Rehabilitation physiotherapy device adopting electromagnetic pulse and ion release and control method

Abstract

The invention discloses a rehabilitation physical therapy device adopting electromagnetic pulse and ion release and a control method, and relates to the technical field of physical therapy equipment, the rehabilitation physical therapy device is used in a control module of the rehabilitation physical therapy device, and the rehabilitation physical therapy device further comprises an electromagnetic pulse generator, an ion release unit and a physiological sensor module, the control method comprises the steps that real-time physiological data, collected by the physiological sensor module, of a user is acquired; preprocessing the physiological data of the user; analyzing the current physiological state of the user according to the physiological data of the user; according to the current physiological state of the user, the control module adjusts the working frequency, intensity and / or working time of the electromagnetic pulse generator and the ion concentration and / or release time released by the ion release unit in real time; generating a personalized treatment scheme based on the historical physiological data of the user and the feedback result; dynamic treatment is achieved by integrating multiple physiological data, and the rehabilitation effect and actual use experience of a patient can be effectively enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field related to physical therapy equipment, and in particular to a rehabilitation physical therapy device and a control method using electromagnetic pulses and ion release. Background Art

[0002] Rehabilitation therapy device is an important equipment in the modern medical field. It combines a variety of modern technologies with traditional Chinese medicine treatment methods to treat various diseases and promote the recovery of patients. The rehabilitation therapy device using electromagnetic pulses and ion release combines electromagnetic pulses and ion release technology to achieve a more efficient effect of promoting human tissue repair and regeneration. The device can not only generate electromagnetic pulses, but also release negative ions, using the synergistic effect of these two physical effects to enhance the effect of rehabilitation treatment.

[0003] The inventors found that although the existing rehabilitation therapy devices have good therapy effects, in actual use, they can only select parameters such as the use time of electromagnetic pulses and ion release or the working intensity of the equipment based on the user's subjective feelings or the experience of medical staff. This will to a certain extent fail to achieve the expected rehabilitation therapy effects, resulting in unstable therapy effects, which in turn affects the user's actual usage experience. Summary of the invention

[0004] In order to solve the defects of the prior art, the present invention provides a rehabilitation therapy device and a control method using electromagnetic pulses and ion release.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] The present invention provides a control method for a rehabilitation therapy device, which is used in a control module of the rehabilitation therapy device. The rehabilitation therapy device also includes an electromagnetic pulse generator, an ion release unit and a physiological sensor module. The control method includes:

[0007] Acquire real-time physiological data of the user collected by the physiological sensor module;

[0008] Preprocessing the user's physiological data;

[0009] Analyze the user's current physiological state based on the user's physiological data;

[0010] According to the current physiological state of the user, the control module adjusts the operating frequency, intensity and / or operating time of the electromagnetic pulse generator, as well as the ion concentration and / or release time released by the ion release unit in real time;

[0011] Generate personalized treatment plans based on the user's historical physiological data and feedback results.

[0012] As a preferred technical solution of the present invention, the control module includes:

[0013] A physiological data processing unit, used to receive the real-time physiological data of the user collected by the physiological sensor module and pre-process and analyze various physiological data of the user;

[0014] a physiological state determination unit configured to determine the physiological state of the user according to the data processed by the physiological data processing unit;

[0015] The control unit is configured to provide corresponding treatment methods according to the user's physiological state.

[0016] As a preferred technical solution of the present invention, the physiological sensor module includes a heart rate sensor, an electromyography sensor and a skin galvanic response sensor.

[0017] As a preferred technical solution of the present invention, the physiological data processing unit is used for:

[0018] Data cleaning of heart rate sensor data, heart rate peak detection, heart rate calculation and obtaining a standardized heart rate list;

[0019] Data cleaning of EMG sensor data and obtaining effective values;

[0020] And data cleaning of skin galvanic response sensor data and obtaining skin conductivity change information.

[0021] As a preferred technical solution of the present invention, a classification model and a regression model are selected to judge the user's physiological state based on the data collected by the heart rate sensor, the electromyography sensor and the galvanic skin response sensor.

[0022] As a preferred technical solution of the present invention, random forest is selected as the classification model, and linear regression is selected as the regression model;

[0023] The user's physiological state includes a relaxed state, a slightly tense state, a moderately tense state, a high-pressure state, a fatigue state, and a recovery state.

[0024] As a preferred technical solution of the present invention, the feedback result is the physiological data of the user and the change of the physiological state of the user after one of the treatment schemes has worked for a period of time Δt.

[0025] As a preferred technical solution of the present invention, the working steps of the control unit are reinforcement learning framework, policy optimization target and policy gradient algorithm;

[0026] Reinforcement Learning Framework:

[0027] In the reinforcement learning framework, we need to construct a state vector s to represent the current user's physiological state.

[0028]

[0029] The strategy for selecting action a in a given state s is as follows:

[0030] π(as;θ)

[0031] Where a represents the selectable treatment mode, and θ represents the parameter of the strategy;

[0032] After executing action a, we get the feedback result and construct the reward function:

[0033] R(s,a)

[0034] Strategy optimization goals:

[0035] In reinforcement learning, the goal of the control unit is to optimize the policy to maximize the expected reward and construct the formula:

[0036]

[0037] J(π) represents the expected return under strategy π, E represents the average return calculated over all possible states and actions, t represents the current time, starting from 0 and up to T, γ represents the discount factor, whose value range is [0,1], which determines the importance of future rewards, and r t Denote as the immediate reward received at time step t.

[0038] As a preferred technical solution of the present invention, the steps of the policy gradient algorithm include collecting experience, calculating policy gradients and updating policy parameters;

[0039] After selecting an action and observing the state and reward at each time step t, the gradient of the policy is calculated using the following formula:

[0040]

[0041] represents the impact of changes in the strategy parameter θ on the expected return J, π(a t ∣s t ; θ) is expressed as t Next select action a t The probability of t Denotes the treatment mode selected at time step t, R t Expressed as the total return starting from time step t, it can be calculated as:

[0042]

[0043] R tIt is represented as the total return starting from time step t, represents the cumulative reward from time step t to T, k is represented as a counting variable, represents the time step, which can be any time step between t and T, and r k is represented as the reward received at time step k, kt represents the number of time steps past time step t, and the exponent of the discount factor;

[0044] Finally, the policy parameters are updated according to the calculated gradients:

[0045]

[0046] α is the learning rate, which controls the amplitude of parameter update.

[0047] The present invention also provides a rehabilitation therapy device using electromagnetic pulses and ion release, comprising an electromagnetic pulse generator, an ion release unit, a physiological sensor module and a control module;

[0048] The control module is used to execute a control method for a rehabilitation therapy device.

[0049] The beneficial effects of the present invention are:

[0050] The present invention combines modern sensing technology, machine learning algorithms and personalized treatment plans to achieve adaptive rehabilitation treatment. This device can monitor the user's physiological state in real time and adjust the therapy based on its feedback, providing a more scientific and personalized rehabilitation plan. Different from the single treatment mode used by existing rehabilitation physiotherapy devices that use electromagnetic pulses and ion release, this device realizes dynamic treatment by integrating multiple physiological data, which can effectively enhance the patient's rehabilitation effect and actual usage experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0052] In the attached picture:

[0053] Figure 1 The figure is an overall flow chart of the control method of the rehabilitation therapy device.

[0054] Figure 2 The workflow of the physiological data processing unit.

[0055] Figure 3 Analyze the process for the physiological status determination unit.

[0056] Figure 4 Generate a flow chart for feedback mechanisms and personalized treatment plans. DETAILED DESCRIPTION

[0057] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0058] like Figure 1 As shown, the present invention provides a control method for a rehabilitation therapy device, which is used in a control module of the rehabilitation therapy device. The rehabilitation therapy device also includes an electromagnetic pulse generator, an ion release unit and a physiological sensor module. The electromagnetic pulse generator is used to relieve pain and promote rehabilitation. When the device is turned on, the power supply introduces current into the electromagnetic coil, and the rapidly switched current generates high-frequency electromagnetic pulses. These pulses can penetrate the skin, stimulate cells and tissues in the body, improve local blood circulation, and relieve pain. The electromagnetic pulses have an adjustable frequency (1Hz-1000Hz) and are adjusted according to different diseases and individual needs to optimize the treatment effect.

[0059] The ion release unit is used to improve muscle and nerve function and promote blood circulation. The ion generator uses the principle of high voltage discharge to generate negative ions. These negative ions have antioxidant properties, can neutralize free radicals in the body, and improve cell metabolism. Negative ions are quickly diffused to the treatment area through a small fan, increasing the concentration of negative ions in the air, thereby improving the patient's overall comfort and alleviating physical discomfort caused by stress;

[0060] The combination of electromagnetic pulses and negative ions can effectively activate cells and secrete regenerative factors, thereby accelerating wound healing and relieving chronic pain. Electromagnetic pulses can reduce local inflammation, while negative ions can enhance the immune response of cells. The dual effects make the recovery effect more significant.

[0061] The control method comprises:

[0062] Acquire real-time physiological data of the user collected by the physiological sensor module;

[0063] Preprocessing the user's physiological data;

[0064] Analyze the user's current physiological state based on the user's physiological data;

[0065] According to the current physiological state of the user, the control module adjusts the operating frequency, intensity and / or operating time of the electromagnetic pulse generator, as well as the ion concentration and / or release time released by the ion release unit in real time;

[0066] Generate personalized treatment plans based on the user's historical physiological data and feedback results.

[0067] Its working principle is: before use, the device collects the user's initial physiological data, such as resting heart rate, muscle activity level and skin conductivity, through a sensor array. During the treatment process, it continuously monitors these physiological signals to capture the user's physical reactions and changes, and uses machine learning algorithms to analyze real-time data, extract key information, and evaluate the user's recovery status. According to the data analysis results, the device adjusts the treatment parameters in real time, such as the intensity, frequency and application time of electromagnetic stimulation, and generates a personalized treatment plan based on the user's historical data and instant feedback. If the user fails to achieve the expected effect under a certain treatment, the system will automatically adjust to another treatment mode to optimize the recovery process.

[0068] Furthermore, the control module includes:

[0069] A physiological data processing unit, used to receive the real-time physiological data of the user collected by the physiological sensor module and pre-process and analyze various physiological data of the user;

[0070] a physiological state determination unit configured to determine the physiological state of the user according to the data processed by the physiological data processing unit;

[0071] The control unit is configured to provide a corresponding treatment method according to the user's physiological state and activate the electromagnetic pulse generator and / or the ion release unit.

[0072] Further, if Figure 2 As shown, the physiological sensor module includes a heart rate sensor, an electromyography sensor and a galvanic skin response sensor. The heart rate sensor provides the user's heart rate data, usually expressed in beats per minute (BPM). Heart rate is an important indicator for evaluating physical activity, stress level and overall health status; the electromyography sensor monitors the electrical activity of the muscle, and is usually used to evaluate the activity state, fatigue level and other health problems of the muscle. The characteristics provided by the EMG signal include the contraction force and fatigue level of the muscle; the galvanic skin response sensor reflects the user's emotional state, stress level and other psychological and physiological reactions by measuring the electrical conductivity of the skin. The galvanic skin response can provide information on the user's psychological state, which is helpful for evaluating diseases related to psychological stress. The data signals collected by the heart rate sensor, the electromyography sensor and the galvanic skin response sensor are recorded as HR, EMG and GSR in this embodiment;

[0073] The physiological data processing unit is used for:

[0074] Data cleaning of heart rate sensor data, heart rate peak detection, heart rate calculation and obtaining a standardized heart rate list;

[0075] Data cleaning of EMG sensor data and obtaining effective values;

[0076] And data cleaning of skin galvanic response sensor data and obtaining skin conductivity change information.

[0077] Further, if Figure 3 As shown, based on the data collected by the heart rate sensor, the electromyography sensor, and the galvanic skin response sensor, a classification model and a regression model are selected to judge the user's physiological state, and a threshold-based rule is used for detection, such as fatigue detection, setting a threshold of the feature value, and determining the fatigue state when the RMS index is lower than the preset value;

[0078] Preferably, random forest is selected as the classification model, and linear regression is selected as the regression model;

[0079] Random forest is an integrated learning method that constructs multiple decision trees and uses voting to perform the final classification. It has high accuracy and stability. Physiological signal data such as heart rate, EMG and GSR may contain multiple features. Random forest can effectively process high-dimensional data. Random forest is highly robust to noise and outliers, and is suitable for physiological signals, which are often accompanied by noise.

[0080] The working process of random forest is as follows:

[0081] Data collection: collect and process sensor data of heart rate, EMG, and GSR. The data set should contain input features (such as heart rate, EMG, and GSR values) and their corresponding labels (such as state classification: relaxed, tense, fatigued, etc.);

[0082] Data preprocessing: standardize or normalize the data and divide the data set into training and test sets, usually in a ratio of 70% training set and 30% test set;

[0083] Random forest model training randomly extracts samples from the training set with replacement to generate multiple subsets. Each subset can be used to train a decision tree. When building each decision tree, a few features are randomly selected from a specific number of features for optimal splitting. A hyperparameter m is set to represent the number of features considered at each split.

[0084] Decision tree construction, each tree is trained using self-service samples, and after the construction is completed, a complete decision tree is formed;

[0085] Prediction and integration, when predicting a new sample, each tree outputs a category label, and the final prediction result is determined by voting on the prediction results of all trees, and the category with the highest frequency is selected as the output of the model;

[0086] Model evaluation uses a test set to evaluate model performance. Common indicators include accuracy, precision, recall, etc., to evaluate the influence of different features on the model and guide subsequent analysis and feature optimization.

[0087] In a random forest, the prediction of each subtree can be expressed as:

[0088]

[0089] Where: y^ t is the prediction result of the tth tree, f t is the decision function of the t-th tree, X represents the input feature, and T is the total number of decision trees;

[0090] The final prediction result is:

[0091]

[0092] Linear regression is a very intuitive regression method that can easily understand and explain the relationship between input features and output targets.

[0093] Linear regression works as follows:

[0094] Data collection: Collect data from heart rate, EMG, and GSR sensors, including target variables (such as fatigue level) and input features (such as heart rate, EMG, GSR, etc.);

[0095] Data preprocessing, standardizing or normalizing features, and dividing data into training and test sets;

[0096] Linear regression model training, assuming that the target variable y and the feature X = [x1, x2, x3, ..., x n ]There is a linear relationship between:

[0097] y=β0+β1x1+β2X2+...+βnx+∈

[0098] y is the dependent variable, x1,x2,,...,x n is the independent variable, β1, β2, ...βn are the regression coefficients, and ∈ is the error term.

[0099] Parameter estimation, regression coefficients are obtained by minimizing the following loss function:

[0100]

[0101] Where m is the number of samples, y i is the true value of the i-th sample, xj i is the jth feature of the i-th sample;

[0102] Prediction, use the trained model to predict the new sample, and the predicted value is:

[0103]

[0104] Finally, the model is evaluated using the test set to assess the performance of the model.

[0105] By selecting random forest as the classification model and linear regression as the regression model, and using data collected by heart rate, EMG and GSR sensors, the user's physiological state can be effectively analyzed. Through clear steps of dividing data, training models, and verifying models, combined with reasonable physiological state judgment, personalized health management suggestions can be provided to users. This comprehensive method is not only widely used, but can also continuously optimize itself based on new data.

[0106] Furthermore, the user's physiological state includes a relaxed state, a slightly tense state, a moderately tense state, a high-pressure state, a fatigue state, and a recovery state.

[0107] In this embodiment, an explanation is given, in the relaxed state: normal heart rate, low GSR and low EMG activity level;

[0108] Mild stress: heart rate is slightly higher, GSR is slightly increased, but EMG activity level is still low;

[0109] Moderate stress state: heart rate increases, GSR increases significantly, and EMG activity increases;

[0110] High-pressure state: heart rate increases significantly, GSR increases significantly, and EMG activity is intense;

[0111] Fatigue state: normal or low heart rate, significantly reduced EMG activity (manifested as low RMS and low MF), GSR performance depends on individual differences;

[0112] Recovery state: Heart rate gradually decreases, GSR returns to normal, and EMG activity levels return to those of the relaxed state.

[0113] For different physiological states (relaxed state, slightly tense state, moderately tense state, high pressure state, fatigue state, recovery state), some reasonable values ​​are provided according to the general range of heart rate (in beats / minute), electromyogram (EMG, in microvolts (μV)) and galvanic skin response (GSR, in microSiemens (μS)). The following are the reference values ​​of these three physiological signals in various states:

[0114]

[0115] Heart rate (bpm):

[0116] Relaxed state: heart rate is lower, usually between 60 and 70 bpm;

[0117] Mild stress: Heart rate rises slightly, reflecting mild psychological stress;

[0118] Moderate tension: heart rate increases significantly, indicating a moderately tense emotional state;

[0119] High-pressure state: heart rate continues to rise, often seen in states of tension or stress;

[0120] Fatigue: Heart rate may drop to normal range or slightly lower, indicating that the body may be working hard to recover;

[0121] Recovery state: The heart rate returns to the normal range, indicating that the body is recovering.

[0122] EMG (RMS value):

[0123] Relaxed state: low muscle activity, RMS value remains at a low level;

[0124] Slight tension: muscle activity increases slightly and the RMS value rises;

[0125] Moderate tension state: muscle tension, RMS value increases significantly;

[0126] High-pressure state: muscles are very tense and EMG values ​​reach their maximum value;

[0127] Fatigue state: muscle contraction ability weakens and the amplitude of EMG signal decreases;

[0128] Recovery state: Similar to the relaxation state, the EMG value gradually decreases.

[0129] GSR(μS):

[0130] Relaxed state: The skin electrical response is the lowest, indicating a lower physiological response;

[0131] Mild tension state: GSR rises slightly, indicating a mild emotional experience;

[0132] Moderate stress state: GSR increases significantly, reflecting stronger emotional stress;

[0133] High-pressure state: GSR reaches high values, indicating an extreme pressure state;

[0134] Fatigue state: GSR changes vary, and individual differences can affect the results;

[0135] Recovery state: GSR value returns to the level of relaxation state.

[0136] The above values ​​are based on reference values ​​of physiological signals under normal circumstances. The specific values ​​may vary due to individual differences, environmental factors and other physiological conditions.

[0137] In practical applications, it is necessary to collect and analyze a large amount of data to establish a more accurate model for specific populations or individuals.

[0138] Furthermore, the feedback result is the physiological data of the user and the change of the physiological state of the user after one of the treatment schemes has been in effect for a period of time Δt.

[0139] The following is an example of a feedback result:

[0140] First, before the treatment plan is implemented, record the user's baseline physiological data as a basis for comparison: Heart rate baseline: HR baseline , EMG baseline: EMG baseline , GSR baseline: GSR baseline；

[0141] Assume that a treatment plan starts and lasts for Δt. After Δt ends, the values ​​of each sensor are recorded: HR (Δt), EMG (Δt) and GSR (Δt).

[0142] The change in physiological state is calculated as follows:

[0143] ΔHR=HR(Δt)-HR baseline ;

[0144] ΔEMG=EMG(Δt)-EMG baseline ;

[0145] ΔGSR=GSR(Δt)-GSR baseline ;

[0146] Based on the above changes, the user's physiological state can be comprehensively evaluated. For example, the following classification can be used:

[0147] Improved physiological status: If ΔHR<0, ΔEMG<0 and ΔGSR<0, it means that the user's physiological status has improved;

[0148] Physiological state deterioration: If ΔHR>0, ΔEMG>0 and ΔGSR>0, it means that the user's physiological state has deteriorated;

[0149] If the changes are close to zero, it means that there is no significant change in the state;

[0150] Feedback result example:

[0151] After the treatment plan was executed for Δt=30 minutes, the user's physiological data were as follows:

[0152] Heart rate: HR(Δt)=70bpm(HR baseline =80bpm), ΔHR = -10bpm, indicating that the user's relaxation state is enhanced;

[0153] Electromyography: EMG(Δt)=150μV(EMG baseline =180μV), ΔEMG = -30μV, indicating that muscle tension decreased;

[0154] Skin galvanic response: GSR(Δt)=5.0μS(GSR baseline =6.0μS), ΔGSR=-1.0μS, indicating that the anxiety is reduced to a certain extent.

[0155] Through such detailed feedback, users can clearly understand the impact of treatment on their physiological state, thereby making more effective adjustments and decisions in health management, which is also beneficial to the work of the control unit.

[0156] It should be noted that the above only describes the objective data obtained by the device based on the user's various physiological data. On the other hand, the user can self-rate through the interface, such as changes in pain level or mobility, and further adjust the treatment strategy. For example, when the user still feels uncomfortable after a period of treatment, he or she can click the button on the screen (such as: good, good, moderate, uncomfortable, very uncomfortable) to feedback the user's subjective feelings to the machine, which is more conducive to the machine making more effective adjustments and decisions in health management.

[0157] Further, if Figure 4 As shown, the working steps of the control unit are reinforcement learning framework, policy optimization target and policy gradient algorithm;

[0158] Reinforcement Learning Framework:

[0159] In the reinforcement learning framework, we need to construct a state vector s to represent the current user's physiological state.

[0160]

[0161] The strategy for selecting action a in a given state s is as follows:

[0162] π(as;θ)

[0163] Where a represents the selectable treatment mode (e.g., electromagnetic pulse mode 1, electromagnetic pulse mode 2, ion release mode 1, ion release mode 2, different treatment modes correspond to different working frequencies, intensities and / or working times of the electromagnetic pulse generator, and different ion concentrations and / or release times of the ion release unit), θ represents the parameters of the strategy, such as weights and biases, which are updated as the learning process progresses;

[0164] After executing action a, we get the feedback result and construct the reward function:

[0165] R(s,a)

[0166] Strategy optimization goals:

[0167] In reinforcement learning, the goal of the control unit is to optimize the policy to maximize the expected reward and construct the formula:

[0168]

[0169] J(π) represents the expected return under strategy π, E represents the average return calculated in all possible states and actions, t represents the current time, starting from 0 and up to T, γ represents the discount factor, its value range is [0,1], which determines the importance of future rewards, such as: when it is close to 1, it emphasizes future rewards, and when it is close to 0, it emphasizes current rewards; r t Denoted as the immediate reward received at time step t;

[0170] The steps of the policy gradient algorithm include collecting experience, calculating policy gradients, and updating policy parameters;

[0171] After selecting an action and observing the state and reward at each time step t, the gradient of the policy is calculated using the following formula:

[0172]

[0173] represents the impact of changes in the strategy parameter θ on the expected return J, π(a t ∣s t ; θ) is expressed as t Next select action a t The probability of t Denotes the treatment mode selected at time step t, R t Expressed as the total return starting from time step t, it can be calculated as:

[0174]

[0175] R t It is represented as the total return starting from time step t, represents the cumulative reward from time step t to T, k is represented as a counting variable, represents the time step, which can be any time step between t and T, and r k is represented as the reward received at time step k, kt represents the number of time steps past time step t, and the exponent of the discount factor;

[0176] Finally, the policy parameters are updated according to the calculated gradients:

[0177]

[0178] α is the learning rate, which controls the amplitude of parameter update.

[0179] It realizes accurate analysis of the user's physiological state and generation of personalized treatment plans. Such a system has the advantages of real-time, strong adaptability and high efficiency. It can provide users with better health management services, thereby optimizing treatment effects and improving user experience.

[0180] An embodiment of the present application also provides a rehabilitation therapy device using electromagnetic pulses and ion release. In one example, the rehabilitation therapy device includes an electromagnetic pulse generator, an ion release unit, a physiological sensor module, and a control module.

[0181] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A control method for a rehabilitation therapy device, characterized in that: In a control module for a rehabilitation therapy device, the rehabilitation therapy device further includes an electromagnetic pulse generator, an ion release unit and a physiological sensor module, and the control method includes: Acquire real-time physiological data of the user collected by the physiological sensor module; Preprocessing the user's physiological data; Analyze the user's current physiological state based on the user's physiological data; According to the current physiological state of the user, the control module adjusts the operating frequency, intensity and / or operating time of the electromagnetic pulse generator, as well as the ion concentration and / or release time released by the ion release unit in real time; Generate personalized treatment plans based on the user's historical physiological data and feedback results.

2. The control method of the rehabilitation therapy device according to claim 1, characterized in that: The control module comprises: A physiological data processing unit, used to receive the real-time physiological data of the user collected by the physiological sensor module and pre-process and analyze various physiological data of the user; a physiological state determination unit configured to determine the physiological state of the user according to the data processed by the physiological data processing unit; The control unit is configured to provide corresponding treatment methods according to the user's physiological state.

3. The control method of the rehabilitation therapy device according to claim 2, characterized in that: The physiological sensor module includes a heart rate sensor, an electromyography sensor, and a galvanic skin response sensor.

4. The control method of the rehabilitation therapy device according to claim 3, characterized in that: The physiological data processing unit is used for: Data cleaning of heart rate sensor data, heart rate peak detection, heart rate calculation and obtaining a standardized heart rate list; Data cleaning of EMG sensor data and obtaining effective values; And data cleaning of skin galvanic response sensor data and obtaining skin conductivity change information.

5. The control method of the rehabilitation therapy device according to any one of claims 3-4, characterized in that: Based on the data collected by the heart rate sensor, the electromyography sensor, and the galvanic skin response sensor, a classification model and a regression model are selected to determine the user's physiological state.

6. The control method of the rehabilitation therapy device according to claim 5, characterized in that: Select Random Forest as the classification model and Linear Regression as the regression model. The user's physiological state includes a relaxed state, a slightly tense state, a moderately tense state, a high-pressure state, a fatigue state, and a recovery state.

7. The control method of the rehabilitation therapy device according to claim 1, characterized in that: The feedback result is the physiological data of the user and the change of the physiological state of the user after one of the treatment plans has been working for a period of time Δt.

8. The control method of the rehabilitation therapy device according to claim 7, characterized in that: The working steps of the control unit are reinforcement learning framework, policy optimization objectives and policy gradient algorithm; Reinforcement Learning Framework: In the reinforcement learning framework, we need to construct a state vector s to represent the current user's physiological state. The strategy for selecting action a in a given state s is as follows: π(as;θ) Where a represents the selectable treatment mode, and θ represents the parameter of the strategy; After executing action a, we get the feedback result and construct the reward function: R(s,a) Strategy optimization goals: In reinforcement learning, the goal of the control unit is to optimize the policy to maximize the expected reward and construct the formula: J(π) represents the expected return under strategy π, E represents the average return calculated over all possible states and actions, t represents the current time, starting from 0 and up to T, γ represents the discount factor, whose value range is [0,1], which determines the importance of future rewards, and r t Denote as the immediate reward received at time step t.

9. The control method of the rehabilitation therapy device according to claim 8, characterized in that: The steps of the policy gradient algorithm include collecting experience, calculating policy gradients, and updating policy parameters; After selecting an action and observing the state and reward at each time step t, the gradient of the policy is calculated using the following formula: represents the impact of changes in the strategy parameter θ on the expected return J, π(a t ∣s t ; θ) is expressed as t Next select action a t The probability of t Denotes the treatment mode selected at time step t, R t Expressed as the total return starting from time step t, it can be calculated as: R t It is represented as the total return starting from time step t, represents the cumulative reward from time step t to T, k is represented as a counting variable, represents the time step, which can be any time step between t and T, and r k is represented as the reward received at time step k, kt represents the number of time steps past time step t, and the exponent of the discount factor; Finally, the policy parameters are updated according to the calculated gradients: α is the learning rate, which controls the amplitude of parameter update.

10. A rehabilitation therapy device using electromagnetic pulses and ion release, characterized in that: It includes an electromagnetic pulse generator, an ion release unit, a physiological sensor module and a control module; The control module is used to execute the control method of the rehabilitation therapy device described in any one of claims 1-9.

Images