CN-121725985-B - Limb rehabilitation exercise data evaluation system based on wearable equipment

Abstract

The invention relates to the technical field of medical rehabilitation monitoring and data processing, in particular to a limb rehabilitation movement data evaluation system based on wearable equipment, which comprises a phase space reconstruction unit, a phase space analysis unit, a data analysis unit and a data analysis unit, wherein the phase space reconstruction unit is used for fusing inertial data to construct a state vector and mapping to generate a phase space state track; the invention discloses a method for achieving rehabilitation motion quality, which comprises a track deviation analysis unit, a nonlinear stability analysis unit, a multidimensional characteristic fusion evaluation unit, a reference parameter self-adaptive updating unit and a reference parameter self-adaptive updating unit, wherein the track deviation analysis unit is used for projecting point cloud to a standard reference track and quantifying a distribution state to generate a track boundary crossing index, the nonlinear stability analysis unit is used for calculating a maximum Lyapunov index based on track evolution to generate a motion stability classification label, the multidimensional characteristic fusion evaluation unit is used for decoupling and identifying compensation and pathology and triggering real-time feedback, and the reference parameter self-adaptive updating unit is used for updating pipeline parameters by utilizing topological characteristics.

Inventors

• LI ZHIPENG
• YOU YANG
• WANG CAN

Assignees

• 重庆青尔科技有限公司

Dates

Publication Date

20260512

Application Date

20260226

Claims (5)

1. 1. A wearable device-based limb rehabilitation exercise data assessment system, comprising: The phase space reconstruction unit is used for acquiring limb movement time sequence data acquired by wearing equipment, wherein the time sequence data comprises a triaxial acceleration component and a triaxial angular velocity component, calculating the optimal delay time and an embedded dimension of the time sequence data based on a mutual information method and a false adjacent point method, constructing a state vector based on the optimal delay time and the embedded dimension, mapping the time sequence data of a one-dimensional time domain to a high dimension Xiang Kongjian to generate a phase space state track, loading a preset standard reference track feature matrix, wherein the standard reference track feature matrix comprises a pipeline center track line constructed based on healthy limb movement data and a pipeline radius threshold, projecting the phase space state track into a topological space of the standard reference track, detecting the space distribution state of the phase space state track relative to the inner wall of the standard reference track, quantifying the degree that point cloud distribution exceeds the inner wall of the pipeline, and generating a track boundary crossing index; The nonlinear stability analysis unit is used for calculating a maximum Lyapunov index based on the track evolution characteristic of the phase space state track in the phase space, loading a preset dynamic stability threshold interval, mapping the maximum Lyapunov index to a corresponding stability grade and generating a motion stability classification label; The multi-dimensional feature fusion evaluation unit is used for combining the track boundary crossing index and the motion stability classification label, performing multi-dimensional feature fusion calculation, decoupling and identifying the compensation type and pathological state of limb motion, generating a rehabilitation motion quality evaluation report based on the identification result, and triggering a real-time feedback instruction; The method for generating personalized reference track parameters by utilizing the topological features to update the standard reference track feature matrix in a weighting way comprises the steps of calculating the topological similarity between the current phase space state track and the standard reference track feature matrix; Fusing the spatial distribution characteristics of the current phase space state track into the standard reference track characteristic matrix by utilizing the updating weight factors, carrying out smooth correction on a pipeline center track line, and storing the pipeline center track line as the personalized reference track parameter; The method for generating the motion stability classification label comprises the steps of loading a preset dynamic stability threshold interval, mapping the maximum Lyapunov exponent to a corresponding stability grade, loading a preset first entropy threshold and a preset second entropy threshold, wherein the first entropy threshold is larger than the second entropy threshold, judging that a motion state is chaotic and divergent if the maximum Lyapunov exponent is larger than the first entropy threshold, generating a risk early warning label as the motion stability classification label, judging that the motion state is a stable period if the maximum Lyapunov exponent is smaller than or equal to the first entropy threshold and larger than or equal to the second entropy threshold, generating an effective training label as the motion stability classification label, and judging that the motion state is stiff and convergent if the maximum Lyapunov exponent is smaller than the second entropy threshold, and generating an ineffective motion label as the motion stability classification label; The method for decoupling and identifying the compensation type and the pathological state of the limb movement by combining the track boundary crossing index and the movement stability classification label through multidimensional feature fusion and resolving comprises the steps of loading a preset collision tolerance threshold, judging that the limb movement is stable and compensated if the movement stability classification label is an effective training label and the track boundary crossing index is larger than the collision tolerance threshold, judging that the limb movement is standard rehabilitation movement if the movement stability classification label is an effective training label and the track boundary crossing index is smaller than or equal to the collision tolerance threshold, ignoring the track boundary crossing index if the movement stability classification label is a risk early warning label, directly judging that the limb movement is an out-of-control pathological state, judging that the limb movement is a stiff state if the movement stability classification label is an ineffective movement label, generating a rehabilitation movement quality assessment report based on an identification result, and triggering a real-time feedback instruction.
2. 2. The limb rehabilitation movement data evaluation system based on the wearable device according to claim 1, wherein the method for mapping the time sequence data of a one-dimensional time domain to a high dimension Xiang Kongjian and generating a phase space state track based on the optimal delay time and the embedded dimension is characterized by defining the time sequence data as a time sequence set, performing time delay processing on the time sequence set by utilizing the optimal delay time to generate a plurality of delay sequences, intercepting the numerical values of the time sequence set and the delay sequence at the same time according to the embedded dimension to form a high dimension state vector, arranging the high dimension state vector at all times in time sequence to construct a phase space track matrix, and taking the phase space track matrix as the phase space state track.
3. 3. The limb rehabilitation exercise data evaluation system based on the wearable equipment is characterized in that the method comprises the steps of analyzing a standard reference track characteristic matrix, obtaining a pipeline center track line and a pipeline radius threshold value, calculating Euclidean distance from each point in the phase space state track to the pipeline center track line, comparing the Euclidean distance with the pipeline radius threshold value, identifying escape points with the Euclidean distance larger than the pipeline radius threshold value, counting the number of the escape points, calculating the proportion of the number of the escape points to the total point number of the phase space state track, and determining the proportion as the track boundary crossing index.
4. 4. The limb rehabilitation exercise data evaluation system based on the wearable device according to claim 1 is characterized in that the system is integrated in the wearable device, the wearable device comprises a microprocessor, a storage module and a wireless transmission module, the storage module is used for storing the standard reference track feature matrix and various preset threshold parameters, the wireless transmission module is used for sending the generated rehabilitation exercise quality evaluation report and the real-time feedback instruction to an external terminal, and the phase space reconstruction unit performs phase space mapping operation by using a matrix operation accelerator of the microprocessor.
5. 5. The wearable device-based limb rehabilitation movement data evaluation system according to claim 1, wherein the means for triggering the real-time feedback instruction comprises generating a correction prompt instruction for a power generating muscle group in response to determining that the limb movement is a stable compensatory movement, generating a blocking alarm instruction for immediately stopping training in response to determining that the limb movement is a runaway pathological state, generating an excitation instruction for maintaining training intensity in response to determining that the limb movement is a standard rehabilitation movement, and generating a guide instruction for promoting an increase in movement amplitude in response to determining that the limb movement is a stiff invalid state.

Description

Limb rehabilitation exercise data evaluation system based on wearable equipment Technical Field The invention relates to the technical field of medical rehabilitation monitoring and data processing, in particular to a limb rehabilitation exercise data evaluation system based on wearable equipment. Background With the development of rehabilitation medicine and intelligent wearing technology, the demand for quantitative assessment of limb movement functions is growing. However, the limb movement data contains complex nonlinear dynamics characteristics, so that accurate judgment of rehabilitation states faces a plurality of challenges. Currently, threshold decision methods based on basic kinematic parameters are generally employed, or rely on subjective visual observations by therapists. In the prior art, the motion quality is generally measured by utilizing statistics such as amplitude, variance and the like of acceleration or angular velocity, and whether the motion meets the standard is judged by a simple linear index, however, the dynamic structural characteristics behind the motion time sequence are difficult to deeply excavate by the traditional evaluation method. Algorithms based solely on speed or acceleration thresholds often cannot effectively distinguish pathological tremors from stiff convergence states, and can easily lead to erroneous judgment of rehabilitation states. In addition, the traditional algorithm is difficult to identify stable compensatory movements, namely, the situation that a patient generates force by using wrong muscle groups but the track performance is relatively stable is often misjudged as standard rehabilitation actions, so that hidden pathological risks are difficult to find, and the personalized accurate rehabilitation requirement cannot be met. Therefore, how to deeply mine the dynamic characteristics of the motion data, and accurately decouple and identify the compensatory type and pathological state of the limb motion becomes a problem to be solved in the art. Disclosure of Invention In order to solve the technical problems, the invention provides a limb rehabilitation exercise data evaluation system based on wearable equipment, and specifically, the technical scheme of the invention comprises the following steps: The phase space reconstruction unit is used for acquiring limb movement time sequence data acquired by the wearable equipment, wherein the time sequence data comprises a triaxial acceleration component and a triaxial angular velocity component, calculating the optimal delay time and the embedding dimension of the time sequence data based on a mutual information method and a false adjacent point method, constructing a state vector based on the optimal delay time and the embedding dimension, mapping the time sequence data of a one-dimensional time domain to a high dimension Xiang Kongjian, and generating a phase space state track; The track deviation analysis unit is used for loading a preset standard reference track characteristic matrix, wherein the standard reference track characteristic matrix comprises a pipeline center track line constructed based on healthy limb movement data and a pipeline radius threshold value; The nonlinear stability analysis unit is used for calculating a maximum Lyapunov index based on the track evolution characteristics of the phase space state track in the phase space, loading a preset dynamic stability threshold interval, mapping the maximum Lyapunov index to a corresponding stability grade, and generating a motion stability classification label; The multi-dimensional feature fusion evaluation unit is used for combining the track boundary crossing index and the motion stability classification label, and decoupling and identifying the compensation type and pathological state of limb motion through multi-dimensional feature fusion calculation; The self-adaptive updating unit of the reference parameters is used for responding to the rehabilitation movement quality evaluation report to judge that the movement is in an effective rehabilitation state, extracting the topological feature of the current phase space state track, and carrying out weighted updating on the standard reference track feature matrix by utilizing the topological feature to generate personalized reference track parameters. As a further aspect of the present invention, constructing a state vector based on the optimal delay time and the embedding dimension, mapping the time-series data of the one-dimensional time domain to the high dimension Xiang Kongjian, and generating the phase space state trajectory includes: Defining time sequence data as a time sequence set; performing time delay processing on the time sequence set by utilizing the optimal delay time to generate a plurality of delay sequences; according to the embedding dimension, intercepting the numerical value of the time sequence set and the delay sequence at the same time to form a high-dimensional state