Wavelet enhancement cooperative activation method for FES multi-target time sequence regulation and control

Abstract

The invention discloses a wavelet enhancement cooperative activation method for FES multi-target time sequence regulation and control, and relates to the technical field of biomedical engineering. According to the model, an overlapped sliding window is analyzed through a multi-channel surface electromyographic signal, a time-frequency energy significant feature matrix is constructed by combining an energy ratio, and meanwhile, an inter-muscle co-activation matrix is constructed in the multi-channel electromyographic signal sliding window; constructing a multi-channel high-dimension matrix by using each single-channel time-frequency energy significant feature matrix and the multi-channel inter-muscle co-activation matrix, and performing data dimension reduction on the multi-channel high-dimension matrix to obtain a multi-channel multi-dimension time-frequency energy-inter-muscle co-activation matrix; and a muscle activation time sequence vector is screened by using a self-adaptive threshold logic combination rule, and finally an activation time sequence is applied to an FES experiment. According to the method, the movement of the hand on the sagittal plane is accurately predicted by analyzing the multi-channel surface electromyogram signals and the multi-target muscle co-activation characteristics, functional electrical stimulation FES parameters are dynamically adjusted, and accurate control over the muscle activity stage is achieved.

Description

Technical Field

[0001] The invention relates to the technical field of biomedical engineering, and in particular to a wavelet enhanced synergistic activation method for FES multi-target timing regulation. Background Art

[0002] Functional electrical stimulation (FES) has shown great potential as an emerging rehabilitation treatment. FES technology uses a low-frequency pulse current of a certain intensity to stimulate one or more groups of muscles through a pre-set program to induce muscle movement or simulate normal voluntary movement. As a technology, FES can be used as a neural prosthesis to restore motor function or as a therapeutic intervention for post-stroke rehabilitation. FES can improve motor performance and reduce muscle loss. By stimulating to overcome the limitations of disabled limbs, it can restore motor function in patients with spinal cord injury, Parkinson's disease, cerebral palsy and stroke. FES can reduce muscle degeneration during inactivity. FES-induced repetitive exercises of single-joint movements, such as flexion and extension, play a vital role in preventing the degeneration of muscle fibers supporting the joints and are used to treat patients with spinal cord injury and stroke. In addition, through FES, damaged nerves and muscles can be stimulated to promote the recovery of neuromuscular function and improve patients' motor ability and quality of life.

[0003] By precisely controlling the current stimulation parameters, FES can simulate the natural stimulation of the nervous system, promote muscle contraction and the recovery of neuromuscular function, thereby helping patients restore their motor ability and improve their quality of life. As an advanced neuromuscular activation method, by adjusting specific stimulation parameters (such as pulse frequency, pulse width, and current intensity, etc.), it can indeed effectively stimulate muscles to produce movement, but there are still many challenges in accurately controlling the coordinated movements of various muscles and joints in the hand. The hand, as one of the most delicate and complex organs in the human body, relies on the coordinated work of many muscles and nerves. Every subtle movement, such as flexion and extension of fingers, rotation of wrists, and clenching of palms, requires these muscles and nerves to work together at a specific intensity at a precise time point. This complexity and delicacy is both a challenge and an opportunity for FES technology. At present, FES technology stimulates muscles to produce movement by adjusting stimulation parameters (such as pulse frequency, pulse width, current intensity, etc.). However, when simulating natural hand movements, the adjustment of these parameters is often difficult to achieve the desired effect. Because the neuromuscular signals of hand movements are complex and changeable, and the stimulation patterns that FES technology can provide are often relatively single. This results in that when simulating natural hand movements, FES technology often cannot accurately capture all neuromuscular signals, making it impossible to achieve precise coordinated movements between the muscles and joints of the hand.

[0004] Most existing FES systems use open-loop control. In an open-loop control system, the stimulation parameters (such as stimulation amplitude, frequency, waveform, etc.) of the functional electrical stimulation output are fixed, and these parameters depend on the experience of the rehabilitation instructor. Since a fixed pulse sequence is used for stimulation, excess charge may remain, causing muscle fatigue, or insufficient stimulation intensity may make it difficult for the muscles to contract accordingly to complete the prescribed training exercise. In addition, in an open-loop control system, it is difficult to achieve the ideal rehabilitation effect because of the lack of real-time feedback and precise control of the patient's actual performance (such as trajectory, angle, angular velocity, joint torque and other parameters) during the rehabilitation process. Summary of the invention

[0005] The purpose of the present invention is to propose a wavelet-enhanced synergistic activation method for FES multi-target timing regulation, to construct a comprehensive muscle activity analysis framework by integrating time-frequency analysis and muscle synergy characteristic quantification, and to more accurately predict the muscle activation trend by dynamically adjusting the energy threshold index, so as to achieve precise regulation of the hand's upper limb movement in the sagittal plane.

[0006] To achieve the above purpose, the present invention proposes a wavelet-enhanced synergistic activation method for FES multi-target timing regulation, and the specific steps are as follows:

[0007] S1, collecting multi-channel surface electromyographic signals and bending angles, and preprocessing the collected electromyographic signals;

[0008] S2, performing wavelet denoising and reconstruction on the electromyographic signals of each channel preprocessed in S1; performing two continuous wavelet transforms on the reconstructed electromyographic signals of each channel, and adaptively obtaining the time-frequency energy significant feature matrix by combining the energy proportion and the overlapping sliding average window;

[0009] S3, combining the multi-channel electromyographic signals preprocessed in S1 in pairs to construct multiple groups of muscle pairs, aligning and segmenting the electromyographic signals of each group of muscle pairs using overlapping sliding average windows to obtain multiple electromyographic signal pairs, and calculating multi-faceted co-activation feature indexes for each pair of electromyographic signals to obtain an inter-muscle co-activation distribution feature matrix;

[0010] S4, constructing each single-channel time-frequency energy significant feature matrix in S2 and the muscle co-activation matrix in S3 into a multi-channel high-dimensional matrix, and at the same time reducing the dimension by accumulating the variance contribution rate to obtain a multi-dimensional time-frequency energy-muscle co-activation matrix;

[0011] S5, for each component of the multi-dimensional time-frequency energy-muscle co-activation matrix of each channel in S4, a threshold is adaptively set according to the statistical probability, and an adaptive threshold logic combination rule is proposed to optimize and screen out the corresponding muscle activation state time series vector;

[0012] S6. Involve the activation timing vectors of multiple muscles in the stimulation parameter design of FES experiments, and conduct corresponding experiments.

[0013] Preferably, in S1, an experimental design is performed in sagittal plane movement, and during the hand clenching movement, multi-channel surface electromyographic signals of the forearm muscles and the finger bending angle are synchronously collected, and the collected multi-channel electromyographic signals are preprocessed.

[0014] Specifically, the surface electromyographic signals of the ulnar extensor carpi, extensor digitorum, superficial flexor digitorum, palmaris longus, and radial flexor carpi of the right side of a normal person are collected, and the curvature of the right index finger is collected at the same time. The collected EMG is band-pass filtered and notch filtered according to the breakpoints of different angle values ​​of the action to obtain the pre-processed electromyographic signals (V1, .., Vi).

[0015] Preferably, in S2, each single-channel electromyographic signal after preprocessing in S1 is subjected to a 4th-order Daubechies 6-level decomposition wavelet decomposition, and a threshold of each signal is set based on the signal length and a logarithmic function to remove noise, and then the wavelet coefficients of each frequency band are reconstructed to obtain a multi-channel reconstructed signal (v1, v2,…, vi); wherein vi is the reconstructed signal.

[0016] Specifically, it includes: using the 4th-order Daubechies wavelet (Db4) to perform 6-level wavelet decomposition to extract the time-frequency characteristics of the electromyographic signal (V1,..,Vi). This process divides the signal into an approximate part (low frequency) and a detail part (high frequency), and further decomposes the approximate part. After decomposition, the high-frequency detail coefficient is suppressed or removed by calculating the threshold, and then the wavelet basis function is used for inverse convolution to reconstruct the signal (v1,v2,…,vi).

[0017] Preferably, the multi-channel reconstructed signal (v1, v2, ..., vi) is scaled from 1 to 128 to obtain the overall coefficient distribution C of each channel electromyographic signal i (a,b), by calculating each C i The square of (a, b) is used to obtain the reconstructed energy, and then the time-frequency scale energy matrix E of the entire signal is obtained. i (a, b); on this basis, calculate the energy ratio, that is, calculate each energy distribution E i The proportion of each scale of (a, b) in the total energy is determined, and it is determined that when the energy proportion exceeds 10%, it is significant, thereby achieving m i Adaptive selection of energy-significant scales; where a is the scale parameter, b is the location parameter, and m i Select m energy-significant scales for the i-th scale;

[0018] Preferably, an overlapping sliding average window is used to divide the time scale of each channel electromyographic signal vi in ​​the multi-channel reconstructed signal (v1, v2, ..., vi) and obtain n electromyographic signal segments, and m i The energy significant scale is used as the scale selection standard after the continuous wavelet transform of the electromyographic signal after sliding window n. The energy value of each significant scale after each window is used to construct the time-frequency energy significant feature matrix (m i ×n), where n is an integer.

[0019] Preferably, in S3, multiple groups of muscle pairs are constructed by pairwise combination of the preprocessed multi-channel reconstructed signals (v1, v2, ..., vi) in S2, and overlapping sliding average windows n are performed on the electromyographic signals of each group of muscle pairs, and multi-level co-activation feature indicators of each pair of electromyographic signals after each window are calculated, including the overlap degree cci of the muscle pair combination, the time point correlation tc, and the information exchange degree cmi, so as to obtain an inter-muscle co-activation distribution feature matrix (cci; tc; cmi)×n containing all combinations.

[0020] Specifically, after the preprocessed signals (v1, v2, ..., vi) are overlapped and averaged over windows n, the common contraction index cci1, ..., cci between muscles after each window is calculated. a , time correlation tc1,..,tc b and cross mutual information cmi1,…,cmi c , these characteristic values ​​quantify the degree of co-activation and information exchange between the EMG envelopes of two muscles at a time point, as the co-activation distribution feature matrix of vi (cci; tc; cmi)×n.

[0021] Preferably, in S4, each single-channel time-frequency energy significant feature matrix (m i ×n) and the intermuscular co-activation matrix (cci; tc; cmi) in S3 to obtain the high-dimensional data matrix (cci; tc; cmi; m i )×n, extract the most important features based on the correlation between variance and eigenvectors, select the principal component with a cumulative feature variance contribution rate of 95%, and reduce the high-dimensional data matrix to a multi-dimensional time-frequency energy-muscle co-activation matrix k i ×n, the multidimensional time-frequency energy-muscle co-activation matrix will be used as the energy threshold indicator; where k i The row number index.

[0022] Specifically, the reconstructed energy values ​​and co-activation characteristic values ​​of the entire signal of multiple muscles (v1, v2, ..., vi) are integrated into a multidimensional data set to obtain the high-dimensional data matrix (cci; tc; cmi; m i )×n, each row represents a time point, each column represents a feature, including energy value, co-activation characteristic value, calculate the covariance matrix of the standardized data for multidimensional data to determine the correlation between the variables, solve the eigenvalues ​​and corresponding eigenvectors of the covariance matrix, and select the principal component with a cumulative variance contribution rate of more than 95% according to the size of the eigenvalue. i The principal component constructs the eigenvector to obtain the multi-dimensional time-frequency energy-intermuscle co-activation matrix k i ×n, each eigenvector corresponds to the muscle activity state at a time point.

[0023] Preferably, in S5, for each channel in S4, the multi-dimensional time-frequency energy-muscle co-activation matrix k i Each component k of ×n i , the threshold is adaptively set according to the maximum value of the probability density; in order to provide additional principal components related to muscle activation, considering the simultaneous influence of multiple principal components, a threshold logic combination rule is proposed to perform activation judgment on the n points of the multi-dimensional time-frequency energy-intermuscle co-activation matrix to judge the muscle activation state. Only when the value of each principal component exceeds their respective thresholds at the same time, the muscle is considered to be activated in a specific time period, thereby obtaining the activation timing of different channels g1,..,g i ; Among them, g i is the activation timing vector of the ith channel.

[0024] Specifically, the principal component is a linear combination of the original data set, representing the main variation direction in the data. The first principal component usually contains the most variation information, so it can be used as a strong indicator to monitor changes in muscle activation status. In addition to the first principal component, other principal components can also be used to set thresholds, especially when they provide additional information related to muscle activation. The multidimensional matrix k of each channel i Each component k of ×n i The threshold is adaptively set according to the maximum value of the probability density. Each principal component can have its own threshold. These thresholds work together to determine the muscle activation state. By proposing an adaptive threshold logic combination rule, multiple principal components are considered at the same time. The logic rule is that only when the values ​​of multiple principal components exceed their respective thresholds at the same time, the muscle is considered to be activated, and the activation timing n is obtained. i , reconstruct the signal (v1, v2, …, vi) to obtain the activation timing of different channels g1, .., g i .

[0025] Preferably, in S6, the electromyographic electrode is fixedly attached to the relevant muscle area of ​​the patient's left arm, and the electrical stimulation electrode is kept at a set position, and the electromyographic acquisition device collects surface electromyographic signals for subsequent analysis, and the activation timing of multiple muscles g1, .., g i The stimulation parameters designed for the FES experiment, including the frequency, amplitude and pulse width of the current, stimulate the muscles on the affected side and enable the patient's affected hand to complete a fist movement.

[0026] Therefore, the present invention proposes a wavelet-enhanced synergistic activation method for FES multi-target timing regulation, and its beneficial effects are as follows:

[0027] (1) The present invention is a wavelet-enhanced synergistic activation method for FES multi-target timing regulation. By integrating time-frequency analysis and muscle synergy characteristic quantification, a comprehensive muscle activity analysis framework is constructed, which dynamically adjusts the energy threshold index and can flexibly adapt to changes in muscle activity, thereby more accurately predicting muscle activation trends.

[0028] (2) The present invention provides a wavelet-enhanced co-activation method for FES multi-target timing regulation. Based on the activation timing of different channels, it deepens the understanding of muscle co-activation patterns, thereby being able to accurately determine the co-activation state of muscles within a specific time period, providing more comprehensive and rich muscle activity information, while improving the accuracy of muscle activation judgment and providing a new perspective for the quantitative analysis of muscle co-activation.

[0029] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a flow chart of a preferred embodiment of the functional electrostimulation control strategy based on multi-target timing regulation of the present invention;

[0031] Figure 2 Schematic diagram of the collection site of the forearm of the dominant hand of the present invention;

[0032] Figure 3 Schematic diagram of the finger curvature acquisition module of the present invention; wherein a is a schematic diagram of the finger curvature acquisition device, and b is a schematic diagram of the finger curvature data acquisition;

[0033] Figure 4 This is a display diagram of the visualization interface for collecting finger curvature of the present invention;

[0034] Figure 5 It is a schematic diagram of the muscle timing activation segment of the present invention. DETAILED DESCRIPTION

[0035] In order to make the technical solutions, advantages and purposes of the present invention clearer, the technical solutions of the embodiments of the present invention are clearly and completely described below. The described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the protection scope of this application.

[0036] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0037] like Figure 1 As shown, the present invention provides a wavelet enhanced synergistic activation method for FES multi-target timing regulation, and the specific steps are as follows:

[0038] S1. Get the data set. The electromyography acquisition device is delsys, and its sampling frequency is 2000HZ. Figure 3 As shown in a, the bending angle is collected by the Flex sensor. Before the experiment, the skin surface is wiped with alcohol to increase conductivity and reduce interference. Figure 2 As shown, the electrodes of the electromyographic acquisition device and the functional electrical stimulation electrodes are attached to the right extensor carpi ulnaris, extensor digitorum, flexor digitorum superficialis, palmaris longus, and flexor carpi radialis of a normal person. Figure 4 As shown, at the beginning of the experiment, the electromyography acquisition and the bending degree are synchronously acquired and the action start point and the maintenance stage point are recorded, as shown in FIG. Figure 3 As shown in b, first, the experimental subject wears a curvature glove and lets his arm hang naturally with his palm relaxed. The surface electromyographic signal of the user is collected when he is at rest for 20 seconds. Then the experimental subject begins to grasp with slight force, which is displayed through the finger curvature collection visualization interface. The modes of curvature 0 to 30 degrees, 0 to 60 degrees, and 0 to 90 degrees are respectively used as a group. Each group performs 40 actions, and the experimental subject maintains the current action angle for a period of time after reaching the angle. According to the different angle values ​​of the action as breakpoints, the collected EMG is band-passed (cut-off frequency is 15Hz~250Hz) and notch filtered (frequency is 50Hz) to obtain the pre-processed electromyographic signal (V1,..,Vi).

[0039] S2. The 4th-order Daubechies (6-level decomposition) is selected as the mother wavelet for the preprocessed EMG signal (V1, .., Vi) to extract the time-frequency features. The wavelet decomposition process starts with convolving the original EMG signal with the Db4 wavelet basis function to extract the approximate and detail components of the signal. In each level of decomposition, the signal is divided into an approximate part (containing low-frequency information) and a detail part (containing high-frequency information). The approximate part is further decomposed. After 6 levels of decomposition, a series of wavelet coefficients in different frequency bands are obtained. At the same time, the threshold of each signal is calculated based on the signal length and the logarithmic function, the high-frequency detail coefficients are removed, and then the inverse process of the wavelet decomposition is reconstructed. For the wavelet coefficients of each frequency band, the corresponding wavelet basis function is used for inverse convolution to reconstruct the signal (v1, v2, …, vi).

[0040] S3. Perform continuous wavelet transform on the reconstructed signal (v1, v2, …, vi). The scale of the transform ranges from 1 to 128 to capture the time variation of different frequency components. It provides the coefficient distribution of the signal at different scales. The distribution is usually represented by a three-dimensional image, where two dimensions are time t and frequency 1 / a, and the third dimension is the amplitude C of the coefficient. i (a,b), by calculating each C i The square of (a, b) is used to obtain the reconstructed energy, and then the time-frequency scale energy matrix E of the entire signal is obtained. i (a,b); the continuous wavelet transform formula is as follows:

[0041]

[0042] Among them, W f (x) is the result of the wavelet transform, which is two-dimensional and depends on the scale a and the position b. f(x) is the original signal; ψ(X) is the wavelet function, usually called the mother wavelet, ψ(X) * is the complex conjugate of the wavelet function; a is the scale parameter, which controls the expansion of the wavelet and is inversely proportional to the frequency; b is the position parameter, which indicates the position of the wavelet in the signal; τ is the center position of the wavelet function.

[0043] S4. The time dimension t in the two dimensions is divided into time windows, and the overlapping sliding average window method is used to provide time continuity, which helps to capture signals that cross the window boundaries. The time window size is set to 50 sampling points, and the moving step is set to 25 sampling points. The sliding window can be moved on the time axis to cover each time window to calculate the eigenvalues, and the sliding window takes into account the impact of signal continuity on the eigenvalue calculation.

[0044] S5, based on the energy E of the reconstructed signal vi in ​​S3 i (a,b), for each scale a, calculate the energy E over the entire time seriesi (a), according to the calculated E i (a), on this basis, the energy proportion is calculated to identify the energy-significant scale in the energy distribution. According to each energy distribution E i Each scale of (a, b) accounts for the total energy, and it is determined that when the energy share exceeds 10%, it is significant. Select m i Then, according to the time window divided by S4, the overlapping sliding average window is used to select a window size of 50 sampling points and a moving step of 25 sampling points. i , starting from the time dimension t, apply the first window (1-50), calculate the energy value distribution of all sampling points in the window for each selected scale, move the window forward 25 sampling points (26-75), calculate the energy value in the window again, and repeat this process until the entire time dimension is covered, and obtain E(b1) to E(bm) for each selected scale i ).

[0045] S6, after overlapping sliding average window n of the preprocessed whole segment signals (v1, v2, …, vi) between multiple muscles in S5, calculate the co-activation characteristic value of each time window, calculate the co-activation characteristic value between the five muscles, consider each pair of muscle combinations, and use the combination formula Given, represents the number of ways to select 2 muscles from e muscles, which means that there are The muscle combination needs to be calculated; the specific calculation process is as follows:

[0046] S61, calculate the co-contraction index, calculate the normalized envelope of the EMG signal envelopes of the two electromyograms in the sliding window, and then calculate the sum of the point-by-point minimum values ​​to obtain cci1,…,cci a , and its calculation formula is as follows:

[0047]

[0048] Among them, x norm and norm The CCI is the envelope normalized by the average of 98-99% of the respective antagonist EMG envelopes; the CCI quantifies the degree of overlap of the two muscle envelopes, with higher values ​​indicating a greater degree of co-contraction;

[0049] S62, calculate the time correlation, for the EMG signal envelopes of the two electromyograms in the sliding window, calculate the average value of their products at all sample points, and obtain tc1, .., tc b , and its calculation formula is as follows:

[0050]

[0051] Among them, x and y are the envelopes of the respective antagonistic EMG signals;

[0052] S63, calculate the cross mutual information, quantify the degree of information exchange between the myoelectric envelopes of two different muscles, and obtain cmi1,…,cmi c , and its calculation formula is as follows:

[0053] I(x,y,τ)=H(x)+H(y)-H(x,y,τ);

[0054] Where H represents Shannon entropy, H(x) and H(y) come from the discrete probability distribution of x(n) and y(n), and H(x, y, τ) is the entropy function of the joint probability distribution based on the time lag τ;

[0055] These characteristic values ​​quantify the degree of co-activation and information exchange between the EMG envelopes of two muscles at a given time point.

[0056] S7. Construct a multidimensional data set and construct a feature vector. The specific steps are as follows:

[0057] S71, convert the energy value E(b1) to E(bm) of each selected scale calculated by reconstructing the signal vi in ​​S5 i ) and the co-activation characteristic values ​​calculated in S6 were integrated into a multidimensional data set, where each row represents a time point and each column represents a feature, including energy value and co-activation characteristic value, to obtain a high-dimensional data matrix (cci; tc; cmi; m i )×n;

[0058] S72, standardizing each feature in the multidimensional data set to ensure that each feature has zero mean and unit variance;

[0059] S73, calculating the covariance matrix of the data standardized in S72 to determine the correlation between the variables;

[0060] S74, solve the eigenvalues ​​and corresponding eigenvectors of the covariance matrix in S73. The eigenvalues ​​represent the variance contribution of each principal component, while the eigenvectors represent the direction of the principal component. According to the size of the eigenvalues, the principal component with a cumulative contribution rate of 95% is selected, and the k of the selected ri is used. i The principal components construct feature vectors, each of which corresponds to the muscle activity state at a time point V = [v1, v2, ..., vk i ].

[0061] S8. Draw a histogram of the eigenvalues ​​of each principal component. The eigenvalues ​​of the principal component obey the standard normal distribution; Figure 5As shown in the figure, the threshold can be defined as the interval with the highest frequency of data occurrence. Once the threshold of each principal component is determined, the following logical rule is proposed to determine whether the muscle is activated at a specific time point. The selected k i If the values ​​of the principal components all exceed their thresholds, the muscle is considered to be activated at that time point, and the activation time sequence n is obtained. i , reconstruct the signal (v1, v2, …, vi) to obtain the activation timing of different channels g1, .., g i .

[0062] S9, the electromyographic electrode is fixedly attached to the relevant muscle area of ​​the patient's left arm, and the electrical stimulation electrode is kept at the set position. The electromyographic acquisition device collects the surface electromyographic signal for subsequent analysis, and converts the activation timing of multiple muscles in S8 g1, .., g i The stimulation parameters designed for the FES experiment include the frequency, amplitude and pulse width of the current to stimulate the muscles on the affected side and enable the patient's affected hand to complete a fist-making action.

[0063] Therefore, the present invention provides a wavelet-enhanced co-activation method for FES multi-target timing regulation. By integrating time-frequency analysis and muscle synergy characteristic quantification, a comprehensive muscle activity analysis framework is constructed, which not only deepens our understanding of muscle co-activation patterns, but also can flexibly adapt to changes in muscle activity by dynamically adjusting the energy threshold index, thereby more accurately predicting muscle activation trends.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A wavelet-enhanced synergistic activation method for FES multi-target timing regulation, characterized in that: The specific steps are as follows: S1, collecting multi-channel surface electromyographic signals and bending angles, and preprocessing the collected electromyographic signals; S2, performing wavelet denoising and reconstruction on the electromyographic signals of each channel preprocessed in S1; The reconstructed electromyographic signals of each channel are transformed twice continuously, and the time-frequency energy significant feature matrix is ​​adaptively obtained by combining the energy proportion and overlapping sliding average window. S3, combining the multi-channel electromyographic signals preprocessed in S1 in pairs to construct multiple groups of muscle pairs, aligning and segmenting the electromyographic signals of each group of muscle pairs using overlapping sliding average windows to obtain multiple electromyographic signal pairs, and calculating multi-faceted co-activation feature indexes for each pair of electromyographic signals to obtain an inter-muscle co-activation distribution feature matrix; S4, constructing each single-channel time-frequency energy significant feature matrix in S2 and the muscle co-activation matrix in S3 into a multi-channel high-dimensional matrix, and at the same time reducing the dimension by accumulating the variance contribution rate to obtain a multi-dimensional time-frequency energy-muscle co-activation matrix; S5, for each component of the multi-dimensional time-frequency energy-muscle co-activation matrix of each channel in S4, a threshold is adaptively set according to the statistical probability, and an adaptive threshold logic combination rule is proposed to optimize and screen out the corresponding muscle activation state time series vector; S6. Involve the activation timing vectors of multiple muscles in the stimulation parameter design of FES experiments, and conduct corresponding experiments.

2. A wavelet enhanced synergistic activation method for FES multi-target timing control according to claim 1, characterized in that: In S1, an experiment design was conducted in sagittal plane motion. During the hand clenching motion, multi-channel surface electromyographic signals of the forearm muscles and the finger bending angle were synchronously collected, and the collected multi-channel electromyographic signals were preprocessed.

3. The wavelet enhanced synergistic activation method for FES multi-target timing control according to claim 1, characterized in that: In S2, each single-channel electromyographic signal preprocessed by S1 is decomposed by a 4th-order Daubechies 6-level decomposition wavelet, and the threshold of each signal is set based on the signal length and logarithmic function to remove noise. Then, the wavelet coefficients of each frequency band are reconstructed to obtain a multi-channel reconstructed signal (v1, v2,…, vi); where vi is the reconstructed signal.

4. The wavelet enhanced synergistic activation method for FES multi-target timing control according to claim 3, characterized in that: The multi-channel reconstructed signal (v1, v2, ..., vi) is scaled from 1 to 128 to obtain the overall coefficient distribution C of each channel electromyographic signal. i (a,b), by calculating each C i The square of (a, b) is used to obtain the reconstructed energy, and then the time-frequency scale energy matrix E of the entire signal is obtained. i (a, b); on this basis, calculate the energy ratio, that is, calculate each energy distribution E i The proportion of each scale of (a, b) in the total energy is determined, and it is determined that when the energy proportion exceeds 10%, it is significant, thereby achieving m i Adaptive selection of energy-significant scales; where a is the scale, b is the position, and m i Select m energy-significant scales for the i-th scale.

5. The wavelet enhanced synergistic activation method for FES multi-target timing control according to claim 4, characterized in that: The overlapping sliding average window is used to divide the time scale of each channel electromyographic signal vi in ​​the multi-channel reconstructed signal (v1, v2, ..., vi) and obtain n electromyographic signal segments. i The energy significant scale is used as the scale selection standard after the continuous wavelet transform of the electromyographic signal after sliding window n. The energy value of each significant scale after each window is used to construct the time-frequency energy significant feature matrix (m i ×n), where n is an integer.

6. The wavelet enhanced synergistic activation method for FES multi-target timing control according to claim 1, characterized in that: In S3, multiple groups of muscle pairs are constructed by pairwise combination of the preprocessed multi-channel reconstructed signals (v1, v2, …, vi) in S2, and overlapping sliding average windows n are performed on the electromyographic signals of each muscle pair. The multi-level co-activation feature indicators of each pair of electromyographic signals after each window are calculated, including the overlap degree cci of the muscle pair combination, the time point correlation tc, and the information exchange degree cmi, and then a muscle co-activation distribution feature matrix (cci; tc; cmi)×n containing all combinations is obtained.

7. The wavelet enhanced synergistic activation method for FES multi-target timing control according to claim 1, characterized in that: In S4, each single channel time-frequency energy significant feature matrix (m i ×n) and the myointerstitial coactivation matrix in S3 (cci; tc; cmi)×n to construct the high-dimensional data matrix (cci; tc; cmi; mi)×n, extract the most important features based on the correlation between variance and eigenvectors, select the principal component with a cumulative feature variance contribution rate of 95%, and reduce the high-dimensional data matrix to a multi-dimensional time-frequency energy-intermuscle co-activation matrix k i ×n, the multidimensional time-frequency energy-muscle co-activation matrix will be used as the energy threshold indicator; where k i The row number index.

8. The wavelet enhanced synergistic activation method for FES multi-target timing control according to claim 1, characterized in that: In S5, the multidimensional time-frequency energy-muscle co-activation matrix k for each channel in S4 i Each component k of ×n i , the threshold is adaptively set according to the maximum value of the probability density; the combination rules are formulated for the principal components that provide additional information related to muscle activation. Considering the simultaneous influence of multiple principal components, the threshold logic combination rules are proposed to perform activation judgment on the n points of the multi-dimensional time-frequency energy-intermuscle co-activation matrix to judge the muscle activation state. Only when the value of each principal component exceeds their respective thresholds at the same time, the muscle is considered to be activated in a specific time period, thereby obtaining the activation timing of different channels g1,..,g i ; Among them, g i is the activation timing vector of the i-th channel.

9. The wavelet enhanced synergistic activation method for FES multi-target timing control according to claim 1, characterized in that: In S6, the activation timing of multiple muscles g1,..,g i The stimulation parameters designed for the FES experiment, including the frequency, amplitude and pulse width of the current, stimulate the muscles on the affected side and enable the patient's affected hand to complete a fist movement.

Images