Auxiliary rehabilitation system for postoperative exercise recovery

Abstract

The invention relates to the technical field of medical rehabilitation, in particular to an auxiliary rehabilitation system for postoperative exercise recovery, comprising a data acquisition module which is worn on an ankle or an instep, is internally provided with a three-axis accelerometer and / or a three-axis gyroscope, sets an initial sampling frequency, and performs range configuration and static calibration; the data processing module is used for carrying out coordinate system conversion, noise elimination and data calibration on the collected data, carrying out specific calculation on key parameters and finally executing a Kalman wave algorithm fusion algorithm to obtain a user attitude estimation result; the feedback module is used for extracting and analyzing the step length, the step frequency and the swing phase of the user through modeling and feeding back motion data guidance information to the patient; and the power supply module is internally provided with a rechargeable lithium battery for supplying power and is used for providing stable power support. The method has the advantages that the motion state of the patient can be monitored in real time, the artificial intelligence end-side large model is used for processing data, and personalized rehabilitation guidance is provided for the patient.

Description

Technical Field

[0001] The present invention relates to the technical field of medical rehabilitation, and particularly to an auxiliary rehabilitation system for postoperative motor recovery.

Background Art

[0002] In the hospital, stroke patients and limb-disabled patients can receive professional rehabilitation guidance and treatment well. Rehabilitation exercise is a long-term process. After discharge, patients still need to continue rehabilitation treatment. However, since the rehabilitation effect is closely related to the patient's subjective initiative, and due to the lack of professional rehabilitation guidance at home, the rehabilitation effect of patients is greatly reduced, affecting the healing of the patient's condition.

[0003] Postoperative motor recovery is crucial for patient rehabilitation. Traditional rehabilitation methods mainly rely on doctor guidance and the patient's own exercise, and have the following deficiencies: 1. Lack of objective data support: Medical staff cannot grasp the patient's motion state in real time, making it difficult to accurately evaluate the rehabilitation effect and adjust the rehabilitation plan; 2. The rehabilitation process is boring: Patients lack effective interaction and feedback, and are prone to slack off, affecting the rehabilitation effect; 3. Low personalization level: The unified rehabilitation plan is difficult to meet the individual needs of different patients.

[0004] Therefore, an auxiliary rehabilitation system for postoperative motor recovery is proposed, which can improve the system's autonomous learning ability by real-time monitoring the patient's motion state and using an artificial intelligence edge large model to process data, provide personalized rehabilitation guidance for patients, improve the rehabilitation efficiency, and formulate a further reasonable rehabilitation training plan for patients.

Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the deficiencies of the prior art and provide an auxiliary rehabilitation system for postoperative motor recovery, which can improve the system's autonomous learning ability by real-time monitoring the patient's motion state and using an artificial intelligence edge large model to process data, provide personalized rehabilitation guidance for patients, improve the rehabilitation efficiency, and formulate a further reasonable rehabilitation training plan for patients.

[0006] To solve the above technical problems, the present invention adopts the following technical solutions:

[0007] An auxiliary rehabilitation system for postoperative motor recovery, comprising

[0008] a data acquisition module, worn on the ankle or instep, internally provided with a three-axis accelerometer and / or a three-axis gyroscope, setting an initial sampling frequency, and performing range configuration and static calibration;

[0009] The data processing module performs coordinate system conversion, noise elimination, and data calibration on the collected data, calculates specific key parameters, and finally executes the Kalman wave algorithm fusion algorithm to obtain the user's posture estimation result;

[0010] The feedback module is used to identify gait characteristics, extract and analyze the user's step length, step frequency, and swing phase through modeling, and feedback motion data such as motion postures, time, and frequency guidance information to the patient. At the same time, it customizes a personalized rehabilitation plan for the patient;

[0011] The power supply module is internally equipped with a rechargeable lithium battery for power supply to provide stable power support;

[0012] The feedback module determines the user's motion posture based on the data collected by the data acquisition module and analyzes the posture estimation result of the processing result. The formula for the posture estimation result analysis is:

[0013]

[0014] In the formula, X t represents the state value of the three-axis gyroscope at time t; X t-1 represents the state value of the three-axis gyroscope at time t-1; ω x represents the swing amplitude of the foot in the x-axis direction; ω y represents the swing amplitude of the foot in the y-axis direction; ω z represents the swing amplitude of the foot in the z-axis direction; Δt represents the time interval; y t represents the noise elimination factor; Z t represents the state value of the three-axis accelerometer at time t; a x represents the motion acceleration of the foot in the x-axis direction; a y represents the motion acceleration of the foot in the y-axis direction; a z represents the motion acceleration of the foot in the z-axis direction; υ t represents the integral value of the acceleration after removing gravity.

[0015] Preferably, the static calibration includes calculating the offset between the sensor coordinate system and the world coordinate system according to the wearing position of the data acquisition module, recording the static offset of the data acquisition module, and calculating the static attitude angle. The formula is as follows: Among them, θ acc is the initial pitch angle of the gravity component when the user is stationary, and Φ acc is the initial roll angle of the gravity component when the user is stationary.

[0016] When abnormal values occur during motion, the abnormal values include falls, limping, and dragging gait trajectories, and abnormal jump values are detected through a sliding window.

[0017] First, calculate the window statistic, and the formula is

[0018]

[0019] where N is the window length, and x t-k is the current t - k point.

[0020] According to the sampling rate selection, for example, with a sampling rate of 100Hz, the window length N = 10, corresponding to 0.1 second of data.

[0021] Then, determine the jump value. If for the current point |x t - μ window | > k·σ window , it is marked as a jump value, where k takes the value of 3 or 2.5.

[0022] Preferably, use the acceleration integration method to calculate the movement speed, and reset the data in time when touching the ground to eliminate the drift error. The formula is: a motion = a world - [0, 0, g] T , where a motion is the gravity component separated during the movement process in the updated three - axis accelerometer, a world is the gravity component separated during the movement process in the initial three - axis accelerometer, and g is the gravitational constant.

[0023] Preferably, perform a Fourier transform operation on the acceleration signal on the z - axis to extract the main frequency. The formula is where X(f) is the acceleration signal in the vertical direction.

[0024] Preferably, calculate the acceleration vector sum, extract the movement peak value, and analyze the rotational angular velocity of the three - axis gyroscope to obtain the amplitude of the foot swing angle. a motion is the acceleration amplitude, and ω is the angular velocity amplitude.

[0025] Preferably, use the Kalman wave algorithm and fuse the data of the three - axis accelerometer and the three - axis gyroscope to finally obtain the three - dimensional attitude data value of the foot.

[0026] After identifying the motion posture and motion mode, use the SVM algorithm to detect abnormal states. The SVM algorithm is a support vector machine (Support Vector Machine, SVM), which is a class of generalized linear classifiers for binary classification of data in a supervised learning manner.

[0027] The specific operation steps of the algorithm are as follows:

[0028] (1). Data collection

[0029] (2) Data annotation:

[0030] Normal gait: Marked as class 0, such as walking, running.

[0031] Abnormal events: Marked as class 1, such as falling, limping, shuffling.

[0032] Synchronous recording: Ensure the correspondence between data and actions by video recording or manually marking timestamps.

[0033] (3) Dataset division

[0034] Training set: 70% - 80% of the data, used for model training.

[0035] Validation set: 10% - 15% of the data, used for parameter tuning.

[0036] Test set: 10% - 15% of the data, used for final evaluation.

[0037] (4) Data processing and SVM model training

[0038] Select the kernel function, tune the parameters, and use the pre - encapsulated SVM tool provided by the machine learning module scikit - learn developed based on the Python language for model training.

[0039] (5) Model evaluation and deployment

[0040] Evaluate by calculating precision and recall.

[0041] Export the trained SVM model parameters and deploy them to the mobile app side, or consider directly deploying them to the device side.

[0042] It also includes a voice vibration reminder module, which is used to encourage the user by voice when the user's exercise state is normal and continuous, and vibrate to remind and issue a warning when in an abnormal state;

[0043] A sports report generation module, which is used to generate a visual report. The visual report includes the attainment of the exercise amount and the standard degree of the exercise posture;

[0044] A rehabilitation plan generation module, which is used to generate the daily exercise pattern and time arrangement.

[0045] Advantages of adopting the present invention:

[0046] 1. Real-time monitoring and guidance: By collecting motion state data in real time through an acceleration sensor and a gyroscope, and analyzing and processing it using an artificial intelligence large model on the edge side, it can provide patients with real-time guidance information such as motion postures, time, and frequency, ensuring that patients maintain the correct motion state during the rehabilitation training process and improving the rehabilitation effect.

[0047] Personalized rehabilitation plan: According to the patient's motion data and rehabilitation progress, a personalized rehabilitation plan is customized for the patient to meet the rehabilitation needs of different patients and improve the rehabilitation efficiency.

[0048] Lightweight operation: The wearable device is designed to be lightweight and convenient to wear, without affecting the patient's normal activities, improving the patient's usage experience.

[0049] Reduce rehabilitation costs: Reduce the dependence on medical staff, lower labor costs, and at the same time improve the accessibility and convenience of training.

[0050] 2. First, calculate the window statistic, and the formula is

[0051]

[0052] where N is the window length, and x t-k is the current t - k point.

[0053] According to the sampling rate selection, such as a sampling rate of 100Hz, the window length N = 10, corresponding to 0.1 second of data.

[0054] Then, judge the jump value. If the current point |x t - μ window | > k·σ window , it is marked as a jump value, where k takes the value of 3 or 2.5. Reduce the influence of variant values on the recognition of the user's motion trajectory, improve the recognition accuracy, and improve the computing power accuracy.

[0055] 3. Use the acceleration integration method to calculate the motion speed, and reset the data in time when touching the ground to eliminate the drift error. The formula is: a motion = a world - [0, 0, g] T , where a motion is the gravity component separated in the updated three-axis accelerometer during the motion process, a world is the gravity component separated in the initial three-axis accelerometer during the motion process, and g is the gravitational constant. Reduce the influence of drift error on the system accuracy and improve the algorithm precision.

[0056] These features and advantages of the present invention will be detailedly disclosed in the following specific embodiments and drawings.

Description of the Drawings

[0057] The present invention will be further described below in conjunction with the accompanying drawings:

[0058] Figure 1 It is the overall schematic diagram of the auxiliary rehabilitation system of the present invention.

Specific Embodiment

[0059] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0060] The concepts involved in the present application will be described below first in conjunction with the accompanying drawings. It should be noted here that the descriptions of the following concepts are only for making the content of the present application easier to understand, and do not represent the limitation of the protection scope of the present application.

[0061] Embodiment 1:

[0062] An auxiliary rehabilitation system for postoperative motor recovery, as Figure 1 shown, includes

[0063] a data acquisition module, worn on the ankle or instep, internally provided with a three-axis accelerometer and / or a three-axis gyroscope, setting an initial sampling frequency, and performing range configuration and static calibration;

[0064] a data processing module, performing coordinate system conversion, noise elimination, data calibration on the acquired data, and specifically calculating key parameters, and finally executing the Kalman wave algorithm fusion algorithm to obtain the user posture estimation result;

[0065] At the same time, an auxiliary rehabilitation method for postoperative motor recovery can be used, including S1. cleaning the data, correcting the incorrect data formats and outliers in the data, for duplicate records, determining whether it is a data entry error or data duplication, and taking operations of re-entry or deletion, and converting different measurement standards of the data into one standard measurement;

[0066] S2. extracting feature values, and identifying the features that are most important for predicting patient data through statistical tests, principal component analysis, and correlation analysis;

[0067] S3. Data analysis: Use different machine learning algorithms and deep learning algorithms to conduct in-depth predictive analysis on the patient's rehabilitation data, including various physiological data of the patient monitored by monitoring devices and the rehabilitation movements of the patient captured by cameras, to obtain the patient's health prediction results, evaluate the patient's movement accuracy, stability, and range of motion, and ensure that the patient performs rehabilitation actions correctly; analyze various physiological data of the patient to predict the patient's rehabilitation status and physiological changes during the rehabilitation process; set practical short-term and long-term goals for the patient based on the patient's rehabilitation history and current status; automatically adjust the rehabilitation plan according to the patient's response and progress to ensure that the patient receives the most suitable rehabilitation guidance.

[0068] Feedback module: Used to identify gait characteristics, extract and analyze the user's step length, step frequency, and swing phase through modeling, and provide the patient with guidance information on movement data such as movement posture, time, and frequency, while customizing a personalized rehabilitation plan for the patient;

[0069] In the feedback module, an abnormal movement warning threshold x can be customized for the patient according to medical knowledge and the patient's specific condition and rehabilitation progress. This threshold is dynamic and can be adjusted as the patient's rehabilitation situation changes. The data processing and analysis module analyzes and predicts the patient's movement data including heart rate, exercise intensity, exercise mode, etc. to predict whether there are potential abnormalities such as abnormal heart rate, abnormal exercise volume, or fall risk. The real-time monitoring and feedback module monitors the data after obtaining the prediction data and compares it with the preset threshold x to determine whether the data exceeds the threshold x. When the data exceeds the threshold x, the real-time monitoring and feedback module determines that the patient is performing abnormal movements. The real-time monitoring and feedback module will independently send out a warning signal to the front-end terminal through the background server. This warning may be in the form of sound, vibration, or visual cues, etc. The adoption of this warning method is set by the real-time monitoring and feedback module by judging the patient's current environment, aiming to immediately attract the attention of the patient or people around. The warning will continue until the monitored value of the patient's movement data returns to the normal value.

[0070] During the continuous warning period, the real-time monitoring and feedback module will provide some immediate intervention suggestions, such as prompting the patient to slow down the movement speed, stop the current activity, or take specific safety measures. The system dynamically adjusts the warning threshold according to the patient's actual rehabilitation situation, rather than the fixed warning threshold in the traditional medical system, which provides more accurate monitoring and warning; at the same time, the system uses multiple warning methods such as sound, vibration, and visual cues, which helps to ensure that patients can obtain warning information in a timely manner in different situations.

[0071] Power supply module: It is internally equipped with a rechargeable lithium battery for power supply to provide stable power support;

[0072] The feedback module determines the user's motion posture based on the data collected by the data acquisition module, and analyzes the posture estimation result of the processing result. The formula for the posture estimation result analysis is as follows:

[0073]

[0074] In the formula, X t represents the state value of the three-axis gyroscope at time t; X t-1 represents the state value of the three-axis gyroscope at time t-1; ω x represents the swing amplitude of the foot in the x-axis direction; ω y represents the swing amplitude of the foot in the y-axis direction; ω z represents the swing amplitude of the foot in the z-axis direction; Δt represents the time interval; y t represents the noise cancellation factor; Z t represents the state value of the three-axis accelerometer at time t; a x represents the motion acceleration of the foot in the x-axis direction; a y represents the motion acceleration of the foot in the y-axis direction; a z represents the motion acceleration of the foot in the z-axis direction; υ t represents the integral value of the acceleration after removing gravity.

[0075] By collecting motion state data in real time through the acceleration sensor and gyroscope, and using the artificial intelligence edge large model for analysis and processing, it can provide real-time guidance information such as motion posture, time, and frequency for patients, ensuring that patients maintain the correct motion state during the rehabilitation training process and improving the rehabilitation effect.

[0076] Static calibration includes calculating the offset between the sensor coordinate system and the world coordinate system according to the position where the data acquisition module is worn, recording the static offset of the data acquisition module, and calculating the static attitude angle. The formula is as follows: Among them, θ acc is the initial pitch angle of the gravity component when the user is stationary, and Φ acc is the initial roll angle of the gravity component when the user is stationary.

[0077] The movement of the foot in the air can be regarded as the flight movement of an aircraft. Pitch Angle: It refers to the angle between the transverse axis of the aircraft body (usually the short axis of the aircraft) and the horizontal plane, and is defined as the inclination angle between the nose of the aircraft and the horizontal plane. A positive pitch angle indicates that the nose of the aircraft is tilted downward, and a negative pitch angle indicates that the nose of the aircraft is tilted upward.

[0078] Roll Angle: It refers to the angle between the longitudinal axis of the aircraft body (usually the long axis of the aircraft) and the horizontal plane, and is defined as the angle of rotation of the aircraft body around the longitudinal axis. When one side of the aircraft wing tilts downward, it is a positive roll angle; when one side of the aircraft wing tilts upward, it is a negative roll angle.

[0079] When abnormal values occur in the movement, the abnormal values include falls, limping, and dragging movement trajectories, and abnormal jump values are detected through a sliding window.

[0080] First, calculate the window statistic, and the formula is

[0081]

[0082] where N is the window length, and x t-k is the current t - k point.

[0083] According to the sampling rate selection, such as a sampling rate of 100Hz, the window length N = 10, corresponding to 0.1 seconds of data.

[0084] Then, determine the jump value. If at the current point |x t - μ window | > k·σ window , it is marked as a jump value, where k takes the value of 3 or 2.5. Reduce the influence of variant values on the recognition of the user's movement trajectory, improve the recognition accuracy, and improve the computing power accuracy.

[0085] Use the acceleration integration method to calculate the movement speed, and reset the data in time when touching the ground to eliminate the drift error. The formula is: a motion = a world - [0, 0, g] T , where a motion is the gravity component in the separated movement of the updated three - axis accelerometer, a world is the gravity component in the separated movement of the initial three - axis accelerometer, and g is the gravitational constant.

[0086] Perform a Fourier transform operation on the acceleration signal on the z - axis to extract the main frequency. The formula is where X(f) is the acceleration signal in the vertical direction.

[0087] Calculate the acceleration vector sum, extract the movement peak value, and analyze the rotational angular velocity of the three - axis gyroscope to obtain the amplitude of the foot swing angle. a motion is the acceleration amplitude, and ω is the angular velocity amplitude.

[0088] After identifying the motion posture and motion pattern, the SVM algorithm is used to detect abnormal states. The SVM algorithm is a support vector machine (Support Vector Machine, SVM), which is a type of generalized linear classifier that performs binary classification on data in a supervised learning manner.

[0089] The specific operation steps of the algorithm are as follows:

[0090] (1) Data collection

[0091] (2) Data annotation:

[0092] Normal gait: Marked as class 0 (such as walking, running).

[0093] Abnormal events: Marked as class 1 (such as falling, limping, shuffling).

[0094] Synchronous recording: Ensure the correspondence between data and actions by video recording or manually marking timestamps.

[0095] (3) Dataset division

[0096] Training set: 70% - 80% of the data, used for model training.

[0097] Validation set: 10% - 15% of the data, used for parameter tuning.

[0098] Test set: 10% - 15% of the data, used for final evaluation.

[0099] (4) Data processing and SVM model training

[0100] Select the kernel function, tune the parameters, and use the pre-packaged SVM tool provided by the machine learning module scikit-learn developed based on the Python language for model training.

[0101] (5) Model evaluation and deployment

[0102] Evaluate by calculating precision and recall.

[0103] Export the trained SVM model parameters and deploy them to the mobile app side, or consider directly deploying them to the device side. This enables the system to quickly establish a model and make a preliminary estimate under the condition of small data volume or limited computing resources, and find the optimal regression coefficients and intercepts by minimizing the error between the predicted value and the true value, thereby improving the estimation accuracy of foot movement.

[0104] It also includes a voice vibration reminder module, which is used to encourage the user verbally when the user's motion state is normal and continuous, and to vibrate and issue a warning when in an abnormal state;

[0105] The exercise report generation module is used to form a visual report, which includes the exercise compliance and the standard degree of exercise postures.

[0106] The rehabilitation plan generation module is used to generate the daily exercise pattern and time arrangement.

[0107] The beneficial effects are as follows: Real-time monitoring and guidance: By collecting the exercise state data in real time through the acceleration sensor and gyroscope, and using the artificial intelligence large model at the edge side for analysis and processing, it can provide real-time guidance information such as exercise postures, time, and frequency for patients, ensuring that patients maintain the correct exercise state during the rehabilitation training process and improving the rehabilitation effect.

[0108] Personalized rehabilitation plan: According to the exercise data and rehabilitation progress of patients, a personalized rehabilitation plan is customized for patients to meet the rehabilitation needs of different patients and improve the rehabilitation efficiency.

[0109] Lightweight operation: The wearable device is designed to be lightweight and convenient to wear, without affecting the normal activities of patients, improving the usage experience of patients.

[0110] Reducing the rehabilitation cost: Reducing the dependence on medical staff, lowering the labor cost, and at the same time improving the accessibility and convenience of training.

[0111] Embodiment 2:

[0112] The difference between Embodiment 2 and Embodiment 1 is that noise cancellation adopts a filtering algorithm, and the filtering algorithm formula is: y t1 =b0y t +b1y t-1 +b2y t-2 +b n y t-n ,y t2 =αy t-1 +α(y t -y t-1 ), where b n is the parameter value of the nth acceleration data, y t-n is the (t - n)th acceleration data, y t1 is the filtered acceleration data, y t2 is the filtered gyroscope data, and α is the parameter value. This embodiment has a small amount of operation data and a fast response speed.

[0113] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can adopt the form of a computer program product implemented on one or more computer-usable non-transitory storage media containing computer-usable program code.

[0114] The present invention can provide computer program instructions to the processors of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices generate means for implementing the system.

[0115] These computer program instructions can also be stored in a computer-readable memory capable of guiding the computer or other programmable data processing devices to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions of the system.

[0116] These computer program instructions can also be loaded onto the computer or other programmable data processing devices, such that a series of operation steps are executed on the computer or other programmable devices to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable devices provide steps for implementing the functions of the system.

Claims

1. An auxiliary rehabilitation system for postoperative exercise recovery, characterized in that, including a data acquisition module, worn on the ankle or the instep, internally provided with a three-axis accelerometer and / or a three-axis gyroscope, setting an initial sampling frequency, and performing range configuration and static calibration; a data processing module, performing coordinate system conversion, noise elimination, and data calibration on the acquired data, specifically calculating key parameters, and finally executing a Kalman wave algorithm fusion algorithm to obtain a user posture estimation result; a feedback module, used to identify gait characteristics, extract and analyze the user's step length, step frequency, and swing phase through modeling, and feedback motion data such as motion postures, time, and frequency guidance information to the patient, and at the same time customize a personalized rehabilitation plan for the patient; a power supply module, internally provided with a rechargeable lithium battery for power supply, used to provide stable power support; the feedback module determines the motion posture of the user through the data collected by the data acquisition module, and analyzes the posture estimation result of the processing result. The posture estimation result analysis formula is: Where X t represents the state value of the triaxial gyroscope at time t; X t-1 represents the state value of the triaxial gyroscope at time t - 1; ω x represents the swing amplitude of the foot in the x-axis direction; ω y represents the swing amplitude of the foot in the y-axis direction; ω z represents the swing amplitude of the foot in the z-axis direction; Δt represents the time interval; y t represents the noise cancellation factor; Z t represents the state value of the triaxial accelerometer at time t; a x represents the motion acceleration of the foot in the x-axis direction; a y represents the motion acceleration of the foot in the y-axis direction; a z represents the motion acceleration of the foot in the z-axis direction; υ t represents the integral value of the acceleration after removing gravity.

2. The auxiliary rehabilitation system for postoperative exercise recovery according to claim 1, wherein the static calibration includes calculating the offset between the sensor coordinate system and the world coordinate system according to the position where the data acquisition module is worn, recording the static offset of the data acquisition module, and calculating the static posture angle. The formula is as follows: where θ acc is the initial pitch angle of the gravity component when the user is stationary, and Φ acc is the initial roll angle of the gravity component when the user is stationary.

3. The auxiliary rehabilitation system for postoperative exercise recovery according to claim 1, wherein The noise cancellation uses a filtering algorithm, and the formula of the filtering algorithm is: y t1 = b0y t + b1y t-1 + b2y t-2 + b n y t-n , y t2 = αy t-1 + α(y t - y t-1 ), where b n is the parameter value of the nth acceleration data, yt -n is the (t - n)th acceleration data, yt 1 is the filtered acceleration data, yt 2 is the filtered gyroscope data, and α is the parameter value.

4. An auxiliary rehabilitation system for postoperative exercise recovery according to claim 2 or 3, characterized in that, when an abnormal value occurs in the motion, the abnormal value includes a fall, a limp, and a dragging motion trajectory, and an abnormal jump value is detected through a sliding window.

5. An auxiliary rehabilitation system for postoperative exercise recovery according to claim 2 or 4, characterized in that The acceleration integration method is used to calculate the motion speed, and data reset is performed in a timely manner when touching the ground to eliminate the drift error. The formula is: a motion = a world -[0,0,g] T , where a motion is the gravity component during the separation motion in the updated triaxial accelerometer, a world is the gravity component during the separation motion in the initial triaxial accelerometer, and g is the gravitational constant.

6. The auxiliary rehabilitation system for postoperative exercise recovery according to claim 5, characterized in that, performing a Fourier transform operation on the acceleration signal on the z-axis to extract the main frequency.

7. The auxiliary rehabilitation system for postoperative exercise recovery according to claim 6, characterized in that calculating the acceleration vector sum, extracting the motion peak value, and analyzing the rotational angular velocity of the three-axis gyroscope to obtain the foot swing angle amplitude.

8. The auxiliary rehabilitation system for postoperative exercise recovery according to claim 7, characterized in that, utilizing the Kalman wave algorithm and fusing the data of the three-axis accelerometer and the three-axis gyroscope to finally obtain the foot three-dimensional posture data value.

9. An auxiliary rehabilitation system for postoperative exercise recovery according to any one of claims 1 to 8, characterized in that, after identifying the motion posture and the motion mode, using the SVM algorithm to detect the abnormal state.

10. An auxiliary rehabilitation system for postoperative exercise recovery according to any one of claims 1 to 8, characterized in that, also including a voice vibration reminder module, used to encourage the user verbally when the user's motion state is normal and continuous, and vibrate to remind and issue a warning when in an abnormal state; a motion report generation module, used to form a visual report, and the visual report includes the exercise compliance degree and the motion posture standard degree; a rehabilitation plan generation module, used to generate the daily motion mode and time arrangement.

Images