Intelligent adjustment method and system of shared massage chair based on big data

Abstract

The application relates to the technical field of intelligent health equipment, and discloses an intelligent adjustment method and system of a shared massage chair based on big data, wherein the intelligent adjustment method of the shared massage chair based on big data comprises the following steps: collecting sleep-related physiological indexes of a user to generate a preliminary sleep state feature vector; constructing a personalized spine curve digital model to generate a spine biomechanics feature vector; applying a hidden Markov model to identify the sleep state of the user in real time to generate a complete sleep state vector; establishing a sleep spine collaborative optimization model to construct a state conversion function, a spine mechanics response function and a sleep quality evaluation function; realizing real-time optimization and dynamic adjustment of massage parameters to form a closed-loop control system; the application introduces a state machine and an adaptive trajectory planning algorithm, automatically switches an optimal massage mode, continuously optimizes the strategy through a sleep quality feedback mechanism, and significantly improves the comfortable experience and health management level of the user.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent health devices, more specifically, it relates to an intelligent adjustment method and system for a shared massage chair based on big data. BACKGROUND

[0002] With the acceleration of modern social life rhythm, sleep disorders and spinal health problems are increasingly prominent, and more and more people are facing sleep quality decline, spinal fatigue and chronic pain and other health problems in daily life and work. Massage chairs, as an important device for assisting sleep improvement and spinal health, have been widely used in families, public places and medical rehabilitation and other scenarios. Traditional massage chairs mainly rely on mechanical structure and preset program to provide basic massage and support function, but there are obvious deficiencies in dealing with individual differences and dynamic physiological needs of users.

[0003] In recent years, with the development of biosensor technology and artificial intelligence algorithm, intelligent health devices based on multi-modal physiological signal collection and analysis have gradually emerged. By integrating multiple sensors, physiological parameters such as user's breathing, heart rate and body movement can be obtained in real time, providing data basis for health management and personalized service. However, existing massage chairs have not realized the deep integration and dynamic response of user sleep state and spinal biomechanical characteristics in sleep assistance and spinal regulation, and it is difficult to meet the individual needs of users in different sleep stages and spinal states.

[0004] In addition, the existing technology generally lacks real-time evaluation and feedback mechanism of sleep quality, and cannot continuously optimize massage strategy according to user historical data and subjective experience, resulting in a large space for improvement of massage effect and user experience.

[0005] Therefore, an intelligent massage chair system capable of integrating multi-modal physiological signals and spinal pressure distribution data, with adaptive control and closed-loop optimization capability, is needed to realize the synergistic optimization of sleep quality improvement and spinal health protection. SUMMARY

[0006] The present application provides an intelligent adjustment method and system for a shared massage chair based on big data, which solves the technical problems that multi-modal physiological signals and spinal pressure distribution data cannot be integrated, individualized dynamic control and sleep quality feedback closed-loop mechanism are lacking, and sleep quality improvement and spinal health synergistic optimization are difficult to realize in related technologies.

[0007] The present application provides an intelligent adjustment method for a shared massage chair based on big data, comprising the following steps:

[0008] Collecting user sleep-related physiological indicators to generate a preliminary sleep state feature vector;

[0009] Based on the preliminary sleep state feature vector, the pressure distribution data of the user's spine and the massage chair contact surface is collected, a personalized spine curve digital model is constructed, and a spine biomechanics feature vector is generated;

[0010] Based on the preliminary sleep state feature vector and the spine biomechanics feature vector, a hidden Markov model is applied to identify the user's sleep state in real time, and a complete sleep state vector is generated;

[0011] Based on the complete sleep state vector and the spine biomechanics feature vector, a sleep-spine collaborative optimization model is established, a state transition function, a spine mechanics response function and a sleep quality evaluation function are constructed;

[0012] Based on the sleep-spine collaborative optimization model, real-time optimization and dynamic adjustment of massage parameters are realized, forming a closed-loop control system.

[0013] In a preferred embodiment, the step of collecting user sleep-related physiological indicators includes:

[0014] Integrating millimeter wave radar sensors, far infrared sensors and sound sensors to form a multi-modal sensing system to collect user sleep-related physiological indicators;

[0015] The collected raw biological signals are preprocessed, including noise filtering, signal separation and feature extraction;

[0016] The processed multi-modal physiological features are fused to generate a preliminary sleep state feature vector.

[0017] In a preferred embodiment, the step of collecting pressure distribution data of the user's spine and the massage chair contact surface includes:

[0018] Deploying a high-density pressure sensor array in the massage chair back area to collect pressure distribution data of the user's spine area;

[0019] Based on the collected pressure distribution data, a thin-plate spline interpolation algorithm is applied to reconstruct the user's spine curve;

[0020] Based on the reconstructed spine curve and pressure distribution, a spine biomechanics feature vector is constructed.

[0021] In a preferred embodiment, the step of applying a hidden Markov model to identify the user's sleep state in real time includes:

[0022] The preliminary sleep state feature vector is subjected to time series feature extraction and fusion;

[0023] A sleep state classifier based on a hidden Markov model is constructed to determine the user's sleep stage in real time;

[0024] Based on the identified sleep state sequence, calculate the sleep quality score and characteristic indicators;

[0025] Combine the sleep state classification results with other physiological characteristics to form a complete sleep state vector.

[0026] In a preferred embodiment, the step of establishing a sleep-spine co-optimization model includes:

[0027] Construct a state transition function to predict the trend of the current sleep state under a specific massage parameter;

[0028] Construct a spine mechanics response function to predict the effect of a specific massage parameter on the biomechanical state of the spine;

[0029] Perfect the sleep quality evaluation function to achieve comprehensive evaluation of the sleep state;

[0030] Construct the overall objective function of sleep-spine co-optimization.

[0031] In a preferred embodiment, the overall objective function is formally expressed as: maximize the lifting effect of the massage parameter on the sleep quality under the premise of meeting the safety constraints of the spine.

[0032] In a preferred embodiment, the step of implementing real-time optimization and dynamic adjustment of the massage parameter includes:

[0033] Construct a sleep-spine state machine to define the ideal spine support configuration for different sleep stages;

[0034] Based on the current sleep state and spine characteristics, use an adaptive trajectory planning algorithm to dynamically optimize the massage trajectory;

[0035] Implement multi-dimensional collaborative control of the massage parameters to form a complete massage experience;

[0036] Introduce a sleep quality feedback mechanism to continuously optimize the control strategy through sleep effect evaluation.

[0037] In a preferred embodiment, the sleep-spine state machine includes:

[0038] Define the sleep-spine combined state space, including the combination of sleep stages and spine support modes;

[0039] Based on sleep physiology and biomechanics knowledge, establish the mapping relationship between sleep stages and ideal spine support configuration;

[0040] Define the trigger conditions and smooth transition mechanism of state transition to ensure the comfort during massage mode switching;

[0041] According to the user's spine characteristics and feedback history, optimize the state machine parameters to achieve personalized support configuration.

[0042] In a preferred embodiment, the sleep quality feedback mechanism comprises:

[0043] Based on the sleep state change trend, the immediate effect of the massage intervention is evaluated in real time;

[0044] After each use, the effectiveness of the current massage program is evaluated according to the overall sleep quality score;

[0045] Based on the evaluation results, the massage strategy library and the optimization algorithm parameters are automatically adjusted;

[0046] Combining user subjective feedback and objective sleep data, a comprehensive evaluation is formed to guide the system to continuously optimize.

[0047] In a preferred embodiment, an intelligent adjustment system for a shared massage chair based on big data is used to perform an intelligent adjustment method for a shared massage chair based on big data, comprising:

[0048] A non-contact biological parameter monitoring unit is used to collect user sleep-related physiological indicators and generate a preliminary sleep state feature vector;

[0049] A pressure distribution sensing unit acquires pressure distribution data of the user's spine and the contact surface of the massage chair based on the preliminary sleep state feature vector, constructs a personalized spine curve digital model, and generates a spine biomechanics feature vector;

[0050] A sleep state recognition unit applies a hidden Markov model to identify the user's sleep state in real time based on the preliminary sleep state feature vector and the spine biomechanics feature vector, and generates a complete sleep state vector;

[0051] A collaborative optimization modeling unit establishes a sleep-spine collaborative optimization model based on the complete sleep state vector and the spine biomechanics feature vector, constructs a state transition function, a spine mechanics response function, and a sleep quality evaluation function;

[0052] An adaptive trajectory planning unit realizes real-time optimization and dynamic adjustment of massage parameters based on the sleep-spine collaborative optimization model, forming a closed-loop control system.

[0053] The beneficial effects of the present application are:

[0054] By fusing multi-modal physiological signals and spine pressure distribution data, precise recognition of user sleep state and spine characteristics is achieved, which can dynamically adapt to individual differences of different users, significantly improving the intelligent level of the massage chair. The system can automatically optimize massage parameters according to the user's real-time sleep stage and spine biomechanics state, provide personalized spine support and sleep assistance, and enhance the user's comfort experience and sleep quality.

[0055] The sleep spine cooperative optimization model breaks through the limitation of single target optimization of traditional massage chairs, realizes the cooperative regulation of sleep quality improvement and spine health protection, and automatically switches the optimal massage mode in different sleep stages through the introduction of state machine and adaptive trajectory planning algorithm, so as to meet the dynamic physiological needs of users.

[0056] A sleep quality feedback closed-loop mechanism is established, which can continuously optimize the massage strategy based on user historical data and subjective feedback, realize the self-learning and long-term evolution ability of the system, improve the effectiveness of massage intervention, and bring a more scientific and reliable health management experience to users. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 is a flow chart of an intelligent adjustment method of a shared massage chair based on big data of the present application;

[0058] Figure 2 is a column chart of sleep quality improvement comparison of the present application;

[0059] Figure 3 is a radar chart of spine health index improvement of the present application;

[0060] Figure 4 is a line chart of sleep state recognition accuracy of the present application;

[0061] Figure 5 is a sankey diagram of spine support configuration in different sleep stages of the present application;

[0062] Figure 6 is an area chart of new user adaptability of the present application;

[0063] Figure 7 is a column chart of system response time comparison of the present application. DETAILED DESCRIPTION

[0064] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that the discussion of these implementations is merely meant to provide a better understanding of the subject matter described herein and can be changed in function and arrangement without departing from the scope of the present description. Various processes or components can be omitted, substituted, or added according to desired implementations. Additionally, features described in some examples can be combined in other examples.

[0065] At least one embodiment of the present application discloses an intelligent adjustment method of a shared massage chair based on big data, as shown in Figure 1 The method comprises the following steps:

[0066] Step 1: Collecting user sleep-related physiological indicators and generating a preliminary sleep state feature vector.

[0067] Specifically, the following steps are included:

[0068] Step 1.1: Biological parameter collection;

[0069] A multi-modal sensing system composed of millimeter wave radar sensors, far infrared sensors, and sound sensors is used to collect sleep-related physiological indicators such as respiratory rate, heart rate variability, body movement frequency, and respiratory sound.

[0070] Millimeter wave radar sensor: By emitting 24GHz or 60GHz millimeter waves and receiving reflected waves, it detects small chest and abdominal movements using the Doppler effect principle to collect respiratory frequency and amplitude data.

[0071] Far infrared sensor: Captures changes in user body surface temperature, monitors the association between body temperature regulation and sleep cycle changes.

[0072] Sound sensor: Collects respiratory sound features to assist in sleep state analysis.

[0073] The sensing system is configured in the headrest and backrest of the massage chair, maintaining an effective sensing distance of 20 to 50 cm, ensuring high-quality physiological signals are obtained without touching the user's body.

[0074] Step 1.2: Signal preprocessing;

[0075] The collected raw biological signals are preprocessed, including:

[0076] Noise filtering: Apply adaptive filtering algorithms to eliminate environmental noise and motion artifacts, improving signal-to-noise ratio.

[0077] Signal separation: Use independent component analysis technology to separate overlapping physiological signals, such as separating respiratory signals and heartbeat signals from composite signals.

[0078] Feature extraction: Calculate key time and frequency domain features such as respiratory rate, heart rate, heart rate variability indicators (SDNN, RMSSD, LF / HF ratio), body movement frequency, etc.

[0079] Step 1.3: Preliminary sleep state feature vector generation;

[0080] The processed multi-modal physiological features are fused to generate a preliminary sleep state feature vector. This vector contains the following dimensions:

[0081] Respiratory feature sub-vector: Contains respiratory frequency, depth, regularity, and other parameters.

[0082] Heart rate feature sub-vector: Contains average heart rate, variability, and other parameters.

[0083] Body motion feature sub-vector: contains body motion frequency, amplitude, etc.

[0084] Temperature feature sub-vector: contains body surface temperature and its trend.

[0085] By connecting these sub-vectors and normalizing, a preliminary sleep state feature vector is formed, providing a data basis for subsequent sleep state classification.

[0086] As shown in Figure 2 , the comparison data of sleep quality indicators between the system and traditional massage chairs is shown, which intuitively shows the sleep quality improvement effect brought by the system.

[0087] Step 2, based on the preliminary sleep state feature vector, collect the pressure distribution data of the user's spine and the contact surface of the massage chair, construct the personalized spine curve digital model, and generate the spine biomechanics feature vector;

[0088] Specifically, the following steps are included:

[0089] Step 2.1, pressure distribution data collection;

[0090] Deploy a high-density pressure sensor array in the massage chair backrest area to collect pressure distribution data of the user's spine area:

[0091] Sensor array configuration: arrange a 12×24 flexible pressure sensor array along the longitudinal direction of the backrest, with a sensor unit spacing of 1.5 cm, covering the complete spine area from the cervical vertebra to the sacrum.

[0092] Pressure data sampling: collect pressure distribution matrix data at a frequency of 20Hz, with a pressure value range of 0-1000g / cm² for each sensor unit, and an accuracy of ±5g / cm².

[0093] Initial calibration: when the user uses it for the first time, the system guides the user to perform standard posture calibration collection and records the baseline pressure distribution.

[0094] Step 2.2, spine curve reconstruction;

[0095] Based on the collected pressure distribution data, apply the thin-plate spline interpolation algorithm to reconstruct the user's spine curve:

[0096] Pressure centerline extraction: extract the pressure centerline along the spine by analyzing the local maximum points of the pressure distribution matrix.

[0097] Spline curve fitting: use the extracted pressure center points as control points, and apply the cubic spline interpolation algorithm to generate a continuous spine curve.

[0098] Curve parameterization: Convert the reconstructed spinal curve into a parameterized representation, including the curvature radii and relative positions of the cervical, thoracic, and lumbar curves.

[0099] Step 2.3, Spinal Biomechanics Feature Vector Generation;

[0100] Based on the reconstructed spinal curve and pressure distribution, construct the spinal biomechanics feature vector:

[0101] Curve feature sub-vector: Contains the curvature parameters and relative positions of the cervical, thoracic, and lumbar vertebrae.

[0102] Pressure distribution sub-vector: Contains the pressure distribution characteristics of each spinal segment, such as peak pressure, pressure uniformity, etc.

[0103] Contact area sub-vector: Describes the matching degree of user body shape and massage chair backrest.

[0104] Muscle tension estimation sub-vector: Muscle tension distribution inferred from pressure change patterns.

[0105] By combining the above sub-vectors and performing standardization, a complete spinal biomechanics feature vector is formed, providing constraint conditions for subsequent massage parameter optimization.

[0106] As shown in Figure 3 , the system's improvement effect on multiple spinal health-related indicators is demonstrated, including reduction of spinal pressure peak, improvement of muscle tension balance, and reduction of spinal discomfort.

[0107] Step 3, based on the preliminary sleep state feature vector and spinal biomechanics feature vector, apply the Hidden Markov Model to real-time recognition of user sleep state, generate a complete sleep state vector;

[0108] Specifically including the following steps:

[0109] Step 3.1, sleep feature fusion;

[0110] Time series feature extraction and fusion of preliminary sleep state feature vector:

[0111] Time series feature extraction: Apply sliding window technology (window length 30 seconds, overlap rate 50%), calculate statistical and trend features of each physiological parameter.

[0112] Feature importance evaluation: Apply XGBoost-based feature importance evaluation algorithm to adaptively adjust the weights of different features.

[0113] Multi-modal feature fusion: Use attention mechanism to fuse physiological features of different modalities, and enhance the weights of features with high contribution to sleep state judgment.

[0114] Step 3.2, sleep state classification model;

[0115] Construct a sleep state classifier based on Hidden Markov Model to determine the user's sleep stage in real time:

[0116] State space definition: Define the sleep state space as {awake, sleep transition, light sleep, deep sleep, REM sleep} five discrete states.

[0117] Observation probability modeling: Based on the fused physiological features, construct the observation probability distribution model of each sleep state.

[0118] State transition matrix: According to the rules of sleep physiology, initialize the state transition matrix, and continuously optimize it through user historical data.

[0119] Sequence decoding: Apply Viterbi algorithm to decode the observation sequence to determine the most likely sleep state sequence.

[0120] Step 3.3, sleep quality assessment;

[0121] Based on the identified sleep state sequence, calculate the sleep quality score and characteristic indicators:

[0122] Sleep structure indicators: Calculate indicators such as the time proportion of each sleep stage and transition frequency.

[0123] Sleep continuity indicators: Calculate indicators such as the number of sleep interruptions and average interruption duration.

[0124] Sleep quality comprehensive score:

[0125] ;

[0126] where, represents the final comprehensive sleep quality score; represents the total number of sleep quality indicators included in the evaluation system; represents the th specific sleep quality indicator; represents the th sleep quality indicator corresponding weight coefficient, used to reflect the relative importance of the indicator in the comprehensive score. The sum of all weight coefficients is 1, that is, to ensure the normalization of the score. This scoring method can comprehensively reflect the multi-dimensional sleep quality of the user.

[0127] Step 3.4, complete sleep state vector generation;

[0128] Combine the sleep state classification results with other physiological features to form a complete sleep state vector:

[0129] Sleep stage sub-vector: contains the current sleep stage, time spent in this stage, stage transition probability, etc.

[0130] Sleep quality sub-vector: contains the current sleep quality score, sleep depth index, etc.

[0131] Physiological state sub-vector: contains muscle tone state, autonomic nervous system balance state, etc.

[0132] Sleep trend sub-vector: contains the predicted sleep stage development trend and transition time point.

[0133] The complete sleep state vector provides key input for subsequent massage parameter optimization.

[0134] As shown in Figure 4 , the recognition accuracy of the system in different sleep stages is shown to change over time, reflecting the self-learning ability and adaptability of the system.

[0135] Step 4, based on the complete sleep state vector and the spine biomechanics feature vector, establish the sleep spine collaborative optimization model, construct the state transition function, the spine mechanics response function and the sleep quality evaluation function;

[0136] Specifically, the following steps are included:

[0137] Step 4.1, state transition function establishment;

[0138] Construct the state transition function, which is used to predict the change trend of the current sleep state under specific massage parameters:

[0139] Function definition: construct the state transition function:

[0140] ;

[0141] Where, represents the state transition operator, which is used to describe how the user's sleep state will change under the condition of given current sleep state vector and massage parameter vector ; represents the current sleep state vector; represents the massage parameter vector; represents the expected sleep state vector. This function is used to describe how the user's sleep state changes under the action of specific massage parameters, and is the core prediction tool of the subsequent optimization model.

[0142] Model structure: apply recurrent neural network structure, take sleep state vector and massage parameter vector as input, and predict sleep state at future time point.

[0143] Model Training: Based on historical user data and sleep physiology knowledge, the model is trained using a combination of supervised learning and reinforcement learning.

[0144] Model Validation: The model's prediction accuracy is evaluated through cross-validation methods, and the model parameters are continuously optimized through online learning.

[0145] Step 4.2, Spinal Mechanics Response Function Establishment;

[0146] Establish the spinal mechanics response function to predict the influence of specific massage parameters on the biomechanical state of the spine:

[0147] Function Definition: Establish the response function:

[0148] ;

[0149] where, represents the spinal mechanics response function, which describes the mechanical distribution of the spine under the given massage parameter vector and the spinal biomechanical feature vector . is the spinal biomechanical feature vector, is the massage parameter vector, is the spinal stress distribution vector.

[0150] Model Construction: Based on finite element analysis and biomechanics principles, construct a parameterized spinal mechanics model to simulate the influence of different massage actions on the stress of the spine.

[0151] Fast Calculation Optimization: Through model reduction technology and pre-computation database, achieve millisecond-level real-time response prediction.

[0152] Safety Constraint Definition: Based on biomechanical safety standards and expert knowledge, define the set of spinal biomechanical safety constraints to ensure that massage parameters do not cause discomfort or damage to the spine.

[0153] Step 4.3, Sleep Quality Evaluation Function Optimization;

[0154] Perfect the sleep quality evaluation function to achieve comprehensive evaluation of sleep state:

[0155] Function Improvement: Optimize the sleep quality evaluation function, refine the calculation method of each sleep quality indicator to make the score more comprehensive and reasonable:

[0156] ;

[0157] where, represents the current sleep state vector The higher the value, the better the sleep quality. represents the current sleep state vector, containing the user's multi-dimensional sleep physiological characteristics (such as sleep stage, depth, muscle tension, etc.) at the current time; represents the total number of sleep quality indicators contained in the evaluation system; represents the score function of the th sleep quality indicator, with the input being the current sleep state vector and the output being the score of the indicator (such as sleep depth, continuity, stability, etc.); represents the weight coefficient corresponding to the th sleep quality indicator, reflecting the relative importance of the indicator in the comprehensive score, and all weight coefficients satisfy to ensure the normalization of the score.

[0158] Weight self-adaptation: dynamically adjust the weight coefficients of each indicator according to user characteristics (such as age, gender, sleep habits) and usage scenarios (short rest or overnight sleep) .

[0159] Individualized parameters: combine user historical data to adjust the evaluation function parameters, making them more in line with the user's personal sleep characteristics and needs.

[0160] Step 4.4, collaborative optimization objective function construction;

[0161] Based on the above function, the overall objective function of the sleep spine collaborative optimization is constructed:

[0162] Optimization objective definition: this optimization problem can be formally expressed as , with the constraint condition being , i.e. the stress distribution of the spine under the selected massage parameters needs to meet the safety constraints.

[0163] where represents finding the parameter combination that can maximize the objective function among all possible massage parameter vectors ; represents the massage parameter vector, including massage intensity, frequency, mode, action area, etc. multiple dimensions; represents the current sleep state vector, reflecting the user's current sleep stage, depth, muscle tension, etc. multi-dimensional physiological characteristics; represents the state transition function, with the input being the current sleep state and the massage parameter , and the output being the predicted future sleep state vector ; represents the sleep quality evaluation function, which performs comprehensive scoring on the input sleep state vector, with the higher the value, the better the sleep quality. represents the spine biomechanics feature vector, describing the current biomechanics state of the user's spine; represents the spine mechanics response function, input massage parameters and spine features , output spine stress distribution vector ; represents the set of spine biomechanics safety constraints, which specifies the safety conditions that the spine stress must satisfy. The meaning of this optimization objective is: under the premise of ensuring that the spine stress distribution satisfies the safety constraints , select the optimal massage parameters , so that the user's sleep quality score after state conversion is maximized.

[0164] Multi-objective optimization: combine the two goals of improving sleep quality and protecting spine health through weight coefficients to form a balanced optimization objective.

[0165] Constraint integration: combine various constraint conditions such as spine safety constraints, massage device physical limitations, and user comfort requirements into the optimization framework.

[0166] After completing the model construction, the system obtains a theoretical basis for calculating the optimal massage parameters based on the user's current sleep state and spine features, providing scientific guidance for subsequent adaptive adjustment.

[0167] As shown in Figure 5 , it shows how the system dynamically adjusts the spine support configuration according to different sleep stages, and shows the mapping relationship between sleep state and spine support mode.

[0168] Step 5, based on the sleep-spine collaborative optimization model, realize the real-time optimization and dynamic adjustment of massage parameters, form a closed-loop control system;

[0169] Specifically, the following steps are included:

[0170] Step 5.1, sleep-spine state machine construction;

[0171] Construct a sleep-spine state machine to define the ideal spine support configuration for different sleep stages:

[0172] State definition: define the sleep-spine joint state space, which includes the combination of sleep stages and spine support modes.

[0173] State mapping: based on sleep physiology and biomechanics knowledge, establish the mapping relationship between sleep stages and ideal spine support configurations, such as relaxed support in REM sleep period, stable support in deep sleep period, etc.

[0174] Transition Rules: Define the trigger conditions and smooth transition mechanisms for state transitions, ensuring the comfort during massage mode switching.

[0175] Individualized Adjustment: Optimize state machine parameters based on user's spine characteristics and feedback history, achieving personalized support configuration.

[0176] As Figure 6 shown, the system response time of the system compared with traditional massage chairs in different operating scenarios, showing the advantages of the system in computational efficiency.

[0177] Step 5.2, adaptive trajectory planning algorithm;

[0178] Based on the current sleep state and spine characteristics, the adaptive trajectory planning algorithm is used to dynamically optimize the massage trajectory:

[0179] Trajectory library initialization: Pre-design a basic massage trajectory library for different sleep stages and spine characteristics.

[0180] Real-time trajectory adjustment: Based on the optimization model, real-time optimization and adjustment of the parameters in the trajectory library.

[0181] Trajectory smoothing processing: Apply cubic spline interpolation algorithm to smooth the optimized trajectory, ensuring motion continuity.

[0182] Dynamic response mechanism: Design a dynamic response mechanism for the trajectory, so that the massage trajectory can quickly respond to the mutation of the sleep state.

[0183] Step 5.3, multi-parameter collaborative control;

[0184] Achieve multi-dimensional collaborative control of massage parameters to form a complete massage experience:

[0185] Parameter vector collaborative optimization: Simultaneously optimize force, frequency, trajectory, heat, vibration and other multi-dimensional parameters to achieve overall optimal effect.

[0186] Parameter correlation modeling: Establish a correlation model between different massage parameters to avoid conflicts or conflicts in parameter combinations.

[0187] Hierarchical control strategy: Adopt hierarchical control architecture to separate sleep cycle scale strategy planning and millisecond level motion control, improving system response speed.

[0188] Parameter constraint verification: Before executing specific control, safety constraint verification is performed on the optimized parameters to ensure that they do not exceed the device capacity range and biomechanical safety boundaries.

[0189] Step 5.4, sleep quality feedback closed loop;

[0190] Introduce sleep quality feedback mechanism, continuously optimize regulation strategy through sleep effect evaluation:

[0191] Real-time effect evaluation: based on sleep state change trend, real-time evaluation of immediate effect of massage intervention.

[0192] Post-session evaluation: after each use, evaluate the effectiveness of this massage plan according to overall sleep quality score.

[0193] Strategy adjustment: based on evaluation results, automatically adjust massage strategy library and optimize algorithm parameters.

[0194] User feedback integration: combine user subjective feedback and objective sleep data to form comprehensive evaluation, guide system continuous optimization.

[0195] Through the above steps, the system forms a complete closed-loop control process, which can dynamically optimize massage parameters according to the real-time sleep state and spine characteristics of the user, and continuously improve through continuous feedback learning, realizing truly intelligent and personalized sleep assistance.

[0196] As shown in Figure 7 , it shows the adaptation speed of the system to new users, including the time of constructing personalized model, the degree of completing parameter optimization, and the score of user comfort degree changing with the use time.

[0197] The above describes the embodiments of the present application, but the embodiments are not limited to the specific implementation described above, which is only illustrative and not restrictive. Those skilled in the art can make more forms of equivalent embodiments under the inspiration of the embodiments, which are all within the protection of the embodiments.

Claims

1. A smart adjustment method for a shared massage chair based on big data, characterized in that, Includes the following steps: Collect users' sleep-related physiological indicators to generate preliminary sleep state feature vectors; Based on the preliminary sleep state feature vector, pressure distribution data of the user's spine and the contact surface of the massage chair are collected to construct a personalized digital model of the spinal curve and generate a spinal biomechanical feature vector. The spinal biomechanical feature vector is formed by combining spinal curve feature sub-vectors, pressure distribution sub-vectors, contact area sub-vectors and muscle tension estimation sub-vectors and then standardizing them. Based on preliminary sleep state feature vectors and spinal biomechanical feature vectors, a hidden Markov model is applied to identify the user's sleep state in real time and generate a complete sleep state vector; the complete sleep state vector consists of sleep stage sub-vectors, sleep quality sub-vectors, physiological state sub-vectors, and sleep trend sub-vectors. Based on the complete sleep state vector and spinal biomechanical feature vector, a state transition function and a spinal biomechanical response function are constructed to improve the sleep quality evaluation function; a general objective function for sleep-spinal co-optimization is also constructed; the general objective function for sleep-spinal co-optimization is as follows: The constraints are: ; This means finding, among all the massage parameter vectors P, the option that satisfies the objective function. The parameter combination that achieves the maximum value; S represents the current complete sleep state vector; This represents the state transition function, which takes S and P as inputs and outputs a vector of predicted future sleep states. represents the sleep quality evaluation function, which comprehensively scores the input sleep state vector; the higher the value, the better the sleep quality. B represents the spinal biomechanical feature vector, which describes the current biomechanical state of the user's spine. The input is P and B, and the output is the force distribution vector of the spine; C represents the set of biomechanical safety constraints of the spine, which specifies the safety conditions that the spine must meet under stress. Based on the overall objective function of sleep-spine synergy optimization, massage parameters are optimized and dynamically adjusted in real time, forming a closed-loop control system.

2. The intelligent adjustment method for a shared massage chair based on big data according to claim 1, characterized in that, The steps of collecting users' sleep-related physiological indicators and generating preliminary sleep state feature vectors include: A multimodal sensing system is constructed by integrating millimeter-wave radar sensors, far-infrared sensors, and sound sensors to collect users' sleep-related physiological indicators. The collected raw biological signals are preprocessed, including noise filtering, signal separation, and feature extraction. The processed multimodal physiological features are fused to generate a preliminary sleep state feature vector.

3. The intelligent adjustment method for a shared massage chair based on big data according to claim 1, characterized in that, The steps for generating a complete sleep state vector by applying a hidden Markov model to identify the user's sleep state in real time based on preliminary sleep state feature vectors and spinal biomechanical feature vectors include: Temporal feature extraction and fusion are performed on the initial sleep state feature vector; Construct a sleep state classifier based on a hidden Markov model to determine the user's sleep stage in real time; Based on the identified sleep state sequences, sleep quality scores and characteristic indicators are calculated; The sleep state classification results are combined with other physiological characteristics to form a complete sleep state vector.

4. A smart adjustment system for a shared massage chair based on big data, used to execute the smart adjustment method for a shared massage chair based on big data as described in any one of claims 1-3, characterized in that, include: A non-contact bio-parameter monitoring unit is used to collect users' sleep-related physiological indicators and generate a preliminary sleep state feature vector. The pressure distribution sensing unit, based on a preliminary sleep state feature vector, collects pressure distribution data between the user's spine and the contact surface of the massage chair, constructs a personalized digital model of the spinal curve, and generates a spinal biomechanical feature vector; the spinal biomechanical feature vector is formed by combining spinal curve feature sub-vectors, pressure distribution sub-vectors, contact area sub-vectors, and muscle tension estimation sub-vectors and then standardizing them. The sleep state recognition unit, based on the preliminary sleep state feature vector and spinal biomechanical feature vector, applies a hidden Markov model to identify the user's sleep state in real time and generate a complete sleep state vector; the complete sleep state vector consists of sleep stage sub-vectors, sleep quality sub-vectors, physiological state sub-vectors, and sleep trend sub-vectors. The collaborative optimization modeling unit, based on the complete sleep state vector and spinal biomechanical feature vector, constructs a state transition function, a spinal biomechanical response function, and improves the sleep quality evaluation function; it also constructs an overall objective function for sleep-spine collaborative optimization; the overall objective function for sleep-spine collaborative optimization is... The constraints are: ; This means finding, among all the massage parameter vectors P, the option that satisfies the objective function. The parameter combination that achieves the maximum value; S represents the current complete sleep state vector; This represents the state transition function, which takes S and P as inputs and outputs a vector of predicted future sleep states. represents the sleep quality evaluation function, which comprehensively scores the input sleep state vector; the higher the value, the better the sleep quality. B represents the spinal biomechanical feature vector, which describes the current biomechanical state of the user's spine. The input is P and B, and the output is the force distribution vector of the spine; C represents the set of biomechanical safety constraints of the spine, which specifies the safety conditions that the spine must meet under stress. The adaptive trajectory planning unit, based on the overall objective function of sleep-spine coordination optimization, realizes real-time optimization and dynamic adjustment of massage parameters, forming a closed-loop control system.

Images