posture recognition module to identify the sleep posture and generate a sleep posture label. The bed angle control module outputs an angle control strategy based on the label analysis strategy, and adjusts the bed angle through a pneumatic device to achieve dynamic adjustment and improve sleep comfort and quality. This solves the technical problems of the existing technology that cannot dynamically adjust the bed angle according to changes in the user's sleep state, lacks personalization and dynamic adaptability, and has poor sleep adjustment capabilities. It achieves the technical effect of improving sleep adjustment capabilities by performing personalized bed linkage control through real-time monitoring of the user's sleep state and bed pressure data. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0009] Figure 1 A schematic diagram of the structure of an intelligent bed linkage control system based on sleep state perception provided in an embodiment of the present application;

[0010] Figure 2 This is a flow chart of the intelligent bed linkage control method based on sleep state perception provided in an embodiment of the present application.

[0011] Description of reference numerals: sleeping state sensing module 10 , bed sensing module 20 , sleeping posture recognition module 30 , bed angle control module 40 . DETAILED DESCRIPTION

[0012] The present application provides an intelligent bed linkage control system and method based on sleep state perception, which is used to solve the technical problems that the existing technology cannot dynamically adjust the bed angle according to changes in the user's sleep state, lacks personalization and dynamic adaptability, and has poor sleep regulation ability.

[0013] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only some of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0014] It should be noted that the terms "first", "second", etc. in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices.

[0015] Example 1, as Figure 1 As shown, the present application provides an intelligent bed linkage control system based on sleep state perception, which includes:

[0016] The sleep state sensing module 10 continuously detects sleep state sensing data of a sleeping user, wherein the sleep state sensing data includes snoring sensing data.

[0017] Specifically, the main function of the sleep state sensing module 10 of the present application is to continuously monitor the user's sleep state, especially to judge the user's sleep quality and posture by collecting snoring perception data. The module can be integrated into a bed or mattress, and use a sound sensor (such as a microphone) to collect the user's snoring during sleep in real time. Snoring perception data refers to the snoring signal emitted by the user during sleep, which is collected by the sound sensor. These signals contain characteristic parameters such as the frequency, loudness, and duration of the snoring. For example, frequency refers to the speed of snoring vibration, usually measured in Hertz (Hz); loudness refers to the strength of snoring, usually measured in decibels (dB); and duration refers to the length of time that snoring occurs continuously. By analyzing these characteristic parameters, it is possible to preliminarily judge the user's breathing state and possible sleep problems, such as sleep apnea syndrome.

[0018] In terms of technical implementation, the sensors used in the snoring detection module must have high sensitivity and strong anti-interference capabilities to ensure that they can accurately capture snoring signals and effectively distinguish the user's snoring from other environmental noise interference. For example, using a highly sensitive microphone can accurately detect snoring in noisy environments and convert it into an electrical signal output. In addition, during the audio signal acquisition process, the collected characteristic data such as sound frequency, loudness, and duration can provide key data support for the subsequent sleeping posture recognition module.

[0019] In terms of data processing, the Sleep State Perception Module uses built-in signal processing algorithms to perform noise reduction and filtering on the collected audio signals, eliminating background noise interference and ensuring the high accuracy of the extracted snoring signals. Using common signal processing techniques including time domain analysis and frequency domain analysis, it can extract key information such as frequency and intensity from the audio signals, providing effective input data for subsequent sleep posture analysis.

[0020] In summary, the sleep state sensing module 10 provides the system with important data about the user's sleep state by continuously detecting snoring and extracting key snoring features, thereby supporting the coordinated control of the entire smart bed system.

[0021] Furthermore, the sleep state sensing module 10 further includes a sleep depth analysis module, which is connected to the bed angle control module 40 and is configured to perform the following steps:

[0022] P11: Acquire a multidimensional physiological data set through a physiological signal sensing device, wherein the multidimensional physiological data set includes heart rate changes, respiratory rate and body movement frequency; P12: Perform physiological feature convolution on the multidimensional physiological data set and output a multidimensional physiological feature vector; P13: Fuse the multidimensional physiological feature vector into a time series model for sleep depth identification and output a sleep depth label; P14: Input the sleep depth label into the bed angle control module for constraining the angle control strategy.

[0023] It should be understood that the sleep state perception module 10 further includes a sleep depth analysis module, which is connected to the bed angle control module 40, and can accurately analyze the user's sleep depth, and input the analysis results as constraints into the bed angle control module 40 to optimize the bed angle adjustment strategy, thereby achieving a more humane and precise sleep environment adjustment.

[0024] Specifically, the sleep depth analysis module first obtains a multi-dimensional physiological data set through physiological signal sensing devices. These physiological signal sensing devices include but are not limited to heart rate sensors, respiratory sensors, and body motion sensors, which are used to monitor the user's heart rate changes, respiratory rate, and body motion frequency during sleep. Heart rate changes are captured in real time by the heart rate sensor, which can reflect the user's heart rate and its fluctuations; respiratory rate is monitored by the respiratory sensor to obtain the speed and regularity of the user's breathing; body motion frequency is recorded by the body motion sensor to record the user's physical activities during sleep. Through these physiological data, the sleep depth analysis module can obtain key information about the user's sleep depth.

[0025] Next, the sleep depth analysis module performs physiological feature convolution on the collected multi-dimensional physiological data. The process of physiological feature convolution is to process signals such as heart rate, respiratory rate, and body movement frequency through a convolution algorithm to extract representative physiological features from them. For example, the convolution algorithm will help capture the volatility of heart rate, changes in respiratory rhythm, and the frequency of body movements. These features are crucial for judging the user's sleep depth. Through this feature extraction, the system can eliminate noise, highlight key signals that help determine sleep depth, and form multi-dimensional physiological feature vectors. These vectors provide clear input data for subsequent sleep depth identification.

[0026] Subsequently, the sleep depth analysis module inputs these extracted multi-dimensional physiological feature vectors into the time series model for sleep depth identification. The time series model can identify different sleep stages based on the temporal changes of physiological signals. Specifically, time series analysis algorithms such as long short-term memory networks (LSTM) can process complex time series data and capture the dynamic changes of physiological data over time. By training the model, the system can accurately determine the user's sleep depth, identify whether the user is in deep sleep, light sleep or rapid eye movement (REM) sleep, and generate corresponding sleep depth labels based on these identification results.

[0027] Finally, the sleep depth analysis module transmits the identified sleep depth label to the bed angle control module 40. When the bed angle control module 40 formulates the angle adjustment strategy, the sleep depth label is taken into consideration as a constraint. This means that when the bed angle control module 40 performs the angle adjustment operation, it will not only refer to the sleeping posture information provided by the sleeping posture recognition module 30, but also combine the sleep depth label to ensure that the bed angle adjustment will not have a negative impact on the user's sleep depth. For example, when the user is in a deep sleep stage, the bed angle adjustment strategy will be more cautious to avoid disturbing the user's sleep quality due to excessive angle changes; when the user is in a light sleep stage, the bed angle adjustment can be relatively flexible to improve the user's sleeping posture and reduce snoring and other problems. Through this adjustment strategy combined with sleep depth information, the system can achieve more accurate and personalized sleep environment adjustment, further improving the user's sleep quality and overall user experience.

[0028] Through the above steps, the sleep depth analysis module works in conjunction with the sleep state sensing module 10 and the bed angle control module 40 to provide the user with an intelligent bed linkage control system that can monitor the sleep state in real time and make intelligent adjustments according to the sleep depth.

[0029] The bed sensing module 20 acquires bed pressure sensing data.

[0030] Specifically, the bed sensing module 20 of this application is a key component in the intelligent bed linkage control system for monitoring the contact status between the bed and the human body. Its core function is to obtain real-time bed pressure sensing data. This data can reflect the user's body distribution on the bed and pressure changes, providing an important basis for sleeping posture recognition and bed angle adjustment.

[0031] In practice, the bed sensing module 20 primarily collects data through an array of pressure sensors installed on or within the bed. These pressure sensors can accurately and sensitively detect the pressure distribution at the points of contact between the human body and the bed. The pressure sensor array typically consists of multiple sensor units located in key areas of the bed, such as the head, shoulders, waist, hips, and legs. Through these sensor units, the module can capture the pressure applied by various body parts to the bed in different sleeping positions.

[0032] The process for acquiring bed pressure sensing data is as follows: When a user lies on the bed, various parts of the body come into contact with the bed, generating pressure. Each sensor unit in the pressure sensor array detects pressure changes at its corresponding location in real time and converts these pressure values ​​into electrical signals. These electrical signals are then transmitted to the signal processing unit of the bed sensing module 20. The signal processing unit filters, amplifies, and digitizes the raw pressure signals to eliminate noise interference and extract accurate pressure data. This processed pressure data forms a dataset containing multiple pressure values, namely the bed pressure sensing data.

[0033] Bed pressure sensing data not only reflects the pressure distribution across the user's body, but also can be used to infer changes in the user's sleeping position through changes in pressure distribution. For example, when a user shifts from a supine position to a side-lying position, the shift in body center of gravity causes a change in pressure distribution. The bed sensing module 20 can identify this change in sleeping position through the detected pressure changes. Furthermore, bed pressure sensing data can be used to assess the comfort of the bed. By analyzing the uniformity of pressure distribution, it can be determined whether there are any areas of excessive pressure, thereby providing a reference for optimal bed design.

[0034] In this system, the bed pressure sensing data acquired by the bed sensing module 20 is transmitted to the sleeping posture recognition module 30 for comprehensive analysis along with other sleep state sensing data to achieve more accurate sleeping posture recognition. Simultaneously, this pressure sensing data is also used in strategy analysis by the bed angle control module 40, providing an important reference for automatic bed angle adjustment. For example, if excessive pressure is detected in a certain part of the user's body, the bed angle control module 40 can adjust the local angle of the bed to disperse the pressure, thereby improving the user's sleeping comfort.

[0035] The sleeping posture recognition module 30 inputs the sleeping state perception data and the bed pressure perception data into the sleeping posture recognition module 30 for feature extraction, obtains a state change feature vector and a pressure distribution feature vector, performs sleeping posture recognition based on the state change feature vector and the pressure distribution feature vector, and obtains a first sleeping posture label.

[0036] Furthermore, the sleeping posture recognition module 30 is configured to perform the following steps:

[0037] P31: Perform feature extraction on the sleep state perception data to obtain a state change feature vector, which includes the amplitude envelope, frequency band energy distribution, main frequency change curve and periodic jitter of the snoring sound; P32: Perform feature extraction on the bed pressure perception data to obtain a pressure distribution feature vector, which includes the spatial distribution of pressure, support surface change and center of gravity offset path; P33: Perform sleeping posture recognition on the state change feature vector and the pressure distribution feature vector to obtain a first sleeping posture label.

[0038] Optionally, the main task of the sleeping posture recognition module 30 of the present application is to identify the user's sleeping posture through feature extraction and analysis based on the input sleeping state perception data and bed pressure perception data, and provide the system with a corresponding sleeping posture label.

[0039] First, the sleeping posture recognition module 30 performs feature extraction on the sleep state perception data. The sleep state perception data mainly includes snoring data, which can reflect the user's sleep state and sleeping posture. By analyzing the snoring data, the module extracts a state change feature vector. The feature vector includes multiple important features, such as the amplitude envelope, frequency band energy distribution, main frequency change curve and periodic jitter of the snoring. The amplitude envelope reflects the intensity change of snoring, the frequency band energy distribution provides the energy distribution of snoring in different frequency bands, the main frequency change curve describes the frequency change of snoring over time, and the periodic jitter reflects the fluctuation of snoring in period. Through these features, the sleeping posture recognition module can identify the impact of different sleeping postures on snoring. For example, the snoring frequency is lower and the loudness is higher in the supine posture, and the duration is longer, while snoring is lighter or does not occur in the side-lying posture.

[0040] Next, the sleeping posture recognition module 30 performs feature extraction on the bed pressure sensing data. The bed pressure sensing data provides information on the pressure distribution in different areas of the bed surface, which can be used to determine the user's sleeping posture. By analyzing the bed pressure sensing data, a pressure distribution feature vector is extracted. This feature vector includes the spatial distribution of pressure, the change in the support surface, and the center of gravity offset path. The spatial distribution of pressure reflects the distribution of the user's weight on the bed surface, the change in the support surface describes the changes in the force-bearing area on the bed surface, and the center of gravity offset path indicates whether the user has experienced any changes in body position during sleep and the movement trajectory of the center of gravity.

[0041] Finally, the sleeping posture recognition module 30 comprehensively analyzes the extracted state change feature vector and pressure distribution feature vector to identify the user's sleeping posture, and comprehensively analyzes the two feature vectors. Exemplarily, a pre-trained machine learning model (such as a support vector machine, a decision tree or a deep learning model) is used for sleeping posture recognition. By learning a large amount of sample data, the model can accurately identify the user's current sleeping posture, such as supine, side (left or right), prone, etc., and output the corresponding first sleeping posture label. This label provides accurate sleeping posture information for the subsequent bed angle control module 40, so that the module can formulate corresponding angle adjustment strategies according to different sleeping postures, thereby realizing intelligent linkage adjustment of the bed and improving the user's sleep quality.

[0042] Furthermore, the sleeping posture recognition module 30 is embedded with a sleeping posture recognition model, and performs sleeping posture recognition on the state change feature vector and the pressure distribution feature vector using the sleeping posture recognition model, including:

[0043] P33-1: Obtain the state change feature vector samples and the pressure distribution feature vector samples under the known sleeping posture label, and the known sleeping posture labels include supine, left side, right side and prone; P33-2: After time-series synchronization processing of the state change feature vector samples and the pressure distribution feature vector samples, construct the fusion vector samples under the known sleeping posture label; P33-3: Initialize the architecture of the time-space coupling model; P33-4: Use the optimizer to perform iterative supervised training on the time-space coupling model according to the fusion vector samples under the known sleeping posture label to obtain a trained sleeping posture recognition model.

[0044] In one possible embodiment of the present application, the sleeping posture recognition module 30 implements sleeping posture recognition based on the state change feature vector and the pressure distribution feature vector by embedding a sleeping posture recognition model. The training and construction process of the model includes the following detailed steps:

[0045] First, obtain state change feature vector samples and pressure distribution feature vector samples under known sleeping posture labels. Known sleeping posture labels usually include supine, left side, right side, and prone. The state change feature vector and pressure distribution feature vector are obtained through the aforementioned feature extraction steps. Each sleeping posture corresponds to different snoring characteristics and bed pressure distribution patterns. These sample data with known sleeping posture labels are collected from volunteers in an experimental environment in different sleeping postures, ensuring the diversity and accuracy of the data. For example, in the supine state, the amplitude envelope of the snoring may be large and stable, while the spatial distribution of pressure is mainly concentrated on the back and buttocks; in the side-lying state, the snoring characteristics and pressure distribution will be different. These sample data under known sleeping posture labels will be used as training data sets for subsequent model training.

[0046] Next, the system performs time-series synchronization processing on these state change feature vector samples and pressure distribution feature vector samples. Because different sensors may collect data with different frequencies and timestamps, time-series synchronization is required to ensure the temporal consistency of the two feature vectors. This process uses interpolation, alignment, and other technical means to align the time dimensions of the two feature vectors, thereby constructing a fused vector sample under a known sleeping posture label. The fused vector sample is an organic combination of the state change feature vector and the pressure distribution feature vector, which can more comprehensively reflect the user's physiological and mechanical characteristics in different sleeping postures.

[0047] On this basis, the architecture of the time-space coupling model is initialized. The time-space coupling model is a machine learning model that can simultaneously consider time series characteristics and spatial distribution characteristics. It is suitable for processing complex data structures in fusion vector samples. This model is typically built based on a deep learning framework, such as using a convolutional neural network (CNN) to process spatial features and a recurrent neural network (RNN) or long short-term memory network (LSTM) to process time series features. Through this architecture, the model can capture the temporal and spatial coupling relationship between the state change feature vector and the pressure distribution feature vector, thereby more accurately identifying sleeping posture.

[0048] Finally, an optimizer performs iterative supervised training on the time-space coupling model based on the fused vector samples under known sleeping posture labels. In each round of training, the optimizer adjusts the model parameters based on the error between the current model's predictions and the actual labels, gradually improving recognition accuracy. Through iterative training, after multiple iterations, the model learns the characteristic patterns of different sleeping postures in the state change feature vectors and pressure distribution feature vectors, ultimately achieving a fully trained sleeping posture recognition model.

[0049] In actual applications, when the system obtains new state change feature vectors and pressure distribution feature vectors, the sleeping posture recognition model can quickly and accurately output the first sleeping posture label, providing reliable sleeping posture information for the subsequent bed angle control module, thereby realizing the linkage adjustment of the intelligent bed and improving the user's sleeping experience.

[0050] The bed angle control module 40 inputs the first sleeping posture label into the bed angle control module 40 for strategy analysis to output an angle control strategy. The bed angle control module 40 then controls the bed pneumatic device for coordinated adjustment according to the angle control strategy. Furthermore, the bed angle control module 40 includes a sleeping posture-angle mapping strategy library, obtained by accessing a sleep research literature library, for storing angle adjustment strategies corresponding to different sleeping posture labels.

[0051] It should be understood that the primary function of the bed angle control module 40 of this application is to perform a strategy analysis based on the input first sleeping posture label, output a corresponding angle control strategy, and implement coordinated adjustments by controlling the bed's pneumatic devices to optimize the user's sleep quality. This process relies not only on the module's internal logic analysis but also on its internal sleeping posture-angle mapping strategy library, which draws data from a database of sleep research literature to ensure the scientific and effective adjustment strategies.

[0052] Specifically, after the first sleeping posture label is input into the bed angle control module 40, the module first queries the sleeping posture-angle mapping strategy library. The strategy library is a comprehensive data storage system that obtains rich data resources by accessing the sleep research literature library. The sleep research literature library covers a wide range of sleep research literature, including but not limited to clinical experimental reports, sleep medicine research papers, and sleep monitoring data analysis reports. These literature materials are systematically sorted and analyzed to extract statistical analysis data that identify the relationship between different sleeping posture labels and respiratory status, snoring intensity, and sleep quality. For example, studies have shown that supine posture may cause partial obstruction of the airway, thereby increasing snoring intensity and the risk of apnea; while side-lying posture is relatively helpful in improving respiratory patency, reducing snoring intensity, and thus improving sleep quality.

[0053] Based on these statistical analysis data, the sleeping posture-angle mapping strategy library has developed corresponding angle adjustment strategies for each sleeping posture label. These strategies fully consider the impact of different sleeping postures on the human respiratory system and sleep quality, and aim to optimize the user's sleeping environment by adjusting the angle of the bed. For example, when the first sleeping posture label input is "supine", the strategy library will recommend raising the head of the bed moderately based on the data in the literature library to reduce the possibility of airway obstruction, thereby reducing snoring intensity and improving breathing status; if the sleeping posture label input is "side sleeping", it may be recommended to keep the bed relatively flat or make slight angle adjustments to maintain the natural curve of the body while avoiding unnecessary interference with breathing. For example, as shown in Table 1:

[0054] Table 1: Sleeping posture-angle mapping strategy library category table

[0055]

[0056]

[0057] After obtaining the corresponding angle adjustment strategy, the bed angle control module 40 converts the strategy into specific control instructions and sends them to the bed pneumatic device. The bed pneumatic device is the physical component that performs angle adjustment and usually includes a cylinder, an air pump, and a related control system. These devices accurately adjust the angles of various parts of the bed according to the control instructions received. For example, by controlling the extension and retraction of the cylinder, the head, foot, or side of the bed can be raised or lowered. The entire adjustment process is automated and interconnected, and can respond quickly according to the user's real-time sleeping posture changes, ensuring that the user is always in a comfortable and breathing-friendly sleeping posture throughout the sleep process.

[0058] Furthermore, the bed angle control module 40 possesses a degree of adaptive capability. It can fine-tune and optimize the angle adjustment strategy based on user feedback and long-term sleep data. For example, if the system detects that a user responds well to a particular angle adjustment strategy in a specific sleeping position—that is, snoring intensity is significantly reduced and sleep quality is improved—the module will store and apply this strategy as a priority. This adaptive mechanism enables the system to better adapt to individual differences among users, further enhancing the system's intelligence and user experience.

[0059] Through this precise control mechanism, the bed angle control module 40 can adjust the bed angle in real time to meet the needs of users in different sleeping positions. At the same time, with the help of statistical data from the sleep research literature library, it ensures that each adjustment can scientifically, reasonably and effectively improve the user's sleep experience.

[0060] Furthermore, the bed angle control module 40 is further configured to perform the following steps:

[0061] P41: Receive the sleep depth label, and obtain the matching adjustment step corresponding to the sleep depth label according to the adaptive step model; P42: Input the first sleeping posture label into the bed angle control module for strategy analysis to output the angle control strategy, and then perform gradient adjustment on the angle control strategy according to the matching adjustment step to update the angle control strategy.

[0062] Optionally, the bed angle control module 40 can not only adjust the angle according to the sleeping posture label, but also perform adaptive adjustment according to the sleep depth label to achieve more personalized and refined bed angle control.

[0063] First, the bed angle control module 40 receives the sleep depth label from the sleep depth analysis module. The sleep depth label reflects the user's current sleep stage, such as light sleep, deep sleep or rapid eye movement (REM) sleep. The bed angle control module 40 obtains the corresponding matching adjustment step according to the preset adaptive step model based on the sleep depth label. The adaptive step model is an adjustment strategy optimization mechanism based on sleep depth. Its core idea is to dynamically adjust the step size of the bed angle adjustment according to the user's current sleep depth. For example, in the deep sleep stage, the user is more sensitive to environmental changes, so the adjustment step size should be smaller to avoid disturbing the user's sleep due to excessive angle changes; in the shallow sleep stage, the user has a strong adaptability to environmental changes, and the adjustment step size can be appropriately increased to quickly improve the sleeping posture.

[0064] Specifically, the adaptive step-size model establishes a mapping relationship between sleep depth and adjustment step size by analyzing a large amount of sleep research data and experimental results. This mapping relationship is stored in the internal database of the bed angle control module 40. When a new sleep depth tag is received, the module can quickly query and obtain the corresponding matching adjustment step size. For example, if the sleep depth tag indicates that the user is in deep sleep, the adaptive step-size model may recommend a smaller adjustment step size, such as adjusting the angle no more than 2 degrees at a time; if the user is in light sleep, the adjustment step size may be increased to 5 degrees.

[0065] After obtaining the matching adjustment step, the bed angle control module 40 inputs the first sleeping posture label for strategy analysis and outputs the initial angle control strategy. This strategy is formulated based on the sleeping posture-angle mapping strategy library, and aims to adjust the bed angle according to the user's current sleeping posture to optimize the sleeping environment. However, in order to further improve the accuracy and adaptability of the adjustment strategy, the initial angle control strategy can be gradient-adjusted according to the matching adjustment step. Gradient adjustment is a gradual optimization process that gradually approaches the optimal angle by decomposing the initial adjustment angle into multiple small step adjustments. For example, if the initial angle control strategy recommends raising the head of the bed by 10 degrees, and the matching adjustment step is 2 degrees, then the adjustment process can be decomposed into 5 steps, raising 2 degrees each time, and gradually completing the angle adjustment.

[0066] After each adjustment step, the bed angle control module 40 monitors the user's physiological reactions and sleep state changes in real time to evaluate the adjustment effect. If the user experiences sleep interruptions or other discomfort during the adjustment process, the module automatically adjusts the subsequent adjustment step or pauses the adjustment to ensure that the user's sleep quality is not affected. Through this dynamic gradient adjustment mechanism, the bed angle control module 40 can flexibly adjust the bed angle based on the user's real-time sleep depth and sleeping posture changes, achieving more personalized and precise sleep environment optimization.

[0067] Ultimately, the gradient-adjusted angle control strategy is updated and applied to the bed's pneumatic control system. Based on the updated strategy, the pneumatic system precisely adjusts the angles of various bed components, providing a comfortable and healthy sleeping environment. This process not only considers the user's sleeping position but also fully accounts for the impact of sleep depth on the adjustment strategy, significantly enhancing the intelligence level and user experience of the intelligent bed linkage control system.

[0068] Furthermore, constructing the adaptive step size model includes:

[0069] P41a: Set multiple sleep depth labels and multiple adjustment step intervals corresponding to the multiple sleep depth labels; wherein the multiple sleep depth labels include a first sleep depth label, a second sleep depth label, and a third sleep depth label, and the first sleep depth label is greater than the second sleep depth label and greater than the third sleep depth label; the multiple adjustment step intervals include a first step interval, a second step interval, and a third step interval, and the first step interval is greater than the second step interval and greater than the third step interval. P42a: Collect adjustment response data samples of multiple target users based on the multiple adjustment step intervals under the multiple sleep depth labels, the adjustment response data samples are data that characterize the degree of impact on the user's sleep; P43a: Obtain multiple step optimal solutions corresponding to the multiple adjustment step intervals based on the adjustment response data samples; P44a: Establish a mapping relationship between the multiple sleep depth labels and the multiple step optimal solutions, and output an adaptive step model.

[0070] Specifically, the process of building an adaptive step length model can be further refined to dynamically adjust the step length of bed angle adjustment according to changes in different sleep depth labels, thereby optimizing the user's sleeping experience.

[0071] First, set multiple sleep depth labels and multiple adjustment step intervals corresponding to these sleep depth labels. The sleep depth label includes a first sleep depth label, a second sleep depth label and a third sleep depth label, wherein the first sleep depth label represents a deep sleep state, the second sleep depth label represents a light sleep state, and the third sleep depth label represents a rapid eye movement (REM) sleep state. In order to ensure that the system can be flexibly adjusted according to the needs of different sleep depths, multiple adjustment step intervals are set. Specifically, the first step interval corresponds to the deep sleep stage and has a larger adjustment amplitude; the second step interval corresponds to the light sleep stage and has a moderate step length; the third step interval corresponds to the rapid eye movement (REM) sleep stage and has a smaller step length, that is, the first step interval> the second step interval> the third step interval, thereby ensuring that the user obtains a more delicate angle adjustment in deep sleep, and makes appropriate adjustments in the light sleep and REM stages.

[0072] Next, data samples of the adjustment responses of multiple target users in the corresponding adjustment step intervals under different sleep depth labels are collected. These data samples are key indicators that characterize the degree of impact on the user's sleep, and may include, for example, changes in heart rate, breathing rate, body movement frequency, number of awakenings, etc. By conducting long-term monitoring of the target users in an experimental environment, their physiological and behavioral responses to different step length adjustments at different sleep depths are recorded. For example, when a larger step length is used for adjustment in the light sleep stage (the third sleep depth label), the user's heart rate and body movement frequency may increase significantly, while when a smaller step length is used for adjustment in the deep sleep stage (the first sleep depth label), the user's physiological response is relatively small.

[0073] Subsequently, based on the collected adjustment response data samples, the impact of different step length adjustments on the user's sleep under each sleep depth label is analyzed. Statistical analysis methods (such as variance analysis, correlation analysis, etc.) are used to evaluate the degree of impact of adjustments in different step length intervals on sleep quality. For example, the user's average heart rate change, number of awakenings and other indicators are calculated under each step length interval to determine the optimal step length interval under each sleep depth label. Using these optimal step length intervals as the optimal step length solution can minimize interference with the user's sleep while achieving effective adjustment of the bed angle.

[0074] Finally, based on the above analysis results, a mapping relationship between the sleep depth label and the optimal step length solution is established. For example, the first sleep depth label (deep sleep) is mapped to the third step length interval (smaller step length), the second sleep depth label (medium deep sleep) is mapped to the second step length interval (medium step length), and the third sleep depth label (light sleep) is mapped to the first step length interval (larger step length). Through this mapping relationship, the adaptive step length model can intelligently select the appropriate adjustment step length according to the user's sleep depth label in actual applications to achieve precise adjustment of the bed angle.

[0075] Through the above steps, the constructed adaptive step length model can not only dynamically adjust the adjustment step length of the bed angle according to the user's sleep depth state, but also provide personalized adjustment solutions in different sleep stages, thereby effectively improving the user's sleep quality.

[0076] Furthermore, in the bed angle control module 40, the bed pneumatic device includes multiple pneumatic adjustment units, and the multiple pneumatic adjustment units include a head pneumatic unit, a back pneumatic unit, a waist pneumatic unit and a leg pneumatic unit; the angle control strategy is received and converted through the control interface module, and the pneumatic execution parameters are output, and the pneumatic execution parameters are used to control one or more pneumatic units in the multiple pneumatic adjustment units to perform linkage adjustment.

[0077] Optionally, when executing the angle control strategy, the bed angle control module 40 implements specific bed adjustment functions through its internal bed pneumatic device. This bed pneumatic device is composed of multiple pneumatic adjustment units, specifically including a head pneumatic unit, a back pneumatic unit, a waist pneumatic unit, and a leg pneumatic unit. These pneumatic adjustment units correspond to different parts of the human body and can independently or collaboratively adjust the angle to meet the personalized needs of different sleeping positions and sleep depths.

[0078] After receiving the angle control strategy, the bed angle control module 40 converts the strategy through its built-in control interface module to generate specific pneumatic execution parameters. These pneumatic execution parameters include key information such as the pneumatic unit's adjustment direction (such as raising or lowering), adjustment angle, and adjustment speed. The control interface module's function is to convert the angle control strategy from logical instructions into specific parameters that the pneumatic device can understand and execute, ensuring the accuracy and reliability of the adjustment process.

[0079] Based on the generated pneumatic execution parameters, one or more pneumatic units in the bed's pneumatic system will be activated and perform corresponding adjustment actions. For example, if the angle control strategy requires improving the user's breathing condition, the system may only activate the head pneumatic unit, raising it to a certain angle to reduce airway pressure; if the overall sleeping position needs to be adjusted to improve comfort, the back pneumatic unit and the waist pneumatic unit may be activated at the same time for coordinated adjustment. In this way, the bed's pneumatic system can achieve precise control of different parts of the bed to meet the user's physiological needs at different sleep stages.

[0080] In practice, the bed's pneumatic linkage adjustment function dynamically adjusts to the user's real-time sleeping position and sleep depth. For example, when the user is in deep sleep, the system uses an adaptive step-size model to select a smaller adjustment step size, slowly adjusting the angle of the pneumatic unit to avoid disturbing the user's sleep. In light sleep, the system can use a larger adjustment step size to quickly optimize the sleeping position and reduce snoring or other discomfort symptoms. This flexible adjustment mechanism not only improves the user's sleep quality, but also reflects the humanized design of the intelligent bed linkage control system.

[0081] Furthermore, each pneumatic unit in the bed's pneumatic system is equipped with a high-precision sensor and feedback mechanism, enabling real-time monitoring of pressure changes and angle status during adjustment. This sensor data is fed back to the bed angle control module 40, enabling it to dynamically adjust pneumatic execution parameters to ensure smooth and safe adjustment. For example, if an abnormal pressure is detected in a pneumatic unit during adjustment, the system will immediately suspend the adjustment to prevent discomfort or potential risks to the user.

[0082] In this way, the bed's pneumatic mechanism dynamically adjusts the bed's angle based on the varying sleep position and sleep depth tags, providing comfortable sleep support. This coordinated adjustment of the pneumatic unit ensures the user maintains an optimal sleeping position throughout sleep, improving sleep quality and preventing discomfort caused by an inappropriate bed angle.

[0083] In summary, the embodiments of the present application have at least the following technical effects:

[0084] This application uses a sleep state sensing module to continuously monitor snoring data and pressure data obtained by a bed sensing module. This module then uses the state change feature vectors and pressure distribution feature vectors extracted by a sleep posture recognition module to identify sleep postures and generate a first sleep posture label. The bed angle control module then performs a strategy analysis based on this sleep posture label, outputs an angle control strategy, and uses a pneumatic mechanism to adjust the bed angle, achieving dynamic and personalized bed adjustment to improve the user's sleep quality and comfort.

[0085] The technical effect of real-time monitoring of the user's sleep status and bed pressure data, performing personalized bed linkage control, and improving sleep regulation ability has been achieved.

[0086] The second embodiment is based on the same inventive concept as the intelligent bed linkage control system based on sleep state perception in the previous embodiment. Figure 2 As shown, the present application provides a method for controlling an intelligent bed body linkage based on sleep state perception. The system and method embodiments in the present application are based on the same inventive concept. The method includes:

[0087] The sleep state perception module continuously detects the sleep state perception data of the sleeping user, wherein the sleep state perception data includes snoring perception data; the bed pressure perception data is obtained according to the bed perception module; the sleep state perception data and the bed pressure perception data are input into the sleeping posture recognition module for feature extraction to obtain a state change feature vector and a pressure distribution feature vector, and the sleeping posture is recognized according to the state change feature vector and the pressure distribution feature vector to obtain a first sleeping posture label; the first sleeping posture label is input into the bed angle control module for strategy analysis to output an angle control strategy, and the bed pneumatic device is controlled according to the angle control strategy to perform linkage adjustment.

[0088] Furthermore, the method further comprises:

[0089] A multidimensional physiological data set is obtained through a physiological signal sensing device, wherein the multidimensional physiological data set includes heart rate changes, respiratory rate, and body movement rate; physiological feature convolution is performed on the multidimensional physiological data set to output a multidimensional physiological feature vector; the multidimensional physiological feature vector is fused and input into a time series model for sleep depth identification, and a sleep depth label is output; the sleep depth label is input into the bed angle control module for constraining the angle control strategy.

[0090] Furthermore, the method further comprises:

[0091] Receive the sleep depth label, and obtain the matching adjustment step length corresponding to the sleep depth label according to the adaptive step length model; input the first sleeping posture label into the bed angle control module for strategy analysis to output the angle control strategy, and then gradient adjust the angle control strategy according to the matching adjustment step length to update the angle control strategy.

[0092] Furthermore, the method further comprises:

[0093] Set multiple sleep depth labels and multiple adjustment step intervals corresponding to the multiple sleep depth labels; collect adjustment response data samples of multiple target users based on the multiple adjustment step intervals under the multiple sleep depth labels, and the adjustment response data samples are data that characterize the degree of impact on user sleep; obtain multiple step length optimal solutions corresponding to the multiple adjustment step intervals based on the adjustment response data samples; establish a mapping relationship between the multiple sleep depth labels and the multiple step length optimal solutions, and output an adaptive step length model.

[0094] Furthermore, the multiple sleep depth labels include a first sleep depth label, a second sleep depth label, and a third sleep depth label, and the first sleep depth label is greater than the second sleep depth label, which is greater than the third sleep depth label; the multiple adjustment step intervals include a first step interval, a second step interval, and a third step interval, and the first step interval is greater than the second step interval, which is greater than the third step interval.

[0095] Furthermore, the method further comprises:

[0096] Feature extraction is performed on the sleep state perception data to obtain a state change feature vector, where the state change feature vector includes the amplitude envelope, frequency band energy distribution, main frequency change curve and periodic jitter of the snoring sound; feature extraction is performed on the bed pressure perception data to obtain a pressure distribution feature vector, where the pressure distribution feature vector includes the spatial distribution of pressure, support surface change and center of gravity offset path; sleeping posture recognition is performed on the state change feature vector and the pressure distribution feature vector to obtain a first sleeping posture label.

[0097] Furthermore, the sleeping posture recognition module is embedded with a sleeping posture recognition model, and the sleeping posture recognition is performed on the state change feature vector and the pressure distribution feature vector using the sleeping posture recognition model, including:

[0098] Acquire state change feature vector samples and pressure distribution feature vector samples under known sleeping posture labels, wherein the known sleeping posture labels include supine, left side, right side, and prone; perform time-series synchronization processing on the state change feature vector samples and the pressure distribution feature vector samples, and construct fusion vector samples under the known sleeping posture labels; initialize the architecture of the time-space coupling model; use an optimizer to iteratively supervise the time-space coupling model based on the fusion vector samples under the known sleeping posture labels to obtain a trained sleeping posture recognition model.

[0099] Furthermore, the bed pneumatic device includes multiple pneumatic adjustment units, including a head pneumatic unit, a back pneumatic unit, a waist pneumatic unit and a leg pneumatic unit; the angle control strategy is received and converted through the control interface module, and pneumatic execution parameters are output, and the pneumatic execution parameters are used to control one or more pneumatic units in the multiple pneumatic adjustment units for linkage adjustment.

[0100] Furthermore, the bed angle control module includes a sleeping posture-angle mapping strategy library, which is obtained by accessing a sleep research literature library and is used to store angle adjustment strategies corresponding to different sleeping posture labels.

[0101] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0102] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

[0103] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. An intelligent bed linkage control system based on sleep state perception, characterized in that: The system comprises: a sleep state sensing module, wherein the sleep state sensing module continuously detects sleep state sensing data of a sleeping user, wherein the sleep state sensing data includes snoring sensing data; A bed sensing module, which obtains bed pressure sensing data according to the bed sensing module; a sleeping posture recognition module, inputting the sleeping state sensing data and the bed pressure sensing data into the sleeping posture recognition module for feature extraction to obtain a state change feature vector and a pressure distribution feature vector, performing sleeping posture recognition based on the state change feature vector and the pressure distribution feature vector to obtain a first sleeping posture label; The bed angle control module inputs the first sleeping posture label into the bed angle control module for strategy analysis to output an angle control strategy, and controls the bed pneumatic device to perform linkage adjustment according to the angle control strategy.

2. The system according to claim 1, wherein The sleep state perception module further includes a sleep depth analysis module, which is connected to the bed angle control module and is configured to perform the following steps: Acquiring a multidimensional physiological data set through a physiological signal sensing device, wherein the multidimensional physiological data set includes heart rate changes, respiratory rate, and body movement frequency; Performing physiological feature convolution on the multidimensional physiological data set to output a multidimensional physiological feature vector; fusing the multidimensional physiological feature vectors into a time series model for sleep depth recognition, and outputting a sleep depth label; The sleep depth label is input into the bed angle control module to constrain the angle control strategy.

3. The system according to claim 2, wherein: The bed angle control module is further configured to perform the following steps: receiving the sleep depth label, and obtaining a matching adjustment step length corresponding to the sleep depth label according to an adaptive step length model; After the first sleeping posture label is input into the bed angle control module for strategy analysis to output an angle control strategy, the angle control strategy is gradient-adjusted according to the matching adjustment step to update the angle control strategy.

4. The system according to claim 3, wherein: Constructing the adaptive step size model includes: Setting a plurality of sleep depth labels and a plurality of adjustment step intervals corresponding to the plurality of sleep depth labels; Collecting adjustment response data samples of multiple target users under the multiple sleep depth labels based on the multiple adjustment step intervals, the adjustment response data samples being data representing the degree of impact on the user's sleep; Acquire multiple step-size optimal solutions corresponding to the multiple adjustment step-size intervals according to the adjustment response data samples; A mapping relationship between the multiple sleep depth labels and the multiple step-size optimal solutions is established, and an adaptive step-size model is output.

5. The system according to claim 4, wherein: The multiple sleep depth labels include a first sleep depth label, a second sleep depth label, and a third sleep depth label, and the first sleep depth label is greater than the second sleep depth label and greater than the third sleep depth label; The multiple adjustment step intervals include a first step interval, a second step interval, and a third step interval, and the first step interval is larger than the second step interval, which is larger than the third step interval.

6. The system according to claim 1, wherein: The sleeping posture recognition module is used to perform the following steps: Extracting features from the sleep state perception data to obtain a state change feature vector, wherein the state change feature vector includes an amplitude envelope, frequency band energy distribution, a main frequency change curve, and periodic jitter of the snoring sound; Extracting features from the bed pressure sensing data to obtain a pressure distribution feature vector, wherein the pressure distribution feature vector includes pressure spatial distribution, support surface change, and center of gravity offset path; Sleeping posture recognition is performed on the state change feature vector and the pressure distribution feature vector to obtain a first sleeping posture label.

7. The system according to claim 1, wherein: The sleeping posture recognition module is embedded with a sleeping posture recognition model, and performs sleeping posture recognition on the state change feature vector and the pressure distribution feature vector using the sleeping posture recognition model, including: Obtaining state change feature vector samples and pressure distribution feature vector samples under known sleeping posture labels, wherein the known sleeping posture labels include supine, left side, right side, and prone; After performing time-series synchronization processing on the state change feature vector sample and the pressure distribution feature vector sample, a fusion vector sample under a known sleeping posture label is constructed; Initialize the architecture of the time-space coupling model; An optimizer is used to perform iterative supervised training on the time-space coupling model according to fusion vector samples under known sleeping posture labels to obtain a trained sleeping posture recognition model.

8. The system according to claim 1, wherein: The bed pneumatic device includes a plurality of pneumatic adjustment units, and the plurality of pneumatic adjustment units include a head pneumatic unit, a back pneumatic unit, a waist pneumatic unit and a leg pneumatic unit; The angle control strategy is received and converted through a control interface module, and pneumatic execution parameters are output, and one or more pneumatic units among the multiple pneumatic adjustment units are controlled by the pneumatic execution parameters to perform linkage adjustment.

9. The system according to claim 8, wherein The bed angle control module includes a sleeping posture-angle mapping strategy library, which is obtained by accessing a sleep research literature library and is used to store angle adjustment strategies corresponding to different sleeping posture labels.

10. An intelligent bed linkage control method based on sleep state perception is characterized in that: The method comprises: Continuously detecting sleep state perception data of a sleeping user according to a sleep state perception module, wherein the sleep state perception data includes snoring perception data; Acquire bed pressure sensing data according to the bed sensing module; Inputting the sleep state sensing data and the bed pressure sensing data into a sleeping posture recognition module for feature extraction to obtain a state change feature vector and a pressure distribution feature vector, performing sleeping posture recognition based on the state change feature vector and the pressure distribution feature vector to obtain a first sleeping posture label; The first sleeping posture label is input into the bed angle control module for strategy analysis to output an angle control strategy, and the bed pneumatic device is controlled to perform linkage adjustment according to the angle control strategy.

Images