Wing-shaped shoulder rehabilitation training evaluation system based on simplified signal expansion

Abstract

The invention discloses a wing-shaped shoulder rehabilitation training evaluation system based on simplified signal expansion, which realizes accurate monitoring and evaluation of a wing-shaped shoulder patient through household wearable IMU sensing equipment assisted by EMG + IMU signal fusion processing, can deeply analyze the activity states of muscles and joints, and can improve the rehabilitation training accuracy of the wing-shaped shoulder. And a personalized and standardized rehabilitation training scheme is provided for the patient. The system can assist a patient in more scientific and intelligent rehabilitation training in hospitals and home environments, the rehabilitation effect is improved, discomfort in the rehabilitation period is relieved, and a real-time data feedback and rehabilitation progress evaluation tool is provided for doctors.

Description

Technical Field

[0001] The present invention relates to the field of scapula winging rehabilitation equipment, and in particular to a scapula winging rehabilitation training evaluation system based on simplified signal expansion. Background Art

[0002] Shoulder joint and scapula movement disorders, such as winged scapula, have become one of the common movement dysfunctions in clinical practice. Winged scapula is usually caused by dysfunction or injury of the scapula control muscles (such as serratus anterior and levator scapulae), which seriously affects the range of motion of the shoulder and the normal rhythm of the scapula, causing patients to experience shoulder instability, movement difficulties and other problems in daily activities. Traditional rehabilitation treatment for winged scapula mainly relies on physical therapy, muscle strengthening training and surgical intervention, but most of these treatments rely on complex assessment tools and equipment, and it is difficult to achieve continuous monitoring and personalized training during home rehabilitation.

[0003] The existing winged scapula rehabilitation training equipment generally has the following problems: First, most devices need to rely on a variety of complex sensors, such as surface electromyography devices, inertial measurement units, visual sensor units and pressure sensors, etc. These devices are more convenient to use in hospital environments, but they have limitations on the convenience and continuous evaluation of patients in home rehabilitation. Secondly, although surface electromyography signals can provide detailed information about muscle activation and movement, they require specific wear and high signal acquisition accuracy, and are often difficult to use in daily rehabilitation training, especially during home training, and cannot provide real-time and accurate feedback on the patient's muscle activity.

[0004] In view of the above problems, traditional rehabilitation training equipment and methods have failed to effectively achieve real-time evaluation and accurate guidance of patients during rehabilitation. Therefore, there is an urgent need for a new solution that can provide more convenient equipment for comprehensive rehabilitation evaluation and training guidance through simplified sensor configuration, especially relying only on IMU signals during home rehabilitation.

[0005] The present invention aims to solve the problem of over-reliance on EMG signals in the prior art by designing a set of simplified wearable devices, using inertial sensor signals (IMU) and fusion signal processing technology. When patients are rehabilitating at home, IMU signals are used to evaluate shoulder movement and muscle activation, and accurate rehabilitation evaluation is provided through data correction and signal fusion technology. This new technology not only improves the convenience of home rehabilitation, but also enables better real-time monitoring and feedback of patients' rehabilitation progress and movement standards, thereby providing a more efficient, continuous and personalized rehabilitation training system for patients with winged scapula. Summary of the invention

[0006] The purpose of the present invention is to address the deficiencies of the prior art. By combining the acquisition and data calibration of EMG and IMU signals, the present invention provides a winged scapula rehabilitation training evaluation system based on simplified signal expansion, that is, expansion from inertial sensor signals to surface electromyography signals. The core innovation of the invention is to perform movement evaluation and rehabilitation training for winged scapula patients in a home rehabilitation environment by using only IMU signals, and to provide more accurate patient rehabilitation guidance through the joint correction of EMG signals and IMU signals.

[0007] The objective of the present invention is achieved through the following technical solution: A scapula winging rehabilitation training evaluation system based on simplified signal expansion comprises the following steps: The sensor module is used to collect the electromyographic signals and inertial sensor signals of patients with winged scapula performing rehabilitation training movements under the guidance of a doctor, and analyze the patient's scapular activity and muscle activation level; The data processing module is used to fuse the collected electromyographic signals with the inertial sensor signals to establish an individualized rehabilitation data model for patients with winged scapula; The training evaluation module is used to collect inertial sensor signals when patients with winged scapula perform rehabilitation training alone, and evaluate the standard degree of the patient's rehabilitation training movements and muscle activation based on the rehabilitation data model.

[0008] Furthermore, the inertial sensor signal includes angle data, acceleration data and angular velocity data, which are used to characterize the patient's range of motion and trajectory of motion; the relative motion of the patient's scapula is calculated through angle changes, thereby providing accurate movement standard evaluation for rehabilitation training.

[0009] Furthermore, the electromyographic signal is collected by a multi-channel electromyographic sensor and is used to evaluate the electrical activity of the muscle and the movement of the scapula.

[0010] Furthermore, the training evaluation module includes home rehabilitation equipment, and the patient wears the home rehabilitation equipment to perform rehabilitation training according to the patient data calibrated in the hospital stage; the home rehabilitation equipment is used to monitor the patient's range of motion and angle of motion during the rehabilitation training process in real time, and provide feedback according to preset standards.

[0011] Furthermore, the data processing module includes a signal fusion algorithm unit for performing time alignment and data fusion on the electromyographic signal and the inertial sensor signal to generate a rehabilitation training evaluation model.

[0012] Furthermore, the data processing module also includes a patient calibration unit, which is used to calibrate the individual characteristics of the patient according to the hospital rehabilitation data, and generate an individualized rehabilitation training evaluation model for the corresponding patient as a personalized training evaluation standard.

[0013] Furthermore, the signal fusion algorithm adopts a method based on machine learning or deep learning to identify and evaluate muscle activation patterns and movement standards during patient rehabilitation training.

[0014] Furthermore, the home rehabilitation equipment includes: Wearable inertial sensor modules are used by patients to wear during home rehabilitation and provide real-time feedback by collecting inertial signals; The data analysis unit is used to evaluate the patient's rehabilitation movement standard and muscle activation based on the collected inertial signals, thereby helping the patient to perform rehabilitation training according to an individualized training plan.

[0015] Furthermore, the data analysis unit includes an algorithm module based on signal fusion, which is used to evaluate the quality of the patient's rehabilitation movements by analyzing inertial signals in combination with calibration data in the hospital stage.

[0016] Furthermore, the inertial sensor module includes at least one accelerometer and a gyroscope, which are used to measure the position signals of the patient's shoulders and limbs, and to provide feedback of the patient's movement standards in combination with the data analysis unit.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. Simplified home rehabilitation equipment: The present invention only relies on IMU signals to conduct rehabilitation training and assessment of patients during the home rehabilitation stage, eliminating the complex electromyographic sensors that must be used in traditional methods. This design greatly simplifies the device structure, reduces the difficulty of operation for patients, makes it suitable for daily use, and reduces maintenance costs. Especially for home environments, the portability and user-friendliness of the device have been significantly improved, which promotes the popularization and compliance of rehabilitation training.

[0018] 2. Fusion and correction mechanism based on multimodal data: This invention establishes an accurate motion correction model by collecting and fusing high-dimensional EMG and IMU signals, models the relationship between electromyographic signals and inertial signals, and ensures that high-precision motion assessment can be performed at home with only simplified IMU signals. This data fusion strategy achieves data consistency across environments, so that the accuracy and reliability of the assessment can be maintained even when the equipment and signal sources are simplified.

[0019] 3. High-precision muscle activation and movement standard assessment: The present invention introduces a signal dimension reduction and data fusion method based on regression analysis and neural network models. By constructing an algorithm, the complex relationship between inertial signals and electromyographic signals is modeled and optimized, so that the inertial signal can effectively reflect the degree of muscle activation and the standardization level of movement. This innovative data dimension reduction and mapping mechanism not only improves the accuracy of movement assessment, but also provides real-time dynamic feedback to help patients conduct more scientific and effective rehabilitation training. In addition, the system can adjust the rehabilitation plan according to the evaluation results to further optimize the rehabilitation effect.

[0020] The present invention provides an innovative and convenient rehabilitation training method for patients with winged scapula. Through signal fusion and dimensionality reduction, the configuration of home rehabilitation equipment is greatly simplified, while improving the data accuracy and treatment effect during the rehabilitation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be further introduced below.

[0022] Figure 1 It is a schematic diagram of the implementation process of the assisted winged scapula patient rehabilitation training evaluation system of the present invention.

[0023] Figure 2 Schematic diagram of the sensor marking site in the present invention.

[0024] Figure 3 It is a signal processing flow chart of data calibration and personalized rehabilitation training in the present invention. DETAILED DESCRIPTION

[0025] To make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation methods described herein are only used to explain the present invention and do not limit the scope of protection of the present invention.

[0026] The present invention provides a winged scapula rehabilitation training system based on simplified signal expansion or multi-signal fusion. The following is a detailed description of the implementation method, focusing on the core data processing method, device configuration and application process of the present invention.

[0027] Figure 1 FIG. 1 is a flowchart of an implementation of a scapula winging rehabilitation training evaluation system based on simplified signal expansion provided by an embodiment of the present invention. Figure 1 As shown, the rehabilitation system of the present invention includes three parts: a hospital-stage rehabilitation data acquisition system, a patient home rehabilitation training system, and a data processing and model optimization algorithm.

[0028] Figure 2 It is a schematic diagram of collection sites of a data collection system for patients with winged scapula during their hospital visit provided by an embodiment of the present invention.

[0029] The device provided by the embodiment of the present invention includes: a multi-channel surface electromyography sensor, which is used to collect muscle electrical signals in the hospital stage. The signal range is usually in the microvolt level, which is used to evaluate the activation state of the muscle; an inertial measurement unit, which is used to collect the angle, acceleration and angular velocity data of the patient's shoulder and arm to help monitor the movement trajectory.

[0030] Data collection process at this stage: Doctors assist in installing multi-channel electromyographic sensors at the subject's shoulder, upper arm and forearm to perform electromyographic detection to ensure that the electrical activity of the target muscle group can be detected and reflect the activation state and motor function of the muscle. Each electrode channel receives the corresponding electromyographic signal EMG (such as , The data were collected through an amplifier and analog-to-digital converter (ADC) with a sampling rate of 2kHz.

[0031] Doctors assist patients in wearing IMU devices on their shoulders, elbows, and wrists to collect inertial sensor signals The IMU device includes an accelerometer, a gyroscope, and a magnetometer. It collects the angle, acceleration, and angular velocity data of the movement and provides dynamic information of the shoulder joint motion system (such as , Number of IMU devices when training for patient hospitals).

[0032] The combination of EMG and IMU signals can provide doctors with comprehensive data on the patient's muscle activation and dynamic movement of the shoulder joint. All signals will be recorded synchronously for subsequent analysis and adjustment.

[0033] The subjects performed the predetermined shoulder movements and training actions. The doctor assisted in collecting EMG and IMU signal data and performed motion correction to establish a personalized data model for the patient.

[0034] In this process, the acquisition of EMG signals and IMU signals are processed and converted by the following calculation formulas: IMU signal angle conversion: IMU sensors are generally composed of accelerometers, gyroscopes, and magnetometers, and can provide data such as acceleration, angular velocity, and magnetic field strength. In order to convert IMU signals into meaningful angle information, quaternion or Euler angle conversion methods are usually required. Specifically, assuming that the acceleration information output by the IMU sensor is , then the pitch angle of the shoulder can be calculated by the following formula:

[0035] in, is the pitch angle of the shoulder, is the acceleration value of the IMU accelerometer on three axes. Similarly, the roll angle (roll) and yaw angle (yaw) can be calculated in a similar way, further fusing the data of the accelerometer, gyroscope and magnetometer to obtain a more accurate angle estimate.

[0036] EMG signal processing and action evaluation: EMG signals reflect the electrical activity of muscles. By collecting EMG signals, the degree of muscle activation can be obtained. In the present invention, EMG signals are not only used to evaluate muscle activity, but also to assist in the evaluation and correction of training movements.

[0037] In order to process the EMG signal, filtering and denoising are first performed. The EMG signal is processed using bandpass filtering to remove low-frequency and high-frequency noise. The filtering range is generally 10Hz to 500Hz.

[0038]

[0039] in, is the original EMG signal collected, is the impulse response function of the filter, is the filtered signal. By removing the low-frequency baseline drift and high-frequency noise, an effective electromyographic signal is obtained.

[0040] The main components of muscle activity are extracted through filters, and the time domain and frequency domain characteristics of the electromyographic signal are calculated, such as Wilson amplitude (Wamp), mean absolute value (MAV), root mean square (RMS), frequency variance (FVAR), power spectral density (PSD) and other muscle activation characteristics.

[0041] Wilson amplitude (Wamp) represents the number of times the amplitude of two consecutive sampling points of the surface electromyography signal changes beyond the set threshold. The larger the Wilson amplitude, the more points in this signal with drastic amplitude changes: .

[0042] in, is the total number of signal sampling points, The signals are and The sampling value of the point, To set the threshold, is a unit step function.

[0043] The larger the mean absolute value (MAV), the larger the absolute value of the surface electromyographic signal amplitude, the stronger the signal energy, and the stronger the muscle force: .

[0044] The signal root mean square (RMS) value calculates the degree of muscle activation through integration and square method, representing the effective value of a segment of surface electromyographic signal. The larger the root mean square value, the stronger the energy of the signal: .

[0045] in, is the number of sampling points in the time window, For the window sampling points, For The sample value at The frequency variance (FVAR) of the surface electromyographic signal is the square of the difference between the horizontal coordinate of the corresponding spectrum and the center frequency of the signal, then convolved with the vertical coordinate of the spectrum and finally divided by the sum of the vertical coordinates. The size of the frequency variance reflects the frequency distribution and change trend of muscle activity and can be used to evaluate the muscle movement state: .

[0046] in, is the frequency variable of the EMG signal in the frequency domain, is the center frequency of the spectrum, is the power spectral density.

[0047] At the same time, the envelope signal (EMG profile, PF) of the EMG signal is also introduced as a feature. The envelope of the EMG signal can evaluate the movement of the joint body through the proportional control strategy, and further evaluate the degree of muscle activation of the patient's different movements. By extracting multiple types of EMG signal features, the muscle activation of the patient during shoulder movement can be evaluated, and further judgment can be made as to whether the patient is following the standard rehabilitation movements.

[0048] The equipment provided by the embodiment of the present invention during the home rehabilitation training stage for patients with winged scapula includes: a convenient IMU wearable device, and the patient only needs to wear a simplified IMU device for daily rehabilitation training during home training to monitor the movement angles and movement trajectories of the shoulders and arms.

[0049] During the data collection process at this stage, the patient only wears a simplified IMU device. The IMU device collects motion data signals through the IMU sensor. (like , The number of IMU devices used for home training for patients). The motion data collected by the IMU sensor is used to evaluate the patient's shoulder movement and provide real-time feedback to help the patient complete specific rehabilitation movements. Based on the EMG and IMU signal data collected during the hospital stage, more personalized rehabilitation guidance is provided to the patient.

[0050] The present invention provides a set of evaluation algorithms and rehabilitation movement standard evaluation for rehabilitation training of patients with winged scapula. The collected EMG signals and IMU signals provide baseline data of normal shoulder movement of patients. The correction model constructed based on these data can correct and standardize the signals using only IMU during home rehabilitation. By comparing and analyzing the IMU signals with the EMG signals collected by the hospital, the standard degree of shoulder movement and rehabilitation progress can be evaluated by the patient at home relying only on the IMU signals.

[0051] The specific correction method includes a machine learning model based on EMG and IMU signals. By combining traditional filtering and neural network methods, the signal fusion process is further optimized. Assume that the EMG signals and IMU signals collected in the hospital stage are and , the IMU signal at home is , then the correction formula based on the model is:

[0052] in, The fusion function can correct the home IMU signal according to the EMG and IMU signals of the patient in the hospital stage. In this way, the patient can more accurately assess the degree of muscle activation and the degree of movement standard at home using only the IMU signal.

[0053] Figure 3 The invention provides a data processing and model iteration optimization algorithm for a scapula winging rehabilitation training evaluation method based on simplified signal expansion provided by an embodiment of the invention.

[0054] The EMG and IMU time series signals collected during the hospital visit need to be preprocessed. IMU data and EMG signals, as input sources of the method of the present invention, respectively reflect the movement characteristics and muscle activation of the human body.

[0055] IMU data includes multi-dimensional time series features such as acceleration, angular velocity, and device orientation information, which can reflect the dynamic characteristics of human motion. The IMU feature matrix is ​​as follows:

[0056] in, Represents acceleration data, Represents angular velocity data, each row corresponds to data at a different time step, Indicates the number of time steps.

[0057] In order to improve data quality, Kalman filtering is first applied to the IMU data to remove high-frequency noise, and the accelerometer and gyroscope data are used for posture estimation to extract more accurate shoulder angles and dynamic behavior features. By fusing multiple signals (such as acceleration and angular velocity of inertial data, and joint angles calculated by quaternions), a high-dimensional feature vector that can describe the motion posture is formed.

[0058] The EMG signal is a time series data of multiple channels, which records the electrical activity of muscles during movement. It has time series continuity and multi-channel characteristics and is a direct representation of muscle activity. The EMG signal needs to be processed synchronously, using a bandpass filter to eliminate low-frequency and power frequency interference (such as 0.5Hz to 50Hz), and smoothing the signal waveform by moving average method. In the same time window, the aligned EMG signal is segmented into corresponding target matrices, expressed as:

[0059] in, represents the number of time steps, is the number of channels of the EMG signal, Line List express Time step EMG signal value of the channel.

[0060] After data collection is completed, the IMU and EMG signals are normalized to eliminate the dimensional differences between different data dimensions, reduce the range of eigenvalues, and improve the convergence speed and stability of the network. The normalization formula is as follows:

[0061] in, represents the normalized eigenvalue, is the original eigenvalue, and The normalization operation ensures that all input feature values ​​are within the same numerical range, which helps to eliminate the deviation caused by the difference in feature values ​​and makes the model consistent when processing different modal features.

[0062] The goal of the fusion module is to match the time step features of the IMU matrix and the EMG matrix one by one, providing structured input for subsequent modeling. Since the original data has high-dimensional characteristics, in order to reduce the computational complexity and extract key features, this method uses principal component analysis (PCA) to reduce the dimensionality of the input data.

[0063] First, the aligned IMU and EMG signal data are merged into a unified multi-dimensional time series input matrix , the shape is:

[0064] Among them, n is the number of time steps and m is the feature dimension (including all channels of IMU and EMG).

[0065] PCA extracts the most representative features by maximizing the data variance while reducing redundant information to generate a feature matrix after dimensionality reduction:

[0066] in, is the PCA feature matrix, is the dimension after dimensionality reduction, is the feature matrix after dimensionality reduction. After PCA dimensionality reduction, the dimension of the input data is significantly reduced from m to k, while retaining the main features of the data, providing efficient feature expression for subsequent deep learning model input.

[0067] In order to extract the temporal features in IMU data, this method introduces a framework combining convolutional neural network (CNN) and long short-term memory network (LSTM) after the fusion module, which is used to extract local patterns and model time series dependency of the fusion matrix of IMU and EMG, respectively.

[0068] The CNN module is used to capture the local feature patterns of the fusion matrix. The correlation between time steps is extracted through multiple layers of convolution kernels to generate preliminary feature representations:

[0069] in, is the convolution kernel weight, is the bias term, It is an activation function used to enhance nonlinear expression capabilities.

[0070] The convolutional features are further used to extract time series dependencies through the LSTM layer. The LSTM unit controls the flow of information through the forget gate, input gate, and output gate. The calculation formula is: Forget gate activation value:

[0071] Enter the gate activation value:

[0072] Output gate activation value:

[0073] Candidate cell states: Update cell status:

[0074] Hidden state: in, is the weight matrix of the forget gate, and is the hidden state of the current moment and the previous moment, is the input value of the network, is the bias term of the forget gate, is the Sigmod activation function, is the weight matrix of the input gate, is the bias term of the forget gate, is the weight matrix of the input gate, is the bias term of the forget gate, is the weight matrix of the output gate, is the bias term of the output gate, is the weight matrix of candidate cell states, is the bias term of the candidate cell state, is the hyperbolic tangent activation function. After the LSTM layer, the hidden state of the time series is finally generated to represent , which can be used to capture long-term dependency features.

[0075] In order to further improve the model's sensitivity to time series features, the method of the present invention adds a self-attention mechanism after the LSTM output to adaptively assign weights to different time steps. The normalized weight of the attention score is calculated using the following formula:

[0076] in, is the query vector, which represents the target to be focused on. is the key vector, representing the characteristics of the candidate information, for right The attention weight is used to weight the feature matrix and generate the fused global time series feature representation as .

[0077] The fused features are nonlinearly mapped through the fully connected layer to finally generate the predicted value of the EMG signal. The calculation formula of the fully connected layer is as follows:

[0078] in, is the weight matrix of the fully connected layer, is the bias term.

[0079] This model uses mean square error (MSE) as the loss function to measure the gap between the predicted fusion feature value and the actual fusion feature value:

[0080] in, is the first The next time step The eigenvalues ​​of the features, is the first Time step The eigenvalues ​​of the features, To divide the time steps, is the preset number of features.

[0081] The Adam optimization algorithm is used to update the model parameters to ensure that the model converges quickly and achieves optimal performance during training.

[0082] Use the regression model to map the fusion features to the IMU signal. The regression equation is as follows:

[0083] in, This is the IMU signal collected during the hospital phase. is the first Features, is the first The coefficient of a feature.

[0084] Based on the fused data, the computer program uses a pre-trained model for analysis and comprehensive evaluation. It can determine the recovery progress of the muscle function of patients with winged scapula and the standard degree of the patient's home rehabilitation movements, and provide real-time feedback and guidance.

[0085] The computer device, including a memory, a processor and a computer program running on the processor, can realize multimodal data acquisition and processing, meet the needs of winged scapula rehabilitation diagnosis and treatment, and optimize the patient's rehabilitation plan through iteration.

[0086] The spatial feature information of the shoulder can be obtained through the IMU signal , the trained model can be combined with the patient's historical EMG and IMU data to infer the characteristics of the fusion spatial information and muscle activation degree. :

[0087] in, It is a regression model obtained through training, which can map the IMU signal to the feature value of the fused EMG information and evaluate the degree of muscle activation.

[0088] One of the key innovations of the present invention is to fuse the EMG signals and IMU signals collected in the hospital stage through data processing, and reduce the dimension to IMU signals, so that effective rehabilitation assessment and movement standard degree inference can be carried out only by relying on IMU signals during the home rehabilitation stage.

[0089] During the home rehabilitation stage, the final IMU signal is processed by regression prediction and combined with the shoulder movement angle to evaluate the patient's rehabilitation progress and movement standard. At this time, the home IMU signal can not only reflect the change in shoulder angle, but also indirectly infer the muscle activation state based on the following formula:

[0090] The specific implementation process of effective rehabilitation assessment and movement standard degree inference during home rehabilitation stage relying only on IMU signals is as follows:

[0091] in, The characteristic matrix for home rehabilitation training Features, is the coefficient of the corresponding feature. By predicting the features of the home IMU signal and combining it with the regression model, the patient's shoulder movement standard and muscle activation degree can be evaluated.

[0092] The home system will provide real-time feedback to the patient based on the evaluation results of the IMU signal, indicating whether the patient's rehabilitation movements meet the standards. For example, if the patient's movements are excessive or too relaxed, the system will prompt the patient to adjust his posture.

[0093] The patient's movement progress, shoulder angle changes, and muscle activation are visualized in the system. A graph is used to show the change curve of the shoulder angle and the degree of muscle activation calculated by the regression model.

[0094] As the patient's recovery progresses, the system will continuously adjust the regression model based on new IMU signal data, optimize the patient's rehabilitation path, and automatically update data standards based on the rehabilitation progress to ensure continued rehabilitation effects.

[0095] Through the above data collection and calibration, patients can use convenient IMU devices for continuous rehabilitation training at home after completing data correction in the hospital. Doctors can remotely monitor and guide patients based on real-time feedback of IMU data and their rehabilitation progress, thereby improving rehabilitation effects and accelerating the rehabilitation process.

[0096] The present invention uses wearable IMU sensing devices at home, supplemented by EMG+IMU signal fusion processing, to achieve accurate monitoring and evaluation of patients with winged scapula, and can deeply analyze the activity status of muscles and joints, thereby providing patients with personalized and standardized rehabilitation training programs. The present invention can assist patients in conducting more scientific and intelligent rehabilitation training in hospitals and home environments, improve rehabilitation effects, reduce discomfort during rehabilitation, and provide doctors with real-time data feedback and rehabilitation progress evaluation tools.

[0097] Through the above processing process, the present invention can effectively evaluate the rehabilitation progress and movement standard degree of patients with winged scapula by relying only on IMU signals during the home rehabilitation stage, providing an innovative and efficient rehabilitation training evaluation system for patients with winged scapula.

[0098] The above embodiments are used to illustrate the present invention rather than to limit the present invention. Any modification and change made to the present invention within the spirit of the present invention and the protection scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A scapula winging rehabilitation training evaluation system based on simplified signal expansion, characterized in that: The following steps are involved: The sensor module is used to collect the electromyographic signals and inertial sensor signals of patients with winged scapula performing rehabilitation training movements under the guidance of a doctor, and analyze the patient's scapular activity and muscle activation level; The data processing module is used to fuse the collected electromyographic signals with the inertial sensor signals to establish an individualized rehabilitation data model for patients with winged scapula; The training evaluation module is used to collect inertial sensor signals when patients with winged scapula perform rehabilitation training alone, and evaluate the standard degree of the patient's rehabilitation training movements and muscle activation based on the rehabilitation data model.

2. The scapula winging rehabilitation training evaluation system based on simplified signal expansion according to claim 1, characterized in that: The inertial sensor signal includes angle data, acceleration data and angular velocity data, which are used to characterize the patient's range of motion and trajectory; the relative movement of the patient's scapula is calculated through angle changes, thereby providing accurate movement standard evaluation for rehabilitation training.

3. The scapula winging rehabilitation training evaluation system based on simplified signal expansion according to claim 1, characterized in that: The electromyographic signals are collected through a multi-channel electromyographic sensor and are used to evaluate the electrical activity of the muscles and the movement of the scapula.

4. The scapula winging rehabilitation training evaluation system based on simplified signal expansion according to claim 1, characterized in that: The training evaluation module includes home rehabilitation equipment, and the patient wears the home rehabilitation equipment to perform rehabilitation training according to the patient data calibrated in the hospital stage; the home rehabilitation equipment is used to monitor the patient's range of motion and angle of motion during the rehabilitation training process in real time, and provide feedback according to preset standards.

5. The scapula winging rehabilitation training evaluation system based on simplified signal expansion according to claim 1, characterized in that: The data processing module includes a signal fusion algorithm unit, which is used to perform time alignment and data fusion on the electromyographic signal and the inertial sensor signal to generate a rehabilitation training evaluation model.

6. The scapula winging rehabilitation training evaluation system based on simplified signal expansion according to claim 5, characterized in that: The data processing module also includes a patient calibration unit, which is used to calibrate the individual characteristics of the patient according to the hospital rehabilitation data, and generate an individualized rehabilitation training evaluation model for the corresponding patient as a personalized training evaluation standard.

7. The scapula winging rehabilitation training evaluation system based on simplified signal expansion according to claim 5, characterized in that: The signal fusion algorithm uses a method based on machine learning or deep learning to identify and evaluate muscle activation patterns and movement standards during patient rehabilitation training.

8. The scapula winging rehabilitation training evaluation system based on simplified signal expansion according to claim 4, characterized in that: The home rehabilitation equipment includes: Wearable inertial sensor modules are used by patients to wear during home rehabilitation and provide real-time feedback by collecting inertial signals; The data analysis unit is used to evaluate the patient's rehabilitation movement standard and muscle activation based on the collected inertial signals, thereby helping the patient to perform rehabilitation training according to an individualized training plan.

9. The scapula winging rehabilitation training evaluation system based on simplified signal expansion according to claim 8, characterized in that: The data analysis unit includes an algorithm module based on signal fusion, which is used to evaluate the quality of the patient's rehabilitation movements by analyzing inertial signals in combination with calibration data in the hospital stage.

10. The scapula winging rehabilitation training evaluation system based on simplified signal expansion according to claim 8, characterized in that: The inertial sensor module includes at least one accelerometer and a gyroscope, which are used to measure the position signals of the patient's shoulders and limbs, and to feed back the patient's movement standards in combination with the data analysis unit.

Images