Intelligent lower limb rehabilitation training and interaction system

Abstract

The invention relates to the technical field of medical health, and discloses an intelligent lower limb rehabilitation training and interaction system which is characterized in that by monitoring lower limb muscle electric signal state data of a patient during rehabilitation training in real time, at least one change node of the electric signal state data is marked as a memory point; obtaining a state memory point data set; based on the state memory point data set, acquiring rehabilitation training information data fed back by the state memory point data set through a visual interface, so that a patient and a doctor perform cooperative interaction; according to the feedback rehabilitation training information data, the rehabilitation condition of the lower limb muscle electric signal state of the patient is judged, and a rehabilitation training scheme is adjusted in time; when an abnormal condition occurs in the process of adjusting the rehabilitation training scheme, quick response is made and training is immediately suspended so as to output abnormal early warning information; therefore, the patient can perform lower limb rehabilitation training more conveniently and efficiently at home.

Description

Technical Field

[0001] This invention relates to the field of medical and health technology, and more specifically, to an intelligent lower limb rehabilitation training and interaction system. Background Technology

[0002] Rehabilitation training is a crucial part of postoperative rehabilitation for patients with lower limb disorders or fractures. Currently, postoperative rehabilitation for these patients mainly takes place in hospital orthopedic or rehabilitation departments, resulting in long hospital stays, high medical costs, and significant consumption of medical resources. Patients and their families also experience numerous inconveniences regarding diet and rest. After discharge, they often lack timely rehabilitation guidance from doctors and the assistance of medical devices, leading to difficulties in practicing at home, such as the inability or reluctance to exercise. This can result in adverse consequences such as soft tissue adhesions around the joints, joint stiffness, and impaired limb function recovery, even causing permanent disability. To ensure the quality of patient rehabilitation while shortening hospital stays and reducing medical costs, and to allow patients to receive technical guidance from doctors and the assistance of professional medical devices at home, we have adopted a remote smart healthcare model. This model integrates digital sensing technology into traditional mechanical structures and develops remote control programs to enable online interaction between doctors and patients, practicing the concept of smart healthcare services and creating a convenient and technologically advanced service scenario.

[0003] In conclusion, how to make it more convenient and efficient for patients to conduct lower limb rehabilitation training at home is an issue that the intelligent lower limb rehabilitation training and interaction system urgently needs to address and optimize. Summary of the Invention

[0004] This invention provides an intelligent lower limb rehabilitation training and interaction system, which solves the technical problem of how to make it more convenient and efficient for patients to carry out lower limb rehabilitation training in related technologies.

[0005] To address the aforementioned technical problems, this invention provides an intelligent lower limb rehabilitation training and interaction system, the specific technical solution of which is as follows:

[0006] Real-time monitoring of the patient's lower limb muscle electrical signal status data during rehabilitation training; marking at least one node of change in the electrical signal status data as a memory point to obtain a status memory point dataset.

[0007] Based on the state memory point dataset, rehabilitation training information data is obtained through a visual interface to enable collaborative interaction between patients and doctors.

[0008] Based on the feedback of rehabilitation training information data, the patient's lower limb muscle electrical signal status rehabilitation progress is determined, and the rehabilitation training plan is adjusted in a timely manner.

[0009] When abnormal situations occur during the adjustment of the rehabilitation training program, a rapid response should be initiated and training should be immediately suspended to issue an abnormal warning message.

[0010] Furthermore, the training time and cycle are adjusted in real time according to the patient's muscle fatigue level for different rehabilitation training programs in order to avoid excessive strain on intensity and difficulty;

[0011] Based on the patient's recovery progress, the training process is divided into different difficulty stages, and dynamically adjusted according to the patient's recovery status; and the training difficulty is adjusted individually based on the patient's real-time feedback data.

[0012] By analyzing data on patients' activity status, muscle fatigue, and rehabilitation during rehabilitation training, we can assess their lower limb motor function, muscle recovery, and training progress.

[0013] Furthermore, the activity status of the lower limb muscles of the training patient is monitored in real time, and at least one node of activity status change is marked as a memory point to obtain a state memory point dataset, including:

[0014] Wearable sensors are attached to the skin surface of the patient's lower limbs to collect electrical signals of lower limb muscle activity during rehabilitation training.

[0015] Lower limb activity status data is obtained by acquiring lower limb muscle electrical signals; the lower limb activity status data includes movement posture, gait changes, muscle force distribution, activity duration, and timed training data of the activity items.

[0016] Based on the lower limb activity state data, the change nodes of the lower limb activity state data in rehabilitation training are obtained and the change nodes are marked to obtain a state memory point dataset.

[0017] Based on the state memory point dataset, the changing trends of multiple rehabilitation training sessions can be obtained;

[0018] Based on the aforementioned trends, the current rehabilitation training program will be adjusted to ensure that patients receive personalized rehabilitation training during the rehabilitation process.

[0019] Furthermore, based on the aforementioned state memory point dataset, rehabilitation training information data is obtained through a visual interface to enable collaborative interaction between patients and doctors, including:

[0020] The state memory point dataset is transmitted via Converted into a visual and intuitive interface, where M(t) is the visually displayed state value, and S... i (t) represents the state data of the i-th memory point at time t, w i Here, is the weighting coefficient, n is the total number of memory points, and E(t) is the state memory point dataset;

[0021] The electromyographic signal data in the state memory point dataset is converted into a spectrogram through time-frequency domain analysis, and the power spectral density features are extracted from the spectrogram to obtain the power spectral density features.

[0022] The visual interface and power spectral density features are stored according to time series, and a multi-dimensional data query interface is set up.

[0023] The multidimensional data query interface is used to achieve synchronized data querying between the patient and doctor ends through a cloud computing platform, and the data synchronization delay time is... satisfy: To ensure a real-time interactive experience.

[0024] Furthermore, the rehabilitation training information data includes:

[0025] The patient's lower limb range of motion, frequency, bending angle limits, landing force, and weight-bearing force;

[0026] Acquire the range of motion data of the patient's lower limbs and set the range of motion threshold A. max When the range of motion A exceeds the threshold, an activity range alarm Alert(A) is triggered to assess the range of motion for the patient's rehabilitation training.

[0027] Monitor the movement frequency of the patient's lower limbs and set a movement frequency threshold f. max When the exercise frequency f exceeds the threshold, an exercise frequency alarm Alert(f) is triggered to analyze the patient's rehabilitation training exercise frequency.

[0028] Record the bending angle limit of the patient's lower limb and set the bending angle limit θ. max When the bending angle θ exceeds the threshold, a bending angle alarm Alert(θ) is triggered to obtain the patient's recovery status.

[0029] Measure the impact force of the patient's lower limbs upon landing and set a landing force threshold F. max When the landing force F exceeds the threshold, a landing force alarm Alert(F) is triggered to assess the stability of the patient's gait during rehabilitation training.

[0030] Analyze the weight distribution of the patient's lower limbs and set a weight distribution threshold W. max When the force of gravity W exceeds the threshold, the force distribution alarm Alert(W) is triggered to obtain the force situation of the patient's rehabilitation training.

[0031] Furthermore, based on the patient's lower limb range of motion data, time-series waveform graphs and histograms are provided, and a threshold warning progress bar is designed to display A and A. max The degree of proximity is used to obtain the display accuracy of optimized activity amplitude data;

[0032] Spectral analysis was performed on the patient's lower limb movement frequency to show the dominant frequency components and their comparison with the recommended frequency range. Color coding was used to identify different frequency ranges to obtain frequency stability indicators.

[0033] Based on the lower limb flexion angle limit, the joint range of motion is set, and the joint angle changes are displayed in real time through 3D skeletal animation, with safe activity areas marked. θ =∑(n=1+k)(1-(θ(t)-θ) target (t)) / θ min / θ max ) / m; to obtain the angle control accuracy; where, P θ To ensure angle control accuracy, θ(t) represents the real-time angle. target (t) represents the target angle, m represents the number of sampling points within the evaluation time period, and θ min θ represents the minimum bending angle of the lower limb. max This represents the maximum bending angle of the lower limbs.

[0034] Based on the landing force data, a plantar pressure distribution heatmap is generated, and the force changes within the gait cycle are displayed in combination with the time dimension to obtain the left-right balance pressure index.

[0035] Based on gravity distribution W data, a dynamic visualization of the center of gravity trajectory is constructed to display the patient's center of gravity shift process, through S... W =1-∑(n=1+k)(W(t)-W avg ) / (n·W max ), to obtain the center of gravity stability index; where S W W(t) is the index of center of gravity stability, and W(t) is the distribution of gravity force at time t. avg To achieve an average distribution of gravitational force, W max This is for data on the maximum distribution of gravity.

[0036] Furthermore, based on the feedback rehabilitation training information data, the patient's lower limb muscle electrical signal status rehabilitation progress is assessed, and the rehabilitation training plan is adjusted in a timely manner, including:

[0037] Based on the feedback of rehabilitation training information data, an assessment mechanism for the status of lower limb muscle electrical signals of patients was established to obtain a comprehensive rehabilitation status score in order to assess the patient's rehabilitation status.

[0038] Based on the comprehensive rehabilitation status score, an individual rehabilitation prediction model is established to predict the rehabilitation status at a future time t+Δt; and a dynamic rehabilitation objective function is set to adaptively adjust the rehabilitation training program as the rehabilitation status progresses.

[0039] Based on the adaptive adjustment of rehabilitation training programs, a rehabilitation training program adjustment strategy is generated through a rehabilitation difference comparison algorithm; and a fuzzy logic control system is used to finely adjust the training intensity to obtain a personalized rehabilitation training program.

[0040] Furthermore, when abnormal situations arise during the adjustment of the rehabilitation training program, a rapid response and immediate suspension of training are initiated to output abnormal warning information, including:

[0041] A multi-layered anomaly detection mechanism is established based on the adjustment strategy of the rehabilitation training program, and an abnormal state discrimination function is constructed to obtain the anomaly discrimination result of the adjustment strategy.

[0042] Based on the anomaly detection results of the adjustment strategy, a set of safety thresholds for key physiological parameters, S = s1, s2, ..., s, is set. i Real-time monitoring of physiological parameters x i With safety threshold s i The relationship between the real-time monitored physiological parameters x is used to determine the relationship between the parameters x and x. i Does it exceed the safety threshold s? i ;

[0043] Based on physiological parameter x i With safety threshold s i Based on the relationship, an anomaly urgency assessment function is constructed to evaluate the anomaly urgency level; and an emergency stop control strategy for the training system is set, with a stop response time τ. s top satisfies: τ s top < 100ms; to ensure patient safety;

[0044] A hierarchical early warning information mechanism is established based on the anomaly urgency assessment function, and different levels of early warning are output according to the anomaly urgency assessment function to output different early warning information.

[0045] Furthermore, it also includes:

[0046] Lower limb skeletal muscle sensing module: Acquires electromyographic signals through a high-precision surface electromyography sensor array, and uses differential amplification technology to suppress common-mode interference and bio-potential acquisition algorithm to obtain lower limb skeletal muscle sensing signals;

[0047] Human-computer interaction module: Provides an intuitive graphical user interface, adopts adaptive layout technology to adapt to different terminal devices, realizes data visualization and responds to manual adjustment of various rehabilitation training parameters, including real-time waveform graphs, muscle activity heat maps and rehabilitation progress dashboards;

[0048] Adaptive contraction dynamics module: Based on the Hill muscle model, analyze muscle contraction characteristics and establish a force-velocity relationship framework to achieve automatic identification and classification of muscle contraction patterns;

[0049] Muscle fatigue analysis module: It assesses the degree of muscle fatigue by monitoring the change in the median frequency of the electromyographic signal spectrum, analyzes the degree of muscle fatigue corresponding to the muscle fatigue index, and establishes a fatigue recovery model to predict muscle recovery time for the next rehabilitation training operation.

[0050] Rehabilitation Phase Training Module: Automatically divides training phases according to rehabilitation progress, adopts a step-by-step intensity increase strategy, sets multiple difficulty levels to obtain different advanced rehabilitation training content, until the patient's lower limb rehabilitation meets physiological adaptation.

[0051] Emergency pause module: Employs multi-sensor fusion technology to detect abnormal states and obtains the probability of anomalies based on a Bayesian decision model; when the probability of anomalies exceeds a set safety threshold s... i When this occurs, the motor emergency stop protection mechanism is triggered;

[0052] Training and assessment module: Comprehensively analyzes training data to generate multi-dimensional rehabilitation assessment reports, obtains a comprehensive rehabilitation effect score through quantitative rehabilitation effects, and reveals rehabilitation trends through time-series comparisons, providing data support for doctors' decision-making.

[0053] Furthermore, the human-computer interaction module also includes:

[0054] Patient end and doctor end;

[0055] The patient terminal connects to the system terminal via Bluetooth and executes the established rehabilitation training plan according to the rehabilitation training suggestions set by the doctor terminal. The plan includes set angles, speeds, times, and torque protection. The patient terminal then uploads the training data generated during the rehabilitation training process to the cloud database in real time, and provides feedback on the training status or asks questions to the doctor terminal by uploading text, pictures, and videos, so as to obtain professional rehabilitation guidance and answers to questions from the doctor terminal.

[0056] The doctor's terminal can view the rehabilitation training history data of the patient's terminal; the history data includes rehabilitation training information data such as text, pictures and videos uploaded by the patient's terminal, as well as electrical signal feedback images of lower limb muscle rehabilitation training; thereby the doctor's terminal can provide rehabilitation training suggestions to the patient's terminal and adjust the rehabilitation training plan in a timely manner;

[0057] The patient and doctor terminals enable efficient interaction between patients and doctors in a home-based rehabilitation environment.

[0058] The beneficial effects of this invention are as follows: This invention monitors the electromuscular signal status data of the patient's lower limbs in real time during rehabilitation training, marking at least one node of change in the electromuscular signal status data as a memory point to obtain a state memory point dataset. Based on the state memory point dataset, a visual interface is used to obtain feedback rehabilitation training information data, enabling collaborative interaction between the patient and doctor. Based on the feedback rehabilitation training information data, the rehabilitation status of the patient's lower limb muscle electromuscular signal status is assessed, and the rehabilitation training plan is adjusted promptly. When abnormalities occur during the adjustment of the rehabilitation training plan, a rapid response is initiated and training is immediately paused, outputting an abnormality warning message. This collectively achieves the goal of enabling patients to safely and effectively conduct lower limb rehabilitation training at home, solving many problems currently faced by patients undergoing rehabilitation training at home: shortening hospital stays, reducing medical costs, reducing the occupation of medical resources, improving patient comfort, and preventing adverse consequences such as soft tissue adhesions and joint stiffness caused by improper training, ultimately promoting better recovery of limb function. Through this intelligent system, digital sensing technology is integrated into traditional mechanical structures, and a remote control program is developed to realize the concept of smart healthcare services with online interaction between doctors and patients, providing a more convenient and efficient solution for lower limb rehabilitation training. Attached Figure Description

[0059] Figure 1 This is a schematic diagram of a method flow for an intelligent lower limb rehabilitation training and interaction system provided by an embodiment of the present invention;

[0060] Figure 2 This is a schematic diagram of the process of an intelligent lower limb rehabilitation training and interaction system provided by an embodiment of the present invention;

[0061] Figure 3 This is a schematic diagram illustrating the main functions of the WeChat mini-program for doctors provided in an embodiment of the present invention;

[0062] Figure 4 This is a schematic diagram illustrating the main functions of the patient-side WeChat mini-program provided in an embodiment of the present invention.

[0063] Figure 5 This is a schematic diagram of doctor-patient interaction provided in an embodiment of the present invention. Detailed Implementation

[0064] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, some features described in the examples may be combined in other examples.

[0065] like Figure 1 and Figure 2 As shown, at least one embodiment of the present invention discloses an intelligent lower limb rehabilitation training and interaction system, comprising:

[0066] Step 100: Monitor the electrical signal status data of the patient's lower limb muscles in real time during rehabilitation training, and mark at least one node of change in the electrical signal status data as a memory point to obtain a state memory point dataset;

[0067] Step 200: Based on the state memory point dataset, obtain the rehabilitation training information data fed back from the state memory point dataset through a visual interface, so as to enable patients and doctors to interact collaboratively;

[0068] Step 300: Based on the feedback rehabilitation training information data, determine the patient's lower limb muscle electrical signal status and rehabilitation training plan in a timely manner;

[0069] Step 400: When an abnormal situation occurs during the adjustment of the rehabilitation training program, respond quickly and immediately suspend the training to output an abnormal warning message.

[0070] Specifically, before manipulating the data, preprocessing is necessary. This involves cleaning, organizing, and transforming the raw data to facilitate subsequent analysis, modeling, or machine learning. The goal is to remove noise and anomalies, ensure data format consistency, make the data more suitable for analysis, and transform it into a unified format to ensure it can be correctly processed by algorithms.

[0071] The preprocessed data is then normalized, specifically by scaling the data proportionally to a specified range. The purpose is to eliminate the influence of different units and value ranges on model training and prediction, ensuring that each feature contributes to the model with similar importance. Common normalization methods include: min-max normalization, which scales the data proportionally to a specified interval, typically [0,1]; Z-Score standardization, which converts the data into normally distributed data with a mean of 0 and a standard deviation of 1; maximum absolute value normalization, which scales the data to the interval [-1,1]; and Robust Scaler, which scales based on the median and quartiles of the data and is robust to outliers.

[0072] In this embodiment of the invention, step 100 involves real-time monitoring of the patient's lower limb muscle electrical signals using sensors. This electrical signal data typically comes from surface electromyography (EMG) or electrophysiological monitoring techniques. By analyzing changes in the muscle electrical signals, the system can identify important change nodes, which usually represent muscle fatigue, abnormal contraction, or rehabilitation progress. These change nodes are marked as memory points, used for subsequent dataset construction and training evaluation. This accurately records the changes in the patient's muscle state during rehabilitation, helping medical personnel understand the patient's progress. By marking key change nodes, precise evidence can be provided for subsequent feedback and adjustments, avoiding missing important rehabilitation opportunities.

[0073] Step 200 uses the collected state memory point data to generate feedback information, which is displayed on a visualization interface for both patients and doctors. This interface may include charts, trend analyses, and data points to help doctors understand the patient's recovery status in real time. Through feedback, patients can see their recovery progress, while doctors can adjust training plans based on this feedback, thus providing intuitive and real-time recovery information, enabling patients to understand their progress and boosting their confidence. Doctors can obtain detailed feedback data, facilitating more accurate and timely recovery decisions. This enables collaborative interaction between patients and doctors, promoting communication and joint development of recovery plans.

[0074] In step 300, the doctor assesses the patient's lower limb muscle rehabilitation progress based on feedback data (such as electromyographic signals and trends) obtained from the visualization interface. When the feedback information indicates that the patient's rehabilitation status is not as expected, the doctor can adjust the training intensity, frequency, or method in a timely manner. This process analyzes signal trends through intelligent algorithms to determine whether the predetermined rehabilitation goals have been achieved, thereby realizing a personalized and dynamically adjusted rehabilitation training plan that is more adaptable to the patient's real-time needs. It can maximize rehabilitation effects, avoid overtraining or undertraining, improve the efficiency of rehabilitation training, and ensure that the patient trains under optimal conditions.

[0075] If any abnormalities occur during the adjustment of the rehabilitation training program in step 400 (such as abnormal electromyographic signals, patient pain or fatigue), the system will automatically identify these abnormalities and trigger an alarm mechanism. The system will immediately suspend training and output an abnormality warning message to notify the doctor or rehabilitation personnel to intervene. Abnormality detection may involve real-time monitoring of electromyographic signals, patient physiological feedback, and environmental changes, thereby promptly detecting potential risks or problems and preventing further injury or discomfort. This ensures patient safety during rehabilitation, reduces injuries caused by overtraining or incorrect training methods, improves the safety and effectiveness of the rehabilitation process, and avoids treatment interruptions due to unforeseen circumstances.

[0076] In a preferred embodiment of the present invention, the training time and training cycle are adjusted in real time according to the patient's muscle fatigue level for different rehabilitation training programs, so as to avoid excessive strain on intensity and difficulty.

[0077] Based on the patient's recovery progress, the training process is divided into different difficulty stages, and dynamically adjusted according to the patient's recovery status; and the training difficulty is adjusted individually based on the patient's real-time feedback data.

[0078] By analyzing data on patients' activity status, muscle fatigue, and rehabilitation during rehabilitation training, we can assess their lower limb motor function, muscle recovery, and training progress.

[0079] In this embodiment of the invention, training time and cycle are adjusted in real time according to the degree of muscle fatigue: a surface electromyography bioelectric signal monitoring device is used to collect muscle electrical signal data of patients during training in real time; muscle fatigue indicators such as median frequency decrease, root mean square value change, and muscle contraction pattern are extracted through signal frequency, amplitude and other features; when the muscle fatigue level is detected to exceed the set threshold, the system will automatically adjust.

[0080] Real-time adjustment of training time and cycle: When muscle fatigue is high, the system will shorten the single training time, extend the rest time, or adjust the training interval cycle; when muscle fatigue is low, the training time or training frequency can be appropriately increased to maintain rehabilitation progress; avoid overtraining and reduce the risk of muscle strain or secondary injury; optimize the training rhythm to ensure that patients are rehabilitating within an appropriate intensity range and improve training effectiveness; improve rehabilitation efficiency by precisely adjusting in real time so that patients can train in the best condition and shorten the rehabilitation cycle.

[0081] The training difficulty is adjusted in stages according to the rehabilitation progress: combining the patient's historical training data, the indicators such as muscle strength, range of motion (ROM), and endurance are analyzed to judge the rehabilitation progress; motion sensors, electromyography signal analysis and AI algorithms are used to identify the improvement of the patient's motor ability; different rehabilitation training stages such as primary, intermediate and advanced are set, and each stage corresponds to different training intensities, movement complexity and duration.

[0082] Beginner stage: Simple isometric contraction training (static support), low resistance and low intensity training.

[0083] Intermediate stage: Gradually introduce isotonic contraction training (joint mobility training) and light-load resistance training.

[0084] Advanced stage: Increase dynamic load and complex functional training, such as gait training and balance training.

[0085] Dynamically adjust training programs: By combining muscle fatigue data and patient subjective feedback (such as pain scores), the system automatically adjusts the training difficulty; when a patient adapts well to a certain stage, the system can advance to the next difficulty level ahead of time; if the patient recovers slowly, the current stage is appropriately extended; the system progresses gradually, reducing the risk of injury caused by excessively difficult training; dynamic adjustments ensure that the training intensity matches the patient's recovery level, improving training effectiveness; and personalized rehabilitation plans enhance patient participation and compliance.

[0086] By combining real-time patient feedback data, the system personalizes and adjusts training difficulty: Sensors collect activity status data (gait, joint angles), muscle fatigue data (EMG), and rehabilitation progress data (ROM, muscle strength); combined with subjective patient feedback (such as pain scores and fatigue level questionnaires), a comprehensive data analysis is formed. The system intelligently adjusts training difficulty: through data analysis, the system automatically adjusts training content; when muscle fatigue is detected to be too rapid, the training intensity is reduced or rest time is increased; when the patient reports increased pain, the training difficulty is reduced or the training method is adjusted (such as reducing load or switching to passive training); when the patient is progressing well, the training difficulty is increased, the training load is increased, or the training time is extended; this achieves personalized rehabilitation training, improves training adaptability, reduces patient discomfort, improves training compliance, and makes patients more willing to adhere to the rehabilitation plan; and precisely controls training difficulty, improving rehabilitation effects and shortening rehabilitation time.

[0087] This approach assesses patients' motor function, muscle recovery, and training progress through data analysis: Gait data, electromyography (EMG) data, and joint range of motion data are collected in real time using wearable devices (electromyography sensors, inertial measurement units, IMUs, etc.); machine learning or statistical analysis methods are used to analyze data trends and assess the patient's muscle function recovery; motor function assessment includes calculating stride length, stride speed, and gait symmetry to evaluate lower limb motor function; muscle recruitment patterns are analyzed using EMG signals to determine if muscles have recovered to normal functional levels; combined with training data, the approach predicts the patient's rehabilitation progress and adjusts training goals; accurate rehabilitation assessment reports are provided to help doctors develop more scientific rehabilitation plans; abnormal motor function is automatically detected to identify potential problems and prevent secondary injuries; and training programs are optimized based on data analysis to improve the scientific rigor and accuracy of rehabilitation training.

[0088] In a preferred embodiment of the present invention, step 100 involves real-time monitoring of the activity state of the lower limb muscles of the training patient, marking at least one node of activity state change as a memory point to obtain a state memory point dataset, including:

[0089] Step 101: Wearable sensors are attached to the skin surface of the patient's lower limbs to collect electrical signals of lower limb muscle activity during rehabilitation training.

[0090] Step 102: Obtain lower limb activity status data by acquiring lower limb muscle electrical signals; the lower limb activity status data includes movement posture, gait changes, muscle force distribution, activity duration, and timed training data of the activity items;

[0091] Step 103: Based on the lower limb activity state data, obtain the change nodes of the lower limb activity state data in rehabilitation training, and mark the change nodes to obtain a state memory point dataset;

[0092] Step 104: Based on the state memory point dataset, obtain the changing trends of multiple rehabilitation training sessions;

[0093] Step 105: Adjust the current rehabilitation training program according to the changing trend so that the patient receives a personalized rehabilitation training program during the rehabilitation process.

[0094] In this embodiment of the invention, step 101 involves attaching a wearable sensor to the surface of the patient's lower limb. The sensor can monitor and collect electrical activity signals of the patient's lower limb muscles in real time during rehabilitation training. These electrical signals are typically related to the contraction and relaxation processes of muscles and can reflect the intensity and state of muscle activity. Real-time acquisition of patient muscle activity data provides an accurate biosignal basis for subsequent analysis. The procedure is non-invasive, offering high patient comfort and ease of long-term use. It provides accurate muscle activity data, which helps in a comprehensive assessment of the patient's muscle function.

[0095] Step 102 involves processing and analyzing the lower limb muscle electrical signals collected by sensors to convert them into lower limb activity status data. Specific data includes: movement posture (e.g., standing, walking, squatting, etc.); gait changes (frequency, stride length, and gait consistency); muscle force distribution (force changes in different muscle groups during movement); activity duration (duration of each activity during training); and timed training data (execution status of each training exercise). This provides detailed and comprehensive activity status data, offering a basis for assessing rehabilitation progress. Data-driven movement status data helps doctors or therapists objectively judge the patient's rehabilitation status. By subdividing the data into different training parameters, the patient's rehabilitation process can be comprehensively monitored.

[0096] Step 103, based on the activity status data obtained in Step 102, analyzes the changes in the data and identifies change nodes in lower limb activity—that is, those moments that reflect key changes in the patient's rehabilitation process. For example, moments of change in gait or significant changes in muscle strength are marked as state memory points, forming a dataset. This helps track each important change node in the rehabilitation process, which is key to individualized rehabilitation; it provides clear markers for doctors or therapists to facilitate the assessment of rehabilitation effects and changes in the patient's condition; and through state memory points, patients can better understand their rehabilitation progress and make corresponding adjustments.

[0097] Step 104 analyzes the changing trends in multiple rehabilitation training sessions based on the information in the state memory point dataset. These trends may include improvements in the patient's muscle activity, changes in gait, and even the degree of fatigue during training. By comparing historical data, it analyzes progress, stagnation, or regression during the rehabilitation process; provides a dynamic assessment of rehabilitation effectiveness to help determine the effectiveness of rehabilitation training; by observing trends, it can identify potential bottlenecks or problems in the patient's rehabilitation process, allowing for targeted adjustments; and provides a quantitative and objective evaluation standard for the patient's rehabilitation progress, avoiding errors caused by relying solely on subjective judgment.

[0098] Step 105 involves adjusting the current rehabilitation training program based on the analyzed trend data. This adjustment can optimize training intensity, the order of training programs, or the type of exercise. The goal is to provide a more personalized and precise training program based on the patient's rehabilitation progress. Personalized rehabilitation programs can maximize training effectiveness and accelerate the patient's recovery. They can also avoid overtraining or undertraining, ensuring a safe and effective rehabilitation process. Furthermore, they can enhance the patient's treatment experience, making them feel that the treatment process is more tailored to their individual needs, thereby increasing their trust and compliance.

[0099] In a preferred embodiment of the present invention, step 200, based on the state memory point dataset, obtains rehabilitation training information data fed back from the state memory point dataset through a visual interface to enable collaborative interaction between the patient and the doctor, including:

[0100] Step 201, transfer the state memory point dataset through... Converted into a visual and intuitive interface, where M(t) is the visually displayed state value, and S... i (t) represents the state data of the i-th memory point at time t, w i Here, is the weighting coefficient, n is the total number of memory points, and E(t) is the state memory point dataset;

[0101] Step 202: Convert the electromyographic signal data in the state memory point dataset into a spectrogram through time-frequency domain analysis, and extract the power spectral density features from the spectrogram to obtain the power spectral density features;

[0102] Step 203: Store the visualization interface and power spectral density features according to time series, and set up a multi-dimensional data query interface;

[0103] Step 204: The multidimensional data query interface is used to achieve synchronized data query between the patient and doctor ends through a cloud computing platform. The data synchronization delay time... satisfy: To ensure a real-time interactive experience.

[0104] Specifically, the electromyographic signal data in the state memory point dataset is converted into a spectrogram through time-frequency domain analysis, using P(f) = (FFT(E(t))). 2 Feature extraction is performed; where P(f) is the power spectral density, FFT represents the fast Fourier transform, and f is the frequency;

[0105] The state memory point dataset is stored in time series, and a multi-dimensional data query interface is set up. The multi-dimensional data query includes: Q(t1,t2,p)(E(t))=S i (t)|t1≤t i ≤t2,p∈P i Where Q is the query result set, t1 and t2 are the time ranges, p is the query parameter, and P is the query parameter. i Let i be the parameter set of the i-th memory point;

[0106] Data synchronization between the patient and doctor terminals is achieved through a cloud computing platform, and the data synchronization delay time... satisfy: To ensure a real-time interactive experience; set an alert threshold T in the visual interface, and provide a visual alert when the monitored data exceeds the threshold: Si (t)>T.

[0107] In this embodiment of the invention, step 201 transforms the patient's state memory point dataset (such as gait, muscle activity, etc.) into a visual interface. This interface displays the state data of the i-th memory point at each time point (time t) using graphs, charts, and other formats. Furthermore, weighting coefficients are added to the state data to better represent the importance of different state memory points in the overall rehabilitation process. By transforming the data into a graphical interface, doctors and patients can more intuitively understand and evaluate various indicators during the rehabilitation process. The visual interface allows doctors to quickly browse the data, thereby saving time and improving diagnostic efficiency. The importance of each memory point can be calibrated by setting weighting coefficients, facilitating the customization of more precise treatment plans for patients.

[0108] Step 202 involves time-frequency domain analysis of the electromyographic (EMG) signals in the state memory point dataset. Time-frequency domain analysis transforms the EMG signals from the time domain to the frequency domain, generating a spectrogram. Then, by extracting the power spectral density (PSD) features from the PSD, the distribution information of the signal at different frequencies is obtained. These features help analyze the intensity and patterns of muscle activity; through time-frequency domain analysis and PSD extraction, details of different frequency bands in the EMG signals can be captured, which is crucial for a deeper understanding of muscle activity patterns. This spectrogram and PSD analysis provides physicians with effective tools to monitor patients' muscle function and fatigue levels, helping to quantitatively assess rehabilitation progress. Based on the spectral features, rehabilitation training programs can be adjusted more scientifically to better meet the individualized needs of patients.

[0109] Step 203 involves storing the visual interface and power spectral density characteristic data separately according to time series, and establishing a multi-dimensional data query interface. This interface allows doctors and patients to query based on different data dimensions (such as time, muscle activity, training programs, etc.), facilitating comprehensive and detailed analysis of rehabilitation training. The time series data storage method ensures the timeliness and sequence of the data, facilitating the query and comparison of historical data. Through the multi-dimensional data query interface, doctors and patients can query specific data dimensions according to different needs, such as the muscle activity state at a certain moment, the effect of a certain training program, etc., providing more diverse analytical perspectives. This enables the system to provide highly customized data query functions to adapt to the needs of different users.

[0110] Step 204 enables data synchronization and querying between the patient and doctor's devices via a cloud computing platform. Through the cloud platform, doctors and patients can access and query data in real time, while the data synchronization latency must be controlled within a reasonable range to ensure a real-time interactive experience. The cloud computing platform allows patients and doctors to query the latest rehabilitation data anytime, anywhere, enabling real-time interaction. This reduces the limitations of physical distance and time; data is stored in the cloud and supports multi-device access, allowing doctors and patients to view, discuss, and adjust treatment plans from different locations; doctors can promptly obtain patient rehabilitation data and provide rapid feedback to adjust treatment plans, ensuring maximum rehabilitation effectiveness.

[0111] In a preferred embodiment of the present invention, the rehabilitation training information data includes:

[0112] The patient's lower limb range of motion, frequency, bending angle limits, landing force, and weight-bearing force;

[0113] Acquire the range of motion data of the patient's lower limbs and set the range of motion threshold A. max When the range of motion A exceeds the threshold, an activity range alarm Alert(A) is triggered to assess the range of motion for the patient's rehabilitation training.

[0114] Monitor the movement frequency of the patient's lower limbs and set a movement frequency threshold f. max When the exercise frequency f exceeds the threshold, an exercise frequency alarm Alert(f) is triggered to analyze the patient's rehabilitation training exercise frequency.

[0115] Record the bending angle limit of the patient's lower limb and set the bending angle limit θ. max When the bending angle θ exceeds the threshold, a bending angle alarm Alert(θ) is triggered to obtain the patient's recovery status.

[0116] Measure the impact force of the patient's lower limbs upon landing and set a landing force threshold F. max When the landing force F exceeds the threshold, a landing force alarm Alert(F) is triggered to assess the stability of the patient's gait during rehabilitation training.

[0117] Analyze the weight distribution of the patient's lower limbs and set a weight distribution threshold W. max When the force of gravity W exceeds the threshold, the force distribution alarm Alert(W) is triggered to obtain the force situation of the patient's rehabilitation training.

[0118] Specifically, if excessive exercise or too rapid recovery leads to overload on joints or muscles, an alarm can be set at the upper or lower limit.

[0119] In the initial stage of rehabilitation, the exercise frequency can be set to 2-3 times per day, gradually increasing. The exercise frequency can be set as the number of times within a certain period of time (such as 30 gait training sessions per hour); the alarm will be triggered when the exercise frequency exceeds the set threshold (such as more than 50 times per hour), indicating that there may be a risk of overtraining or inappropriate exercise rhythm.

[0120] The normal range of motion for the knee joint is 0° (extension) to 135° (flexion), but for patients undergoing rehabilitation, the range of motion is 20° to 90°. When the bending angle exceeds the set threshold of less than 20° or greater than 90°, it indicates a risk of excessive bending or excessive joint load, and an alarm should be triggered.

[0121] The impact force of a normal adult's gait is approximately 1.2 to 1.5 times their body weight. For rehabilitation patients, the threshold for landing force is set to be 1.5 times their body weight, which is below the normal range. When the landing force exceeds the set threshold (e.g., more than twice the body weight), an alarm may indicate that the patient is overloading themselves while walking or standing and needs to adjust their gait or undergo gait training.

[0122] In a normal gait, weight distribution is usually quite even, with the force generally distributed fairly evenly between both legs. However, for rehabilitation patients, there is a tendency for one leg to bear more weight. The threshold can be set based on the patient's weight and rehabilitation needs: if the weight distribution on one side exceeds 60% (normal value is 50%), it may indicate uneven weight distribution; if the weight distribution on one side exceeds the threshold (e.g., greater than 60% on one side), it may indicate gait instability or a risk of body shift, requiring an alarm to be triggered for adjustment.

[0123] In this embodiment of the invention, the range of motion of the patient's lower limbs is monitored, and a range of motion threshold is set: the range of motion (i.e., the range of movement of the lower limbs in each direction) is calculated by real-time monitoring of the patient's lower limb movement trajectory; a range of motion threshold is set (based on the individual patient's situation, such as the postoperative rehabilitation stage or the range of motion recommended by the doctor); when the patient's range of motion exceeds the set threshold, a range of motion alarm is triggered to indicate possible abnormal or overloaded movement; thereby preventing the patient from performing excessively large movements in the early stage of recovery and reducing the risk of secondary injury; quantifying the rehabilitation process: doctors can assess the patient's motor recovery based on changes in range of motion; personalized training adjustment: the training plan is dynamically adjusted based on the patient's range of motion feedback to improve rehabilitation efficiency.

[0124] Monitor the patient's lower limb movement frequency and set a movement frequency threshold: Measure the frequency of the patient's lower limb movements (i.e., the number of steps or swings per unit time) using an accelerometer, gyroscope, or gait analysis system; set a movement frequency threshold, triggering a movement frequency alarm when the movement frequency exceeds the safe range (too fast or too slow); excessively fast movement may lead to imbalance, while excessively slow movement may indicate slow rehabilitation progress; help the patient maintain a stable rhythm and improve training effectiveness; excessively fast or slow gait frequency may indicate gait imbalance, requiring adjustment of the training program.

[0125] Record the bending angle limits of the patient's lower limbs and set bending angle thresholds: Monitor the bending angles of the patient's knee, hip, and ankle joints in real time using joint angle sensors, inertial measurement units, or 3D motion capture technology; Set bending angle thresholds, and trigger a bending angle alarm when the patient's bending angle exceeds the set range (over-bending or under-bending); Prevent joint damage caused by excessive flexion or extension; Gradually increase the angle thresholds to assess the recovery of joint range of motion; Doctors can adjust training methods based on bending angle data to optimize rehabilitation effects.

[0126] Measure the landing force of the patient's lower limbs and set a landing force threshold: Measure the landing force (i.e., the impact force when the foot contacts the ground) of the patient during walking or training using a pressure sensor, pressure sensor, or plantar pressure distribution meter; set a landing force threshold, and trigger a landing force alarm if the landing force is too high or too low; excessive landing force may indicate gait instability; insufficient force may indicate gait weakness; reduce the risk of impact injury: avoid secondary injury caused by excessive landing force, especially for postoperative rehabilitation patients; improve gait stability: doctors can optimize the patient's landing pattern by adjusting gait training, making walking more stable.

[0127] Analyze the weight distribution of the patient's lower limbs and set weight distribution thresholds: Measure the force distribution in different areas of the foot (e.g., forefoot, heel, lateral, medial) during walking using a foot pressure plate, pressure sensing pad, or three-dimensional force sensor; set weight distribution thresholds, and trigger a weight distribution alarm if the force in a certain area is significantly abnormal (e.g., excessive force on one side); avoid gait imbalance caused by uneven force distribution; help patients adjust the force distribution in their feet to make their gait more natural; doctors can adjust training content based on the force distribution, such as strengthening certain muscle groups to improve gait structure.

[0128] In a preferred embodiment of the present invention, based on the patient's lower limb range of motion data, a time-series waveform graph and a histogram are provided, and a threshold warning progress bar is designed to display A and A. max The degree of proximity is used to obtain the display accuracy of optimized activity amplitude data;

[0129] Spectral analysis was performed on the patient's lower limb movement frequency to show the dominant frequency components and their comparison with the recommended frequency range. Color coding was used to identify different frequency ranges to obtain frequency stability indicators.

[0130] Based on the lower limb flexion angle limit, the joint range of motion is set, and the joint angle changes are displayed in real time through 3D skeletal animation, with safe activity areas marked. θ =∑(n=1+k)(1-(θ(t)-θ) target (t)) / θ min / θ max ) / m; to obtain the angle control accuracy; where, P θ To ensure angle control accuracy, θ(t) represents the real-time angle. target (t) represents the target angle, m represents the number of sampling points within the evaluation time period, and θ min θ represents the minimum bending angle of the lower limb. max This represents the maximum bending angle of the lower limbs.

[0131] Based on the landing force data, a plantar pressure distribution heatmap is generated, and the force changes within the gait cycle are displayed in combination with the time dimension to obtain the left-right balance pressure index.

[0132] Based on gravity distribution W data, a dynamic visualization of the center of gravity trajectory is constructed to display the patient's center of gravity shift process, through S... W =1-∑(n=1+k)(W(t)-W avg ) / (n·W max ), to obtain the center of gravity stability index; where S W W(t) is the index of center of gravity stability, and W(t) is the distribution of gravity force at time t. avg To achieve an average distribution of gravitational force, W max This is for data on the maximum distribution of gravity.

[0133] Specifically, based on the patient's lower limb range of motion data, a time-series waveform graph and histogram are provided, and a threshold warning progress bar is designed to display A and A. max The degree of proximity, and through A display (t)=LP(A(t),α A )=α A ·A(t)+(1-α A )·A display (t-1) is used to optimize the display accuracy of activity amplitude data; where LP is a low-pass filter function, α A A is the adaptive smoothing coefficient. display (t) represents the displayed value, and A represents the patient's lower limb range of motion data. max This refers to the maximum range of motion of the patient's lower limbs;

[0134] Spectral analysis of the patient's lower limb motor frequencies was performed, displaying the dominant frequency components and comparing them with the recommended frequency range. Different frequency intervals were identified using color coding. To obtain frequency stability indicators; where S f As a frequency stability index, f i f is the frequency value of the sampling point. avg is the average frequency, k is the number of sampling points, f is the frequency of the patient's lower limb movements, and n is the total number of sampling points within the assessment period;

[0135] Based on landing force data, a plantar pressure distribution heatmap is generated, which, combined with the time dimension, displays the force changes within the gait cycle. This is achieved through B... F =1-F(F left -F right ) / (F left +F right ), to obtain the left-right balance index; where B F For the force balance index, F left and F right These represent the average landing force of the left and right feet, respectively, with F representing the landing force data.

[0136] In this embodiment of the invention, time-series waveforms and histograms of lower limb activity amplitude data are generated: Based on the patient's lower limb activity data, time-series waveforms are generated to display the trend of changes in the patient's activity amplitude over a certain period. Histograms are used to analyze the frequency distribution of activity amplitude, facilitating the observation of common ranges and their distribution characteristics. Time-series waveforms can track the patient's activity dynamics in real time, intuitively reflecting fluctuations in activity amplitude and helping to identify abnormalities or abnormal fluctuations. Histogram analysis provides data support for setting activity range and threshold warnings, thereby optimizing the display accuracy of activity amplitude data.

[0137] Spectral analysis of lower limb movement frequencies: Spectral analysis of the patient's lower limb movement frequencies converts the movement signal into frequency components, extracting the main frequency components and their relative intensities. Simultaneously, the actual frequency is compared with the recommended frequency range, using color coding to represent different frequency intervals. Spectral analysis reveals the frequency characteristics of the patient's movements, helping to monitor whether the movements conform to the recommended frequency range. The use of color coding not only enhances the visualization of the data but also quickly identifies abnormal frequencies or unstable movement patterns, providing a basis for obtaining frequency stability indicators.

[0138] Lower limb flexion angle joint range of motion and 3D skeletal animation display: By real-time acquisition of lower limb joint angle changes, a safe range of joint movement is set, and the movement process of the lower limb joints is displayed in real time using 3D skeletal animation. The animation marks the safe range of motion and tracks changes in joint angles in real time, helping doctors assess the accuracy of the patient's movements. The 3D animation clearly shows changes in joint angles, and color or graphic indicators show whether the patient is within the safe range of motion, thereby reducing the risk of joint injury. Precise angle control helps patients better understand their movement limits and optimize their activity plans.

[0139] Landing force data and plantar pressure distribution heatmap: By collecting plantar pressure data from patients during their gait cycle, a pressure distribution heatmap is generated. The heatmap shows the changes in force during the gait cycle, further analyzing whether the pressure distribution is balanced between the left and right feet. The plantar pressure distribution map can help analyze whether there are unbalanced or abnormal force points in the patient's gait, detect potential gait problems, and provide early warning of possible fall risks or gait abnormalities. At the same time, the pressure distribution heatmap is an effective assessment of balance pressure indicators, helping doctors determine the patient's gait status.

[0140] Dynamic visualization of center of gravity trajectory: By collecting and analyzing the patient's weight distribution data, a dynamic visualization of the center of gravity trajectory is constructed. This image shows the process of the patient's center of gravity shifting during movement, and is analyzed in conjunction with weight distribution data at different time dimensions. Visualizing the center of gravity trajectory helps doctors and patients observe the trajectory of changes in the center of gravity, revealing whether there is instability in the center of gravity shift. This indicator can serve as an important basis for assessing the patient's balance, helping to improve the stability of the patient's gait, especially during rehabilitation, and assisting in adjusting exercise programs to improve stability.

[0141] In a preferred embodiment of the present invention, step 300, based on the feedback rehabilitation training information data, determines the patient's lower limb muscle electrical signal status rehabilitation progress and adjusts the rehabilitation training plan accordingly, including:

[0142] Step 301: Based on the feedback rehabilitation training information data, establish a mechanism for assessing the status of the patient's lower limb muscle electrical signals and obtain a comprehensive rehabilitation status score to assess the patient's rehabilitation status.

[0143] Step 302: Establish an individual rehabilitation prediction model for patients based on the comprehensive rehabilitation status score to predict the rehabilitation status at future time t+Δt; and set a dynamic rehabilitation objective function to adaptively adjust the rehabilitation training program as the rehabilitation status progresses.

[0144] Step 303: Based on the adaptive adjustment of the rehabilitation training program, a rehabilitation training program adjustment strategy is generated through a rehabilitation difference comparison algorithm; and the training intensity is finely adjusted through a fuzzy logic control system to obtain a personalized rehabilitation training program.

[0145] Specifically, an assessment mechanism for the electromuscular signal status of patients' lower limbs is established, and the assessment model includes:

[0146] R = α·E + β·F + γ·A + δ·T + ε·C; where R is the comprehensive score of rehabilitation status, E is the electromyographic signal intensity coefficient, F is the muscle fatigue index, A is the completion rate of the range of motion, T is the training duration coefficient, C is the continuity index, and α, β, γ, δ and ε are the index weight coefficients respectively.

[0147] Establish individual patient rehabilitation prediction models using deep learning neural network algorithms:

[0148] P(t+Δt)=NN(H t H t -1,...,H t -k); where P(t+Δt) is the predicted recovery state at future time t+Δt, NN represents the neural network function, and H t Here, t represents the rehabilitation history data at time t, k represents the retrospective period, and Δt represents the time prediction interval.

[0149] Set a dynamic rehabilitation objective function that is adaptively adjusted as rehabilitation progresses: G(t)=G0·(1+λ·log(1+t / t0)); where G0 is the initial rehabilitation objective value, λ is the adjustment coefficient, t0 is the characteristic time scale, G(t) is the dynamic rehabilitation objective function, and log() is used to record information, debug programs, or monitor system status.

[0150] The rehabilitation training program adjustment strategy is generated based on the rehabilitation difference comparison algorithm: ΔS=η·(G(t)-R(t))·D(t); where ΔS is the adjustment amount of the training program, η is the adjustment gain coefficient, R(t) is the current rehabilitation score, and D(t) is the personalized adjustment coefficient.

[0151] The training intensity I is finely adjusted by a fuzzy logic control system: I = f(R, ΔR / Δt, ∫R·dt); where f is the fuzzy logic mapping function, R is the rehabilitation status score, ΔR / Δt is the rehabilitation rate, ∫R·dt is the cumulative rehabilitation amount, and I is the training intensity.

[0152] In this embodiment of the invention, step 301 involves collecting real-time electromyographic (EMG) signal data of the patient during rehabilitation training and combining it with other relevant physiological indicators to establish an assessment mechanism for the state of lower limb EMG signals. This data may include electromyography (EMG) signals, used to assess muscle activity and fatigue levels. By analyzing these signals, the system can generate a comprehensive rehabilitation status score for the patient, reflecting the patient's muscle recovery progress and rehabilitation effect. The EMG signal status assessment mechanism can track the patient's muscle recovery in real time during rehabilitation and can sensitively capture early signs such as muscle fatigue and functional decline. The comprehensive score provides doctors with a quantitative assessment method, objectively measuring the patient's rehabilitation status and providing data support for developing further training programs.

[0153] Step 302 establishes an individualized rehabilitation prediction model based on the patient's comprehensive rehabilitation status score, using machine learning, statistical analysis, or other prediction algorithms to predict the patient's rehabilitation status at a future point in time. This model can be trained based on the patient's specific data to improve prediction accuracy. Furthermore, by setting a dynamic rehabilitation objective function, the system can automatically adjust the objective based on changes in the patient's rehabilitation status, thereby adaptively adjusting the rehabilitation training plan. Through the individualized rehabilitation prediction model, the system can accurately predict the patient's future rehabilitation progress, helping doctors to assess rehabilitation effects in real time and formulate strategies for adjusting training goals. The dynamic objective function allows rehabilitation goals to be adjusted synchronously with the patient's actual recovery process, avoiding the limitations of fixed plans, effectively improving the effectiveness of personalized training, and reducing overtraining or undertraining.

[0154] Step 303: Based on the individual rehabilitation prediction model established in Step 302, the system uses a rehabilitation difference comparison algorithm to evaluate the effectiveness of the current training program according to the difference between the patient's actual rehabilitation status and the target. Based on the evaluation results, response adjustment strategies are generated. These strategies will guide how the training program is modified to better suit the patient's needs. Furthermore, through the fuzzy logic control system, the intensity, frequency, and duration of training can be finely adjusted to better match the individual needs of the patient; the rehabilitation difference comparison algorithm makes the adjustment of the training program more flexible and precise, avoiding a one-size-fits-all approach. The introduction of the fuzzy logic control system makes the adjustment of training intensity smoother and more intelligent, avoiding overly strenuous or overly lenient training arrangements. This fine adjustment can provide the most suitable training program based on the patient's real-time response and recovery status, thereby effectively improving rehabilitation outcomes and reducing unnecessary risks.

[0155] In a preferred embodiment of the present invention, when an abnormal situation occurs during the adjustment of the rehabilitation training program in step 400, a rapid response is initiated and training is immediately paused to output an abnormality warning message, including:

[0156] Step 401: Based on the rehabilitation training program adjustment strategy, establish a multi-layer anomaly detection mechanism and construct an anomaly state discrimination function to obtain the anomaly discrimination result of the adjustment strategy;

[0157] Step 402: Based on the anomaly detection results of the adjustment strategy, set a set of safety thresholds for key physiological parameters S = s1, s2, ..., s i Real-time monitoring of physiological parameters x i With safety threshold s i The relationship between the real-time monitored physiological parameters x is used to determine the relationship between the parameters x and x. i Does it exceed the safety threshold s? i ;

[0158] Step 403, based on physiological parameter x i With safety threshold s i Based on the relationship, an anomaly urgency assessment function is constructed to evaluate the anomaly urgency level; and an emergency stop control strategy for the training system is set, with a stop response time τ. s top satisfies: τ s top < 100ms; to ensure patient safety;

[0159] Step 404: Establish a hierarchical early warning information mechanism based on the anomaly urgency assessment function, and output different levels of early warning according to the anomaly urgency assessment function to output different early warning information.

[0160] Specifically, a multi-layered anomaly detection mechanism is established, and an anomaly state discrimination function is constructed:

[0161] A(x) = sgn(∑(i = 1 + m)v i ·f i (x)-θ); where A(x) is the anomaly detection result, f i (x) is the function for extracting the i-th anomaly feature, v i The corresponding weights are θ, the threshold parameter is sgn, the sign function is i, the total number of abnormal features is x, the input parameter vector is m, and the number of abnormal features is m.

[0162] Define a set of safety thresholds for key physiological parameters S = s1, s2, ..., s i Real-time monitoring parameter x i With threshold s i Relationship: Alert(x) i ,s i ) = True, if(x i -s i )>ε i ,False,otherwise;where,ε i To define the allowable deviation range, Alert(x) i si ) is a judgment function used to determine the real-time monitored physiological parameter x. i Does it exceed the safety threshold s? i True indicates that parameter x i Exceeding the safety threshold s i False indicates that the parameter x i At the safety threshold s i Within the range, otherwise means parameter x i Within the allowable deviation range, i.e., return False; construct the anomaly urgency assessment function: E(x) = maxw1·d(x1,S1),w2·d(x2,S2),...,w i ·d(x i ,S i ); where d(x) i S i ) represents the parameter vector x i With safety range S i The distance function is given by n, where n is the number of monitored parameters, max is the maximum value, and E(x) is the function for assessing the urgency of the anomaly.

[0163] Implement an emergency stop control strategy for the training system, with a stop response time τ. s top satisfies: τ s top < 100ms; ensure patient safety; establish a tiered early warning information system, outputting different levels of warnings according to the assessment function of the urgency of the abnormality: L = floor(E(x) / E max ·N); where L is the warning level (0 to N), floor is the floor function, and E max N represents the maximum urgency level, and N is the warning level number.

[0164] In this embodiment of the invention, step 401 establishes a multi-layered anomaly detection mechanism based on the rehabilitation training program adjustment strategy. The core of this mechanism is an anomaly state discriminant function, responsible for identifying various abnormal situations that occur during rehabilitation training. These anomalies may include excessive training intensity, inappropriate training methods, or maladaptation of the patient's physiological state. This discriminant function analyzes real-time monitored data (such as heart rate, blood pressure, and electromyography signals) to detect any abnormalities and determine if there are potential problems with the training program. The multi-layered anomaly detection mechanism can monitor the patient's rehabilitation training in real time and comprehensively. Through multi-level analysis of various physiological and training data, the system can promptly identify anomalies that may affect the rehabilitation effect, allowing for early adjustments and interventions to avoid potential health risks. The anomaly state discriminant function provides intelligent decision support for anomaly detection, ensuring the scientific validity and safety of the rehabilitation training program.

[0165] Step 402: Based on the abnormality detection results of the adjustment strategy in step 401, the system sets a set of safe thresholds for key physiological parameters. Key physiological parameters typically include heart rate, blood pressure, electromyography (EMG), and respiratory rate. These parameters are core indicators affecting patient safety and health during rehabilitation training. The system monitors the patient's physiological parameters in real time and compares them with the pre-set safe thresholds to determine if the current physiological state is within the normal range. This step, by setting safe thresholds and monitoring physiological parameters in real time, ensures that the patient will not experience abnormal physiological states exceeding safe limits during rehabilitation training. If any physiological parameter exceeds the set threshold, the system will trigger an alarm and intervene promptly to prevent discomfort caused by overtraining or other factors, ensuring the safety of the training process.

[0166] Step 403, based on real-time monitoring of the relationship between physiological parameters and safety thresholds, constructs an abnormality urgency assessment function to evaluate the urgency of the abnormality according to the degree to which physiological parameters deviate from the safety threshold. For example, a sharp increase in heart rate may be more urgent than a slight increase, and abnormal fluctuations in electromyographic signals may require more urgent treatment than mild fatigue. Based on the assessed urgency, the system will automatically activate the training system's emergency stop control strategy. This strategy stipulates that when danger occurs, the system must shut down within a set response time to ensure patient safety. By constructing the abnormality urgency assessment function, the system can intelligently judge and differentiate the severity of various abnormal situations, thereby providing more accurate and efficient decision support for emergency treatment. The emergency stop control strategy ensures that the system can respond quickly in emergency situations, avoiding unnecessary risks to patients due to delayed treatment, and greatly improving the safety of rehabilitation training.

[0167] Step 404: Based on the urgency assessment function from step 403, the system establishes a tiered early warning mechanism. This mechanism outputs different levels of early warning information according to the urgency of the abnormal situation. For example, minor physiological fluctuations may only trigger a low-level warning, while severe abnormalities will trigger a high-risk warning. The early warning information provides different feedback based on the urgency level, alerting trainers through visual and auditory means, or through digital means to patients or the medical team. The tiered early warning mechanism can provide responsive early warning information based on different emergency situations, enabling rehabilitation training administrators, doctors, or patients to clearly and accurately assess the current risk level. This tiered mechanism avoids information overload and ensures that in emergency situations, the most pressing issues are addressed quickly, thereby improving the efficiency of handling abnormalities and ensuring patient safety.

[0168] In a preferred embodiment of the present invention, it further includes:

[0169] Lower limb skeletal muscle sensing module: Acquires electromyographic signals through a high-precision surface electromyography sensor array, and uses differential amplification technology to suppress common-mode interference and bio-potential acquisition algorithm to obtain lower limb skeletal muscle sensing signals;

[0170] Human-computer interaction module: Provides an intuitive graphical user interface, adopts adaptive layout technology to adapt to different terminal devices, realizes data visualization and responds to manual adjustment of various rehabilitation training parameters, including real-time waveform graphs, muscle activity heat maps and rehabilitation progress dashboards;

[0171] Adaptive contraction dynamics module: Based on the Hill muscle model, analyze muscle contraction characteristics and establish a force-velocity relationship framework to achieve automatic identification and classification of muscle contraction patterns;

[0172] Muscle fatigue analysis module: It assesses the degree of muscle fatigue by monitoring the change in the median frequency of the electromyographic signal spectrum, analyzes the degree of muscle fatigue corresponding to the muscle fatigue index, and establishes a fatigue recovery model to predict muscle recovery time for the next rehabilitation training operation.

[0173] Rehabilitation Phase Training Module: Automatically divides training phases according to rehabilitation progress, adopts a step-by-step intensity increase strategy, sets multiple difficulty levels to obtain different advanced rehabilitation training content, until the patient's lower limb rehabilitation meets physiological adaptation.

[0174] Emergency pause module: Employs multi-sensor fusion technology to detect abnormal states and obtains the probability of anomalies based on a Bayesian decision model; when the probability of anomalies exceeds a set safety threshold s... i When this occurs, the motor emergency stop protection mechanism is triggered;

[0175] Training and assessment module: Comprehensively analyzes training data to generate multi-dimensional rehabilitation assessment reports, obtains a comprehensive rehabilitation effect score through quantitative rehabilitation effects, and reveals rehabilitation trends through time-series comparisons, providing data support for doctors' decision-making.

[0176] Specifically, the lower limb skeletal muscle sensing module: acquires electromyographic signals through a high-precision surface electromyography sensor array, with a sensor sampling frequency of no less than 1000Hz and a resolution of no less than 16 bits, and uses differential amplification technology to suppress common-mode interference, with a signal-to-noise ratio better than 60dB; this module uses a bio-potential acquisition algorithm.

[0177] Human-computer interaction module: Provides an intuitive graphical user interface, adopts adaptive layout technology to adapt to different terminal devices, and realizes data visualization display, including real-time waveform graphs, muscle activity heat maps and rehabilitation progress dashboards;

[0178] Adaptive contraction dynamics module: Based on the Hill muscle model, analyze muscle contraction characteristics and establish a force-velocity relationship equation: (F+a)·(v+b)=(F0+a)·b, where F is the force generated by the muscle, v is the contraction velocity, F0 is the maximum isometric contraction force, and a and b are characteristic parameters; realize automatic identification and classification of muscle contraction patterns;

[0179] Muscle fatigue analysis module: Assesses muscle fatigue level by monitoring changes in the median frequency (MDF) of electromyographic signals, using the fatigue index calculation formula: FI = 1 - (MDF) t / MDF0), where MDF0 is the initial median frequency, MDF t Let be the median frequency at time t, and FI be the fatigue index.

[0180] And based on the fatigue recovery model R(t)=1-e (-t / τ) Predict muscle recovery time, where R(t) is the recovery rate, t is the recovery time, τ is the recovery time constant, and e is the base of the natural logarithm;

[0181] Rehabilitation Phase Training Module: The training phases are automatically divided according to the rehabilitation progress. A step-by-step intensity increase strategy is adopted, and the difficulty coefficient is set as D = d0·(1+k·n), where d0 is the basic difficulty, k is the increment coefficient, and n is the phase number. The training content and intensity of each phase meet the principle of physiological adaptability.

[0182] Emergency Stop Module: Employs multi-sensor fusion technology to detect abnormal states and calculates the abnormal probability based on a Bayesian decision model: P(A|X)=P(X|A)·P(A) / P(X), where A represents the abnormal state, X is a sensor data vector, P(A|X) is the probability of abnormal state A occurring given sensor data vector X, P(X|A) is the probability of observing sensor data vector X under abnormal state A, P(A) is the prior probability of abnormal state A, and P(X) is the probability of observing sensor data vector X. When the abnormal probability exceeds a set safety threshold, the motor emergency stop protection mechanism is triggered, and the control system response time is less than 100ms.

[0183] Training Assessment Module: This module comprehensively analyzes training data to generate a multi-dimensional rehabilitation assessment report, employing a formula to quantify rehabilitation effectiveness. Among them, M g Let g be the measured value of the g-th rehabilitation indicator. 'b' represents the target value, 'b' represents the total number of rehabilitation indicators, and 'E' represents the comprehensive rehabilitation effect score. By comparing time series, we can reveal the rehabilitation trend and provide data support for doctors' decision-making.

[0184] In a preferred embodiment of the present invention, the human-computer interaction module further includes:

[0185] Patient end and doctor end;

[0186] The patient terminal connects to the system terminal via Bluetooth and executes the established rehabilitation training plan according to the rehabilitation training suggestions set by the doctor terminal. The plan includes set angles, speeds, times, and torque protection. The patient terminal then uploads the training data generated during the rehabilitation training process to the cloud database in real time, and provides feedback on the training status or asks questions to the doctor terminal by uploading text, pictures, and videos, so as to obtain professional rehabilitation guidance and answers to questions from the doctor terminal.

[0187] The doctor's terminal can view the rehabilitation training history data of the patient's terminal; the history data includes rehabilitation training information data such as text, pictures and videos uploaded by the patient's terminal, as well as electrical signal feedback images of lower limb muscle rehabilitation training; thereby the doctor's terminal can provide rehabilitation training suggestions to the patient's terminal and adjust the rehabilitation training plan in a timely manner;

[0188] The patient and doctor terminals enable efficient interaction between patients and doctors in a home-based rehabilitation environment.

[0189] like Figure 3 As shown in this embodiment of the invention, the main functions of the doctor's WeChat mini-program include login and registration, displaying patient information, viewing patient rehabilitation exercise information and records, viewing patient questions and question records, doctor messages / responding to patient questions, and setting phased / overall rehabilitation goals. Doctors can use the mini-program to view patients' personal information, mobile phone numbers, causes of illness, and consultation times, as well as relevant rehabilitation training data, such as maximum training angle, minimum training angle, and training duration. Furthermore, doctors can communicate with patients through text, images, and videos, answer patient questions, and develop corresponding training goals and plans based on the patient's rehabilitation progress.

[0190] like Figure 4 As shown, this mainly involves the core functions of the patient-side WeChat mini-program. The mini-program includes login / registration, reference tutorials, device connection, parameter settings, starting training, receiving training plans and messages from doctors, patient submission of supplementary materials, viewing training records, viewing question records and doctor replies, a personal center, and user reviews. Each function has corresponding sub-functions or steps; for example, the login / registration function includes new user registration and existing user login, the reference tutorial includes an introduction to the mini-program's usage and function descriptions, and the device connection includes connecting to the lower-level Bluetooth device.

[0191] like Figure 5The diagram illustrates the communication process between doctors and patients via a WeChat mini-program. First, the doctor's WeChat mini-program can view the patient's training data, develop a rehabilitation training plan, and return the data to the patient's WeChat mini-program, which then uploads and displays the data. Next, the doctor's WeChat mini-program answers the patient's questions and returns the answers to the patient's WeChat mini-program, which then uploads and displays the answers. Finally, the patient's WeChat mini-program asks questions to the doctor, who receives the questions, provides feedback, and then evaluates the service.

[0192] The embodiments of the present invention have been described above. However, the embodiments are not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make more equivalent embodiments under the guidance of the present embodiments, and all of them are within the protection scope of the present embodiments.

Claims

1. An intelligent lower limb rehabilitation training and interactive system, characterized in that, include: Real-time monitoring of the patient's lower limb muscle electrical signal status data during rehabilitation training; marking at least one node of change in the electrical signal status data as a memory point to obtain a status memory point dataset. Based on the state memory point dataset, rehabilitation training information data is obtained through a visual interface to enable collaborative interaction between patients and doctors. Based on the feedback of rehabilitation training information data, the patient's lower limb muscle electrical signal status rehabilitation progress is determined, and the rehabilitation training plan is adjusted in a timely manner. When abnormal situations occur during the adjustment of the rehabilitation training program, a rapid response should be initiated and training should be immediately suspended to issue an abnormal warning message.

2. The intelligent lower limb rehabilitation training and interaction system according to claim 1, characterized in that, The training time and cycle are adjusted in real time according to the patient's muscle fatigue level for different rehabilitation training programs in order to avoid excessive strain on intensity and difficulty; The training process is divided into different difficulty levels according to the patient's recovery progress, and is dynamically adjusted according to the patient's recovery status. In addition, the training difficulty is adjusted in a personalized manner based on the patient's real-time feedback data; By acquiring and analyzing patients' activity status data, muscle fatigue data, and rehabilitation data during rehabilitation training, we can assess patients' lower limb motor function, muscle recovery, and training progress.

3. The intelligent lower limb rehabilitation training and interaction system according to claim 2, characterized in that, Real-time monitoring of the lower limb muscle activity of training patients, marking at least one node of activity state change as a memory point to obtain a state memory point dataset, including: Wearable sensors are attached to the skin surface of the patient's lower limbs to collect electrical signals of lower limb muscle activity during rehabilitation training; the acquired lower limb muscle electrical signals are used to obtain lower limb activity status data; the lower limb activity status data includes movement posture, gait changes, muscle force distribution, activity duration, and timed training data of activity items; Based on the lower limb activity status data, the change nodes of the lower limb activity status data in rehabilitation training are obtained and the change nodes are marked to obtain a status memory point dataset; based on the status memory point dataset, the change trend of multiple rehabilitation training sessions is obtained. Based on the aforementioned trends, the current rehabilitation training program will be adjusted to ensure that patients receive personalized rehabilitation training during the rehabilitation process.

4. The intelligent lower limb rehabilitation training and interaction system according to claim 1, characterized in that, Based on the aforementioned state memory point dataset, rehabilitation training information data is obtained through a visual interface to enable collaborative interaction between patients and doctors, including: The state memory point dataset is transmitted via Converted into a visual and intuitive interface, where M(t) is the visually displayed state value, and S... i (t) represents the state data of the i-th memory point at time t, w i Here, represents the weighting coefficient, n represents the total number of memory points, and E(T) represents the state memory point dataset. The electromyographic signal data in the state memory point dataset is converted into a spectrogram through time-frequency domain analysis, and the power spectral density features are extracted from the spectrogram to obtain the power spectral density features. The visual interface and power spectral density features are stored according to time series, and a multi-dimensional data query interface is set up. The multidimensional data query interface is used to achieve synchronized data querying between the patient and doctor ends through a cloud computing platform, and the data synchronization delay time is... satisfy: To ensure a real-time interactive experience.

5. The intelligent lower limb rehabilitation training and interaction system according to claim 4, characterized in that, The rehabilitation training information data includes: The patient's lower limb range of motion, frequency, bending angle limits, landing force, and weight-bearing force; Acquire the range of motion data of the patient's lower limbs and set the range of motion threshold A. max When the range of motion A exceeds the threshold, an activity range alarm Alert(A) is triggered to assess the range of motion for the patient's rehabilitation training. Monitor the movement frequency of the patient's lower limbs and set a movement frequency threshold f. max When the exercise frequency f exceeds the threshold, an exercise frequency alarm Alert(f) is triggered to analyze the patient's rehabilitation training exercise frequency. Record the bending angle limit of the patient's lower limb and set the bending angle limit θ. max When the bending angle θ exceeds the threshold, a bending angle alarm Alert(θ) is triggered to obtain the patient's recovery status. Measure the impact force of the patient's lower limbs upon landing and set a landing force threshold F. max When the landing force F exceeds the threshold, a landing force alarm Alert(F) is triggered to assess the stability of the patient's gait during rehabilitation training. Analyze the weight distribution of the patient's lower limbs and set a weight distribution threshold W. max When the force of gravity W exceeds the threshold, the force distribution alarm Alert(W) is triggered to obtain the force situation of the patient's rehabilitation training.

6. The intelligent lower limb rehabilitation training and interaction system according to claim 5, characterized in that, Based on the patient's lower limb range of motion data, a time-series waveform graph and histogram are provided. A threshold warning progress bar is designed to display A and A. max The degree of proximity is used to obtain the display accuracy of optimized activity amplitude data; Spectral analysis was performed on the patient's lower limb movement frequency to show the dominant frequency components and their comparison with the recommended frequency range. Color coding was used to identify different frequency ranges to obtain frequency stability indicators. Based on the lower limb flexion angle limit, the joint range of motion is set, and the joint angle changes are displayed in real time through 3D skeletal animation, with safe activity areas marked. θ =∑(n=1+k)(1-(θ(t)-θ) target (t)) / θ min / θ max ) / m; to obtain the angle control accuracy; where, P θ To ensure angle control accuracy, θ(t) represents the real-time angle. target (t) represents the target angle, m represents the number of sampling points within the evaluation time period, and θ min θ represents the minimum bending angle of the lower limb. max This represents the maximum bending angle of the lower limbs. Based on the landing force data, a plantar pressure distribution heatmap is generated, and the force changes within the gait cycle are displayed in combination with the time dimension to obtain the left-right balance pressure index. Based on gravity distribution data W, a dynamic visualization of the center of gravity trajectory is constructed to display the patient's center of gravity shift process, using Sw = 1 - ∑(n = 1 + k)(W(t) - W). avg ) / (n·W max ), to obtain the center of gravity stability index; where S W W(t) is the index of center of gravity stability, and W(t) is the distribution of gravity force at time t. avg To achieve an average distribution of gravitational force, W max This is for data on the maximum distribution of gravity.

7. The intelligent lower limb rehabilitation training and interaction system according to claim 1, characterized in that, Based on the feedback rehabilitation training information data, the patient's lower limb muscle electrical signal status rehabilitation progress is assessed, and the rehabilitation training plan is adjusted in a timely manner, including: Based on the feedback of rehabilitation training information data, an assessment mechanism for the status of lower limb muscle electrical signals of patients was established to obtain a comprehensive rehabilitation status score in order to assess the patient's rehabilitation status. Based on the comprehensive rehabilitation status score, an individual rehabilitation prediction model is established to predict the rehabilitation status at a future time t+Δt; and a dynamic rehabilitation objective function is set to adaptively adjust the rehabilitation training program as the rehabilitation status progresses. Based on the adaptive adjustment of rehabilitation training programs, a rehabilitation training program adjustment strategy is generated through a rehabilitation difference comparison algorithm; and a fuzzy logic control system is used to finely adjust the training intensity to obtain a personalized rehabilitation training program.

8. The intelligent lower limb rehabilitation training and interaction system according to claim 1, characterized in that, When abnormal situations arise during the adjustment of the rehabilitation training program, a rapid response and immediate suspension of training should be initiated, with the output of abnormal warning information, including: A multi-layered anomaly detection mechanism is established based on the adjustment strategy of the rehabilitation training program, and an abnormal state discrimination function is constructed to obtain the anomaly discrimination result of the adjustment strategy. Based on the anomaly detection results of the adjustment strategy, a set of safety thresholds for key physiological parameters, S = s1, s2, ..., s, is set. i Real-time monitoring of physiological parameters x i With safety threshold s i The relationship between the real-time monitored physiological parameters x is used to determine the relationship between the parameters x and x. i Does it exceed the safety threshold s? i ; Based on physiological parameter x i With safety threshold s i Based on the relationship, an anomaly urgency assessment function is constructed to evaluate the anomaly urgency level; and an emergency stop control strategy for the training system is set, with a stop response time τ. s top satisfies: τ s top < 100ms; to ensure patient safety; A hierarchical early warning information mechanism is established based on the anomaly urgency assessment function, and different levels of early warning are output according to the anomaly urgency assessment function to output different early warning information.

9. The intelligent lower limb rehabilitation training and interaction system according to claim 1, characterized in that, Also includes: Lower limb skeletal muscle sensing module: Acquires electromyographic signals through a high-precision surface electromyography sensor array, and uses differential amplification technology to suppress common-mode interference and bio-potential acquisition algorithm to obtain lower limb skeletal muscle sensing signals; Human-computer interaction module: Provides an intuitive graphical user interface, adopts adaptive layout technology to adapt to different terminal devices, realizes data visualization and responds to manual adjustment of various rehabilitation training parameters, including real-time waveform graphs, muscle activity heat maps and rehabilitation progress dashboards; Adaptive contraction dynamics module: Based on the Hill muscle model, analyze muscle contraction characteristics and establish a force-velocity relationship framework to achieve automatic identification and classification of muscle contraction patterns; Muscle fatigue analysis module: assesses the degree of muscle fatigue by monitoring the change in the median frequency of the electromyographic signal spectrum, and analyzes the degree of muscle fatigue corresponding to the muscle fatigue index; And establish a fatigue recovery model to predict muscle recovery time for the next rehabilitation training operation; Rehabilitation Phase Training Module: Automatically divides training phases according to rehabilitation progress, adopts a step-by-step intensity increase strategy, sets multiple difficulty levels to obtain different advanced rehabilitation training content, until the patient's lower limb rehabilitation meets physiological adaptation. Emergency pause module: Employs multi-sensor fusion technology to detect abnormal states and obtains the probability of anomalies based on a Bayesian decision model; When the probability of an anomaly exceeds the set safety threshold s i When this occurs, the motor emergency stop protection mechanism is triggered; Training and assessment module: Comprehensively analyzes training data to generate multi-dimensional rehabilitation assessment reports, obtains a comprehensive rehabilitation effect score through quantitative rehabilitation effects, and reveals rehabilitation trends through time-series comparisons, providing data support for doctors' decision-making.

10. The intelligent lower limb rehabilitation training and interaction system according to claim 9, characterized in that, The human-computer interaction module further includes: Patient end and doctor end; The patient terminal connects to the system terminal via Bluetooth and executes the established rehabilitation training plan according to the rehabilitation training suggestions set by the doctor terminal. The plan includes set angles, speeds, times, and torque protection. The patient terminal then uploads the training data generated during the rehabilitation training process to the cloud database in real time, and provides feedback on the training status or asks questions to the doctor terminal by uploading text, pictures, and videos, so as to obtain professional rehabilitation guidance and answers to questions from the doctor terminal. The doctor's terminal can view the rehabilitation training history data of the patient's terminal; the history data includes rehabilitation training information data such as text, pictures and videos uploaded by the patient's terminal, as well as electrical signal feedback images of lower limb muscle rehabilitation training; thereby the doctor's terminal can provide rehabilitation training suggestions to the patient's terminal and adjust the rehabilitation training plan in a timely manner; The patient and doctor terminals enable efficient interaction between patients and doctors in a home-based rehabilitation environment.

Images