A brain and myoelectric information synchronization system and method for accurately capturing movement intention

Abstract

The present invention discloses a brain-electromyography information synchronization system and method for accurately capturing movement intention, including a steady-state force acquisition module for acquiring the lower limbs of a subject to determine a target force output interval, a brain-electromyography acquisition module for synchronously acquiring the user's EEG signals and EMG signals and having a signal wireless transmission function, a signal preprocessing module for filtering artifacts of EEG signals and EMG signals to improve the signal-to-noise ratio of brain-electromyography signals, and a brain-electromyography information synchronization module for fitting brain-electromyography data using a multivariate model and performing brain-electromyography coherence analysis and brain-electromyography time delay quantification. The present invention uses a multivariate model to fit brain-electromyography data, and performs time delay quantification of brain-electromyography movement intention information based on brain-electromyography coherence analysis, thereby achieving movement intention information synchronization of brain-electromyography signals by removing delay, laying a solid theoretical foundation for brain-electromyography fusion, and improving the effectiveness of the method for identifying human movement intention by brain-electromyography fusion and the robustness of the intention recognition system.

Description

Technical Field

[0001] The present invention belongs to the field of biomechanical and electrical fusion, and relates to a brain and myoelectric information synchronization system and method for accurately capturing movement intention. Background Art

[0002] The existing perception technologies of human lower limb movement intention are mainly divided into three categories according to the different types of signal sources: recognition methods based on mechanical signals, recognition methods based on bioelectric signals, and recognition methods based on mixed signal sources. The recognition method based on mixed signal sources can better make up for the problems of robustness and low accuracy in the recognition process of a single signal source by introducing multiple signal sources for recognition. Among the various signal sources, mechanical signals cannot truly reflect human movement intentions due to their obvious lag; EEG signals are the direct reflection of the brain nerve center on the scalp, and have the characteristics of weakness, aliasing, low signal-to-noise ratio, and predictability; electromyographic signals are closely related to movement, and have the characteristics of high signal-to-noise ratio, high spatial resolution, and unobvious predictability. Since the characteristics of brain electromyographic signals are obviously complementary, the use of a new type of hybrid BCI (Brain Computer Interface) such as brain electromyographic fusion to realize the perception of human movement intention can not only make up for the defects of using only EEG or electromyographic signals for recognition, but also integrate the characteristics of the two to improve the recognition accuracy, and achieve high accuracy and rapid response in the perception of human lower limb movement intention.

[0003] At present, most of the research on hybrid BCI such as brain-electromyography fusion is carried out at the level of brain-electromyography fusion, focusing on the advantages and disadvantages of brain-electromyography fusion recognition of human movement intention at the feature level or decision level. However, in the process of generating human movement intention, due to the inconsistency of the conduction pathway, the movement intention information contained in the brain-electromyography is not synchronized. The movement intention in the brain-electromyography is always ahead of the electromyography. This time delay leads to theoretical defects in the direct use of synchronously collected brain-electromyography for fusion recognition, and it is impossible to accurately capture the human movement intention. The information asynchrony of brain-electromyography will reduce the effectiveness of the method based on brain-electromyography fusion recognition of human movement intention and the reliability of the system, which greatly limits the development of brain-electromyography fusion recognition technology. There is an urgent need for a breakthrough in the method of brain-electromyography information synchronization to improve brain-electromyography synchronization. At present, a large number of studies have shown that there is a strong brain-electromyography coherence phenomenon in steady-state force contraction or isometric contraction of muscles. The synchronously collected brain-electromyography data can be used to quantitatively evaluate the correlation of brain-electromyography to achieve brain-electromyography synchronization. Summary of the invention

[0004] The purpose of the present invention is to solve the problems in the prior art and to provide a brain and myoelectric information synchronization system and method for accurately capturing movement intention.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A method for synchronizing brain and myoelectric information for accurately capturing movement intentions comprises the following steps:

[0007] Collect the plantar MVC of the subject, calculate the plantar target force based on the plantar MVC, and confirm the target lower limb pressure output range based on the plantar MVC;

[0008] Collecting the subject's EEG signal and EMG signal;

[0009] Preprocessing the collected EEG signals and EMG signals to obtain preprocessed EEG and EMG data;

[0010] The coherence analysis of EMG data was performed to obtain the time delay of movement intention information between the subjects' EMG data.

[0011] A further improvement of the present invention is:

[0012] The method for collecting the MVC of the soles of the subjects is: collecting several times and taking the average value to obtain the MVC of the soles of the lower limbs of the left and right legs;

[0013] The target lower limb pressure output range is 0.8MVC±10N.

[0014] The specific method for preprocessing the received EEG signal and EMG signal is as follows:

[0015] The EEG signal is baseline calibrated, and adaptive filtering is used to remove power frequency interference. The independent component analysis algorithm is used to remove eye artifacts. The EMG signal is bandpass filtered and adaptive filtering is used to remove power frequency interference to improve the EMG-to-noise ratio.

[0016] The specific method for analyzing brain myoelectric coherence is:

[0017] Reconstruct the EMG data into high-dimensional data and use multivariate models to fit the EMG data;

[0018] Determine the order of the multivariate model for fitting EMG data;

[0019] After determining the model order, estimate the model coefficients that fit the EMG data;

[0020] The model with determined parameters is used to perform brain myoelectric coherence analysis, determine the brain myoelectric coherence frequency band, and quantify the brain myoelectric delay of the result.

[0021] The specific method of reconstructing EMG data into high-dimensional data and fitting EMG data using multivariate model is as follows:

[0022]

[0023]

[0024]

[0025] Among them, Y t is the reconstructed high-dimensional EMG data; t is the sampling point of EMG data, t = 1, 2, ... T; K is the sum of the number of EMG data channels; M is the number of all trials of the subject; p is the model order; A t-i is the K×K dimensional model coefficient; E t is the residual noise after fitting with covariance ∑.

[0026] The specific calculation method for determining the multivariate model order of brain and myoelectric data is:

[0027]

[0028]

[0029] Among them, p is the model order; M is the number of experimental trials; K is the sum of the number of EMG data channels, N=ML, and L is the time window length.

[0030] The specific method for estimating the coefficients of the model for fitting brain and myoelectric data is:

[0031] Determine the model coefficient matrix Θ:

[0032] Θ=(A1,A2...A P ) (6)

[0033] Where p is the model order, and the calculation method of A is shown in formula (3);

[0034] Determine the model data matrix W n :

[0035] W n =(Y n-1 ,Y n-2 ...Y n-p ) (7)

[0036] Among them, Y n-i is the reconstructed EMG data shown in formula (2), n is the current sampling point;

[0037]

[0038]

[0039] Among them, Z n is the instantaneous error estimation matrix, giving the expected output Y n The error between the expected output and the output; n is the current sampling point; E nis the cumulative error, c is the forgetting factor, and 0 < c < 1 ensures the forgetting of electroencephalomyogram (EEG) data that is far from the current moment;

[0040] The recursive algorithm iteratively calculates Θ|min(E n ), determines the final model coefficient matrix Θ, and calculates the noise covariance matrix ∑ according to the model instantaneous error estimation matrix, as shown in the following formula:

[0041]

[0042] where c is the forgetting factor; n is the current sampling point; M is the number of experimental trials;

[0043] The calculation method of the power spectrum matrix is:

[0044]

[0045] where f is the current EEG coherence frequency; f s is the EEG sampling frequency; A is the mvAR model coefficient; ∑ is the fitting noise covariance, and the iterative calculation method is shown in Equation (10);

[0046] According to Equation (11), the coherence of each EEG channel can be calculated as follows:

[0047]

[0048] where f is the current EEG coherence frequency; i, j are the corresponding EEG channels;

[0049] After determining the model fitting, the central frequency f κ at which significant EEG coherence appears and the time t0 are determined, and the results are output to the EEG time delay quantization unit.

[0050] The specific method for performing time delay quantization is:

[0051] The EEG at the central frequency and time when the subject appears is artificially translated with lagged data, as shown in Equation (13):

[0052] δ = max(Coh ij (f κ )) (13)

[0053] where δ is the finally quantified EEG time delay; Coh ij is the coherence of the i, j channels of the EEG; f κ is the central frequency of the typical EEG coherence;

[0054] After performing several trials, the average value is taken to obtain the time delay of the movement intention information between the corresponding myoelectric channel of the subject and the electroencephalogram, and the synchronization of the EEG movement intention information is completed.

[0055] A brain-electromyography information synchronization system for accurately capturing movement intentions, comprising a steady-state force acquisition module, a brain-electromyography acquisition module, a signal preprocessing module and a brain-electromyography information synchronization module;

[0056] The steady-state force acquisition module is used to collect the plantar MVC of the subject, calculate the plantar target force according to the plantar MVC, and confirm the target lower limb pressure output interval according to the plantar MVC;

[0057] The brain-electromyography acquisition module is used to collect the subject's electroencephalogram (EEG) signals and electromyography (EMG) signals;

[0058] The signal preprocessing module is used to preprocess the collected EEG signals and EMG signals to obtain preprocessed EEG and EMG data;

[0059] The brain electromyography information synchronization module is used to perform coherence analysis on the brain electromyography data to obtain the movement intention information delay between the subject's brain electromyography.

[0060] Furthermore, the brain and myoelectric information synchronization module includes a brain and myoelectric coherence analysis unit and a brain and myoelectric delay quantification unit;

[0061] The electromyographic coherence analysis unit determines the order of the fitting model and estimates the fitting model coefficients, performs electromyographic coherence analysis using the model with determined parameters, determines the electromyographic coherence frequency band and time, and outputs the result to the electromyographic delay quantification unit for electromyographic delay quantification;

[0062] The EMG delay quantification unit is used to quantify the delay of EMG data and perform artificial translation of the delayed EMG data. The translation amount between EMG and brain corresponds to the EMG delay. After several tests, the average value is taken to obtain the delay of the intention information of the brain area and muscles corresponding to the EMG and EMG channels of the subject. The EMG data after delay offset is used to capture the human body's movement intention.

[0063] Compared with the prior art, the present invention has the following beneficial effects:

[0064] The present invention discloses a brain-electromyographic information synchronization system and method for accurately capturing movement intentions. The brain-electromyographic coherence analysis method based on multivariate model fitting can effectively determine the central frequency and time of brain and muscle information exchange during steady-state force output of the lower limbs, and use it to estimate the time delay between brain-electromyographic signals. At the same time, the fitting of the multivariate model includes the coupling between brain-electromyographic channels, and considers cognition as a whole, which is more in line with the physiological basis of the human cognitive process. Based on brain electromyography coherence analysis and multivariate model fitting, the delay estimation of brain electromyography movement intention is realized. After the delay is offset, the brain electromyography will have the ability to capture the human body's movement intention more accurately. Although this delay is not large, the impact of this delay will also increase under the state of muscle or mental fatigue. The performance improvement brought by eliminating the time difference between the two will be amplified. The brain electromyography after information synchronization contains the same movement intention, which can be used to realize the accurate capture of human body movement intention. It will effectively lay the foundation for the effectiveness of the method of brain electromyography fusion recognition of movement intention, thereby improving the robustness and reliability of the brain electromyography fusion recognition system, and also improving the logic of the method of brain electromyography fusion recognition of movement intention, thereby promoting the promotion and development of vital point integration technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0066] Figure 1 It is a schematic diagram of the system structure of the present invention;

[0067] Figure 2 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0069] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0070] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0071] In the description of the embodiments of the present invention, it should be noted that if the terms "upper", "lower", "horizontal", "inner", etc. indicate an orientation or positional relationship based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of the invention is usually placed when in use, it is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0072] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", which does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0073] In the description of the embodiments of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal connection of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0074] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0075] See also Figure 1-2 The embodiment of the present invention discloses a brain-electromyographic information synchronization system for accurately capturing movement intention, including five major modules: steady-state force acquisition, brain-electromyographic acquisition, signal preprocessing, and brain-electromyographic information synchronization. The collected brain-electromyographic data are fitted with a multivariate model, and the delay of the brain-electromyographic movement intention information is quantified based on the brain-electromyographic coherence analysis, thereby realizing the synchronization of the movement intention information of the brain-electromyographic signal by removing the delay.

[0076] The steady-state force acquisition module is used to collect the user's plantar pressure and use it to evaluate whether the user is in a steady-state force output state. The result is output to the brain-electromyography acquisition module, which is used to intercept the brain-electromyography signal when the lower limbs are in steady-state force output;

[0077] The brain-electromyography acquisition module is used to synchronously acquire the user's EEG signals and EMG signals, and wirelessly send the EEG signals and EMG signals to the signal preprocessing module;

[0078] The signal preprocessing module is used to filter out artifacts from EEG signals and EMG signals, and after obtaining EEG signals and EMG signals with high signal-to-noise ratio, synchronize EEG and EMG information;

[0079] The EMG information synchronization module is used to fit the EMG data using a multivariate model, perform EMG coherence analysis, and quantify the EMG delay. The output result is the estimated movement intention information delay between EMGs.

[0080] The steady-state force acquisition module includes a pressure acquisition unit and a steady-state force judgment unit. The pressure acquisition unit is used to collect the MVC of the sole of the subject's foot, and the average value of multiple acquisitions is the MVC of the sole of the subject's foot; the steady-state force judgment unit is used to confirm the target lower limb pressure output interval according to the sole MVC, and to determine the moment when the output pressure of the subject's lower limb meets the target output interval and send it to the brain-electromyography acquisition module for extracting the brain-electromyography data synchronously collected within the target output force interval;

[0081] The brain-electromyography acquisition module includes an electroencephalography acquisition unit and an electromyography acquisition unit; the electroencephalography acquisition unit includes an electroencephalography acquisition subunit and a wireless transmission subunit, and the electroencephalography acquisition subunit is connected to the electroencephalography cap; the electromyography acquisition unit includes an electromyography acquisition subunit and a wireless transmission subunit, and the electromyography acquisition subunit is connected to the electromyography electrode; the electroencephalography acquisition subunit and the electromyography acquisition subunit both exchange data with the signal preprocessing module through the wireless transmission subunit;

[0082] According to the positions defined by the international 10-20 standard electrode placement method, the FCz, FC1, FC2, C1, C2, C3, C4, CP1, CP2 and Cz channels of the motor cortex were collected using an EEG cap. The electromyographic electrodes were placed on the rectus femoris, vastus lateralis, vastus medialis, tibialis anterior, lateral gastrocnemius and medial gastrocnemius muscles of the user's lower limbs to collect the electromyographic signals of the corresponding muscles;

[0083] The collected EEG and myoelectric data are preprocessed in the signal preprocessing module. The signal preprocessing module includes an artifact removal unit and a noise removal unit, which are used to filter out 50Hz power frequency interference from EEG and EMG and filter out electrooculographic artifacts in EEG, and finally obtain EEG and myoelectric data with a high signal-to-noise ratio.

[0084] A brain and myoelectric information synchronization module, including a brain and myoelectric coherence analysis unit and a brain and myoelectric delay quantification unit;

[0085] The EMG coherence analysis unit includes a model order determination subunit and a model coefficient estimation subunit; the model order determination subunit determines the optimal order of the mvAR model, taking into account both the simplicity and fitting accuracy of the model; after determining the model order, the model coefficient estimation subunit estimates the model coefficients using a recursive least squares algorithm with a forgetting factor, and uses the estimated parameters to perform EMG coherence analysis. After determining the EMG significant coherence frequency band, the results are output to the EMG delay quantification unit for EMG delay quantification;

[0086] The EMG delay quantification unit quantifies the EMG data based on the maximum coherence rule and performs artificial shifting of the delayed EMG data. The shift amount with the maximum coherence between EMG and brain corresponds to the EMG delay. The average value is taken after multiple experiments to obtain the delay of the motor intention information of the brain area and muscles corresponding to the EMG and EMG channels of the subject.

[0087] The embodiment of the present invention also discloses a method for synchronizing brain and myoelectric information for accurately capturing movement intention, comprising the following steps:

[0088] Step 1: Collect the MVC of the plantar pressure of each subject's left and right legs. The subject keeps the ankle dorsiflexed and presses the plantar pressure sensor with the bottom of the heel as hard as possible for at least 10 seconds. The plantar pressure of the left and right legs is obtained by taking the average value after multiple collections. The target plantar force of the left and right legs can be obtained based on the plantar MVC of the left and right legs, that is, the output range of 0.8MVC floating up and down, and the floating range is 0.8MVC±10N, where N represents Newton.

[0089] Step 2: Use an EEG cap and an EMG electrode to synchronously collect EEG signals and EMG signals of the subject when the subject performs steady-state force output of the lower limbs within the target output range, and wirelessly transmit the EEG signals and EMG signals to a signal preprocessing module;

[0090] Step 3: The signal preprocessing module preprocesses the received EEG and EMG signals; baseline calibration is performed on the EEG signals, 1-100Hz bandpass filtering is performed, 50Hz power frequency interference is removed by adaptive filtering, and electrooculographic artifacts are removed by independent component analysis algorithm; 1-200Hz bandpass filtering is performed on the EMG signals, and 50Hz power frequency interference is removed by adaptive filtering;

[0091] Step 4: The EMG coherence analysis unit in the EMG signal synchronization module performs coherence analysis on the preprocessed EMG data:

[0092] Step 4-1, reconstruct the multi-channel EMG data into high-dimensional data and fit it with the mvAR model, as shown below:

[0093]

[0094]

[0095]

[0096] Among them, Y t is the reconstructed high-dimensional EMG data; t is the sampling point of EMG data, t = 1, 2, ... T; K is the sum of the number of EMG data channels; M is the number of all trials of the subject; p is the order of the mvAR model; A t-i is the coefficient of the K×K dimensional mvAR model; E t is the residual noise after fitting with covariance ∑.

[0097] Step 4-2: The model order determination subunit uses the Akaike information content (AIC) / Bayesian information content (BIC) criterion to determine the high-dimensional mvAR model order for fitting EMG data. The calculation formula is as follows:

[0098]

[0099]

[0100] Among them, p is the order of the mvAR model; M is the number of experimental trials; K is the sum of the number of EMG data channels; N = ML, and L is the time window length.

[0101] The order p that minimizes AIC / BIC in the fitting process is selected as the final order of the mvAR model.

[0102] Step 4-3: After determining the model order, the model coefficient estimation subunit uses the recursive least squares algorithm with a forgetting factor to estimate the mvAR model coefficients that fit the EMG data. The process is as follows:

[0103] The mvAR model coefficient matrix Θ is determined as follows:

[0104] Θ=(A1,A2...A P ) (6)

[0105] Where p is the order of the mvAR model and the calculation formula of A is shown in (3).

[0106] Determine the mvAR model data matrix W n As shown below:

[0107] W n =(Y n-1 ,Y n-2 ...Y n-p ) (7)

[0108] Among them, Y n-i is the reconstructed EMG data as shown in formula (2); n is the current sampling point.

[0109]

[0110]

[0111] Among them, Z n is the instantaneous error estimation matrix, which gives the error between the expected output Y n and the estimated output; n is the current sampling point; E n is the cumulative error, c is the forgetting factor, and 0 < c < 1 ensures that the electroencephalogram and electromyogram data far from the current moment are forgotten.

[0112] The recursive algorithm iteratively calculates Θ|min(E n ), determines the final mvAR model coefficient matrix Θ, and the noise covariance matrix ∑ can be calculated according to the mvAR model instantaneous error estimation matrix, as shown in the following formula:

[0113]

[0114] Among them, c is the forgetting factor; n is the current sampling point; M is the number of experimental trials.

[0115] And the power spectral matrix is as shown in the following formula:

[0116]

[0117] Among them, f is the current electroencephalogram and electromyogram coherence frequency; f s is the electroencephalogram and electromyogram sampling frequency; A is the mvAR model coefficient; ∑ is the fitting noise covariance, and the iterative calculation method is as shown in formula (10).

[0118] According to formula (11), the coherence of each channel of the electroencephalogram and electromyogram can be calculated as follows:

[0119]

[0120] Among them, f is the current electroencephalogram and electromyogram coherence frequency; i, j are the corresponding electroencephalogram and electromyogram channels.

[0121] After determining the model fitting, the central frequency f κ at which the significant coherence of the electroencephalogram and electromyogram appears and the moment t0 are determined, and the results are output to the electroencephalogram and electromyogram time delay quantization unit;

[0122] Step 4-4: The electroencephalogram and electromyogram time delay quantization unit uses the maximum coherence rule for time delay quantization. The electroencephalogram and electromyogram data at the central frequency and moment where the significant coherence of each subject appears are artificially translated, and the translation amount corresponding to the maximum value of the electroencephalogram and electromyogram coherence is the estimated value of the electroencephalogram and electromyogram time delay, as shown in the following formula:

[0123] δ = max(Coh ij (f κ )) (13)

[0124] Among them, δ is the final quantified EMG delay; Coh ij is the coherence of brain electromyography channels i and j; f κ is the central frequency of typical coherence of brain electromyography;

[0125] After averaging multiple trials, the time delay of the subject's corresponding electromyographic channel and EEG motor intention information can be obtained, that is, the synchronization of brain-electromyographic motor intention information is completed.

[0126] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for synchronizing brain and myoelectric information for accurately capturing movement intention, characterized in that: The following steps are involved: The plantar MVC of the subject is collected, the plantar target force is calculated according to the plantar MVC, and the target lower limb pressure output range is determined according to the plantar MVC; the method for collecting the plantar MVC of the subject is: collecting several times, taking the average value to obtain the plantar MVC of the lower limbs of the left and right legs; the target lower limb pressure output range is 0.8 MVC±10N; Collecting the subject's EEG signal and EMG signal; Preprocessing the collected EEG signals and EMG signals to obtain preprocessed EEG and EMG data; The specific method for preprocessing the collected EEG signals and EMG signals is as follows: The EEG signal was baseline calibrated, and the power frequency interference was removed by adaptive filtering and the eye artifacts were removed by independent component analysis algorithm. Bandpass filtering is performed on the electromyographic signal, and adaptive filtering is used to remove power frequency interference to improve the brain electromyographic signal-to-noise ratio; Conduct coherence analysis on EMG data to obtain the time delay of movement intention information between the subjects' EMG data. The specific method for analyzing brain myoelectric coherence is: Reconstruct the EMG data into high-dimensional data and use multivariate models to fit the EMG data; Determine the order of the multivariate model for fitting EMG data; After determining the model order, estimate the model coefficients that fit the EMG data; The model with determined parameters is used to perform brain myoelectric coherence analysis, determine the brain myoelectric coherence frequency band, and quantify the brain myoelectric delay of the result; The specific method of reconstructing EMG data into high-dimensional data and fitting EMG data using multivariate model is as follows: Among them, Y t is the reconstructed high-dimensional EMG data; t is the EMG data sampling point, t = 1, 2, ... T; K is the sum of the number of EMG data channels, n is the current sampling point; p is the model order; A t-i is the K×K dimensional model coefficient; E t is the residual noise after fitting with covariance ∑; The specific calculation method for determining the multivariate model order of fitting brain and myoelectric data is: Where M is the number of experimental trials; N = ML, L is the time window length; The specific method for estimating the model coefficients of the fitted brain and myoelectric data is: Determine the model coefficient matrix Θ: Θ=(A1,A2...A P ) (6) The calculation method of A is shown in formula (3); Determine the model data matrix W n : W n (And n-1 ,AND n-2 ...AND n-p ) (7) Among them, Y n-i is the reconstructed EMG data shown in formula (2); Among them, Z n is the instantaneous error estimation matrix, which gives the error between the expected output Y n and the predicted output; E n is the cumulative error, c is the forgetting factor, and 0 < c < 1 ensures that the electroencephalomyogram data far from the current moment is forgotten; The recursive algorithm iteratively calculates Θ|min(E n ), determine the final model coefficient matrix Θ, and calculate the noise covariance matrix ∑ according to the model instantaneous error estimation matrix, as shown in the following formula: The calculation method of the power spectrum matrix is: Where, f is the current brain myoelectric coherence frequency; f s is the brain electromyography sampling frequency; A is the mvAR model coefficient; ∑ is the fitting noise covariance, and the iterative calculation method is shown in formula (10); According to formula (11), the coherence of each channel of brain electromyography can be calculated as follows: Among them, Coh ij is the coherence of brain electromyography channels i and j; i and j are the corresponding brain electromyography channels; After model fitting, determine the center frequency f at which significant coherence of EMG occurs κ and time t0, and output the result to the EMG delay quantization unit; The specific method for quantifying the EMG delay is: The central frequency and time of the subject's EMG are artificially shifted to lag data, see formula (13): δ=max(Coh ij (f κ )) (13) Among them, δ is the final quantified EMG delay; f κ is the central frequency of typical coherence of brain electromyography; After conducting several experiments, the average value is taken to obtain the movement intention information delay between the corresponding EMG channel and the EEG of the subject, thus completing the synchronization of EMG movement intention information.

2. A brain and myoelectric information synchronization system for accurately capturing the movement intention of the method described in claim 1, characterized in that: It includes a steady-state force acquisition module, a brain-electromyography acquisition module, a signal preprocessing module and a brain-electromyography information synchronization module; The steady-state force acquisition module is used to collect the plantar MVC of the subject, calculate the plantar target force according to the plantar MVC, and confirm the target lower limb pressure output interval according to the plantar MVC; The brain-electromyography acquisition module is used to collect the subject's electroencephalogram (EEG) signals and electromyography (EMG) signals; The signal preprocessing module is used to preprocess the collected EEG signals and EMG signals to obtain preprocessed EEG and EMG data; The EMG information synchronization module is used to perform coherence analysis on the EMG data to obtain the movement intention information delay between the EMG data of the subject; The brain and myoelectric information synchronization module includes a brain and myoelectric coherence analysis unit and a brain and myoelectric delay quantification unit; The electromyographic coherence analysis unit determines the order of the fitting model and estimates the fitting model coefficients, performs electromyographic coherence analysis using the model with determined parameters, determines the electromyographic coherence frequency band and time, and outputs the result to the electromyographic delay quantification unit for electromyographic delay quantification; The EMG delay quantification unit is used to quantify the delay of EMG data and perform artificial translation of the delayed EMG data. The translation amount between EMG and brain corresponds to the EMG delay. After several tests, the average value is taken to obtain the delay of the intention information of the brain area and muscles corresponding to the EMG and EMG channels of the subject. The EMG data after delay offset is used to capture the human body's movement intention.

Images