Wearable scoliosis rehabilitation monitoring method and system

Abstract

The invention discloses a wearable scoliosis rehabilitation monitoring method which comprises the following steps: acquiring myoelectricity data of muscles on both sides of the trunk and the back of a patient in a rehabilitation training process, limiting the signal frequency of the myoelectricity data, and eliminating power frequency interference and electrocardio noise to obtain a myoelectricity filtering signal; carrying out full-wave rectification and low-pass filtering processing on the myoelectricity filtering signal in sequence to obtain a myoelectricity activation signal; the method comprises the following steps: acquiring myoelectricity data of a patient in a rehabilitation training process, acquiring rehabilitation action amplitude features, rehabilitation action frequency features, rehabilitation action wavelet coefficient features and rehabilitation action activation amplitude features in the myoelectricity data, and inputting the features into a feature selector for screening; and inputting the result into the classifier to obtain a judgment result of the degree of completion of rehabilitation training, so that the rehabilitation training process of the patient is monitored, and the patient is assisted in optimizing the action of rehabilitation training.

Description

Technical Field

[0001] The present invention belongs to the technical field of rehabilitation assistive devices, and in particular relates to a wearable scoliosis rehabilitation monitoring method and system. Background Art

[0002] In recent years, with the application of non-invasive electrodes, the development of electronic computers and the progress of physiological and biochemical research, the research on electromyography has become increasingly in-depth. Electromyography refers to the muscle bioelectric pattern recorded by electromyographs, which is of great significance for evaluating human activities in human-machine systems. Electromyography allows doctors to study the functional activity of patients' muscles in dynamic conditions. In human electromyography monitoring, there are two main technologies for monitoring muscle electrical signals: one is to use metal needle electrodes to pierce the muscles and connect to electrophysiological instruments to collect electromyographic data. Needle electrode electromyography records and analyzes the bioelectric signals generated by each muscle fiber of the motor unit during rest and voluntary contraction by inserting needle electrodes into the muscles. The electromyographic signals obtained in this way are relatively stable, but this type of invasive collection method will bring risks such as pain and bleeding to patients. The second is to use surface gel patch electrodes to collect signals and connect electrophysiological instruments to collect electromyographic data. Compared with needle electrode electromyography, surface electromyography has the advantages of non-invasiveness, non-traumaticity and simple operation in measurement.

[0003] Adolescent-specific scoliosis, or scoliosis for short, is a very common type of scoliosis, and its patients are mainly adolescents. The spinal deformity of scoliosis patients leads to asymmetry of the paraspinal muscles. For example, the erector spinae muscle is shown on the electromyogram as having a higher average amplitude on the convex side of the scoliotic curve than on the concave side. In the early stages of scoliosis, the treatment method is usually to observe the development of the spinal curve or to perform rehabilitation exercise therapy. In rehabilitation exercise therapy, the doctor guides the patient on rehabilitation movements, and the patient needs to perform rehabilitation training independently. Due to the characteristics of scoliosis patients being young, heavy schoolwork, and a long treatment cycle, they often have non-standard movements and insufficient completion during the self-performed rehabilitation training. At the same time, they are unable to monitor whether the rehabilitation movements they have performed have reached the standard level and whether the muscle activation degree of the target muscle group is in place. On the other hand, scoliosis patients cannot improve the standard degree of the rehabilitation movements they have performed after completing the rehabilitation exercises, which makes it difficult for them to correctly perform the rehabilitation movements in the subsequent continuous self-performed rehabilitation training. As a result, the expected rehabilitation effect of rehabilitation training cannot be achieved, and there is currently a lack of more convenient monitoring methods for wearable surface electromyography monitoring devices.

[0004] Existing patent CN117653084A discloses a method for evaluating the rehabilitation status of scoliosis using gait, and uses visual images to obtain gait symmetry scores to evaluate the rehabilitation status of scoliosis. This patent only evaluates the rehabilitation status of the spine, and lacks the judgment of the completion of rehabilitation movements and rehabilitation movement mechanics suggestions for scoliosis rehabilitation training; Existing patent CN115985463A discloses a real-time prediction method of muscle fatigue based on wearable devices. This patent predicts and classifies the fatigue state of the current muscle and gives real-time voice reminders. Muscle fatigue is a natural physiological reaction after exercise, which can be improved by rest and nutritional supplementation, while rehabilitation movements are usually a series of physical therapy or exercise therapy to accelerate the rehabilitation process. This patent lacks discriminant analysis of the completion of bilateral muscle rehabilitation movements; the existing patent CN113368468A discloses an adjustable spinal correction training frame with electromyography monitoring, which is used to monitor whether the user's target muscles are involved in force and the force situation. This patent does not meet the portability requirements and is only applicable to static rehabilitation training processes. It lacks discriminant analysis of the completion of rehabilitation movements in dynamic rehabilitation training processes; the existing paper document "Synergy analysis of back muscle activities in patients with adolescent idiopathic scoliosis based on high-density electromyogram." uses a high-density surface electromyography system to study the differences in back muscle synergy between AIS patients and healthy subjects when performing flexion / extension, lateral bending, axial rotation, sitting and standing. It only analyzes the back muscles in static postures and lacks analysis of the back muscles in dynamic rehabilitation training. Summary of the invention

[0005] The purpose of the present invention is to solve the problems raised in the background technology and to propose a wearable scoliosis rehabilitation monitoring method.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] The present invention provides a wearable scoliosis rehabilitation monitoring method, comprising:

[0008] The electromyographic data of the bilateral muscles of the back of the trunk of the patient during rehabilitation training are obtained, and the signal frequency of the electromyographic data is limited, and the power frequency interference and electrocardiographic noise are eliminated to obtain the electromyographic filter signal, and then the electromyographic filter signal is processed by full-wave rectification and low-pass filtering in sequence to obtain the electromyographic activation signal;

[0009] Preprocessing the electromyographic filter signal and the electromyographic activation signal respectively obtains the electromyographic signal segment and the electromyographic activation segment in turn;

[0010] The EMG signal segments are analyzed by time domain waveform analysis and sample entropy analysis to obtain the amplitude characteristics of the rehabilitation movement, the EMG signal segments are analyzed by power spectrum analysis to obtain the frequency characteristics of the rehabilitation movement, and the EMG signal segments are analyzed by wavelet transform to obtain the wavelet coefficient characteristics of the rehabilitation movement;

[0011] The activation amplitude characteristics of rehabilitation movements are obtained by analyzing the EMG activation segments through time domain waveform analysis.

[0012] Input the rehabilitation movement amplitude feature, rehabilitation movement frequency feature, rehabilitation movement wavelet coefficient feature and rehabilitation movement activation amplitude feature into the feature selector to obtain the filtered features, and then input the filtered features into the classifier at the same time to obtain the judgment result of the completion degree of rehabilitation training;

[0013] When the judgment result is low, the myoelectricity-activation coupling model is established for the bilateral muscles of the patient's back trunk, and the model is as follows:

[0014]

[0015] in,

[0016] u j (t) = αe j (td)-β1u j (t-1)-β2u j (t-2);

[0017] Among them, a j (t) represents the muscle activation degree at time t, e j (t) represents the EMG activation segment at time t, u j (t) represents the EMG activation segment after muscle activation processing, A represents the nonlinear shape factor, which is generally constrained to be -3<A<0, α represents the gain coefficient, β1 and β2 both represent the recursive coefficient, and d represents the EMG delay;

[0018] The obtained a j (t) is compared with the first preset value and the second preset value. j (t) is less than the first preset value or a j (t) is less than a second preset value, information about insufficient muscle activation at the corresponding moment is generated and fed back to the patient.

[0019] Preferably, the method of acquiring electromyographic data of the bilateral muscles of the back of the trunk of the patient during rehabilitation training, limiting the signal frequency of the electromyographic data, and eliminating power frequency interference and electrocardiographic noise to obtain an electromyographic filter signal, and then subjecting the electromyographic filter signal to full-wave rectification and low-pass filtering in sequence to obtain an electromyographic activation signal, includes:

[0020] A bandpass filter is used to limit the signal frequency of the electromyographic data to a preset frequency band, and a comb notch filter is used to eliminate power frequency interference. Then, an adaptive filter is used to filter out the ECG noise to obtain the electromyographic filter signal, which is then processed through full-wave rectification and low-pass filtering to obtain the electromyographic activation signal.

[0021] Preferably, the preprocessing of the electromyographic filter signal and the electromyographic activation signal to obtain the electromyographic signal segment and the electromyographic activation segment in sequence comprises:

[0022] The extended Teague-Kaiser energy operator is used to detect the starting time of the electromyographic filter signal and the electromyographic activation signal respectively, and the start and end time nodes of the muscle contraction are obtained. The segments of the active correction period and the continuous maintenance period of the electromyographic filter signal are segmented as electromyographic signal segments, and the segments of the active correction period and the continuous maintenance period of the electromyographic activation signal are segmented as electromyographic activation segments. The active correction period refers to the time segment when the patient maintains the posture of rehabilitation training and the muscles contract, and the continuous maintenance period refers to the time segment when the patient maintains the posture of rehabilitation training and the muscles relax.

[0023] Preferably, the extended Teague-Kaiser energy operator is calculated by electromyographic activation signal, and the calculation formula is as follows:

[0024] Ψ (d,m,s) [n] = det(X n );

[0025] in,

[0026]

[0027] Among them, X n represents a matrix of d×d dimensions, x[n] represents the value of the nth sampling point of the electromyographic activation signal, s represents the step size, m represents the starting offset, Ψ (d,m,s) [n] represents the extended Teague-Kaiser energy operator, det represents the function for calculating the determinant of the matrix, and d represents X n The order of .

[0028] Preferably, the rehabilitation movement frequency characteristics include: peak frequency, median frequency, average power frequency, low frequency ratio and change rate of each frequency of the electromyographic signal segment;

[0029] The rehabilitation action wavelet coefficient characteristics include: the absolute mean and root mean square value of the rehabilitation action wavelet coefficients of the electromyographic signal segment;

[0030] The activation amplitude characteristics of the rehabilitation movement include: the mean, variance and root mean square value of the electromyographic activation segment;

[0031] The rehabilitation movement amplitude characteristics include: the root mean square value, absolute mean, number of zero crossings, signal segment length, number of slope sign changes, Willison amplitude and rehabilitation movement signal entropy of the electromyographic signal segment, wherein the root mean square value, absolute mean, number of zero crossings, signal segment length, number of slope sign changes and Willison amplitude of the electromyographic signal segment are obtained through time domain waveform analysis, and the rehabilitation movement signal entropy is obtained through sample entropy analysis, and the calculation formula of the rehabilitation movement signal entropy is as follows:

[0032]

[0033] in,

[0034]

[0035] d(X i ,X j )=max k＝0,...,m-1 (x(i+k)-x(j+k));

[0036] X i =[x(i),x(i+1),…,x(i+m-1)];

[0037] Among them, SampEn(m,r,N) represents the rehabilitation action signal entropy, B m (r) represents the ratio of signal vectors of length m that meet the similarity condition, B m+1 (r) represents the ratio of signal vectors of length m+1 that meet the similarity condition, I(d(X i ,X j )<r) represents an indicator function, which is used to determine whether the distance between two signal vectors is less than a threshold value r. i The vector representing the i-th signal segment, X j represents the vector of the jth signal segment, k is the index used to locate within the signal segment, and is used to calculate the difference between the two vectors at each corresponding point.

[0038] Preferably, the feature selector adopts a two-level structure, the first level adopts an elastic network regularization algorithm, and the second level adopts a sequential floating feature selection algorithm;

[0039] The classifier adopts support vector machine algorithm.

[0040] Preferably, the first preset value The muscle activation degree obtained by inputting the electromyographic activation segment corresponding to the sustained maintenance period into the electromyographic-activation degree coupling model, the second preset value The muscle activation degree is obtained by inputting the EMG activation segment corresponding to the standing posture and muscle relaxation into the EMG-activation degree coupling model.

[0041] A wearable scoliosis rehabilitation monitoring system includes an electromyography acquisition device and a host computer, wherein the host computer includes a data processing module and a rehabilitation action guidance feedback module, wherein:

[0042] The electromyographic acquisition device is used to obtain electromyographic data of the bilateral muscles on the back of the patient's trunk during rehabilitation training, wherein the electromyographic acquisition device is attached to the bilateral muscles on the back of the patient's trunk at a preset distance from the midline of the spine, and collects the electromyographic data, and sends the electromyographic data to the data processing module of the host computer;

[0043] The data processing module is used to limit the signal frequency of the received electromyographic data and eliminate power frequency interference and electrocardiographic noise to obtain an electromyographic filter signal, and then process the electromyographic filter signal through full-wave rectification and low-pass filtering in sequence to obtain an electromyographic activation signal;

[0044] The data processing module is further used to pre-process the electromyographic filter signal and the electromyographic activation signal to obtain the electromyographic signal segment and the electromyographic activation segment respectively;

[0045] The data processing module is further used to obtain the rehabilitation movement amplitude characteristics by performing time domain waveform analysis and sample entropy analysis on the electromyographic signal segments, obtain the rehabilitation movement frequency characteristics by performing power spectrum analysis on the electromyographic signal segments, and obtain the rehabilitation movement wavelet coefficient characteristics by performing wavelet transform analysis on the electromyographic signal segments;

[0046] The data processing module is also used to obtain activation amplitude characteristics of rehabilitation movements by analyzing the myoelectric activation segments through time domain waveforms;

[0047] The data processing module is further used to input the rehabilitation movement amplitude feature, rehabilitation movement frequency feature, rehabilitation movement wavelet coefficient feature and rehabilitation movement activation amplitude feature into the feature selector to obtain the filtered features, and then input the filtered features into the classifier at the same time to obtain the judgment result of the completion degree of rehabilitation training, and send the judgment result to the rehabilitation movement guidance feedback module;

[0048] The rehabilitation action guidance feedback module is used to establish an electromyographic-activation coupling model for the bilateral muscles on the back of the patient's trunk when the judgment result is low, and the model is as follows:

[0049]

[0050] in,

[0051] u j (t) = αe j (td)-β1u j (t-1)-β2u j (t-2);

[0052] Among them, a j (t) represents the muscle activation degree at time t, e j (t) represents the EMG activation segment at time t, u j (t) represents the EMG activation segment after muscle activation processing, A represents the nonlinear shape factor, which is generally constrained to be -3<A<0, α represents the gain coefficient, β1 and β2 both represent the recursive coefficient, and d represents the EMG delay;

[0053] The rehabilitation action guidance feedback module is also used to obtain a j (t) is compared with the first preset value and the second preset value. j (t) is less than the first preset value or a j (t) is less than a second preset value, information about insufficient muscle activation at the corresponding moment is generated and fed back to the patient.

[0054] Preferably, the electromyography acquisition device includes an analog front-end module, an analog-to-digital conversion module and a wireless transmission module, wherein:

[0055] The analog front-end module is used to obtain the electromyographic data of the bilateral muscles of the back of the patient's trunk during rehabilitation training, and perform low-pass filtering and amplification processing on the electromyographic data signals in sequence, and has a high common-mode rejection ratio, provides a common-mode bias voltage, reduces the common-mode noise at the input end, and sends the electromyographic data after the common-mode noise is reduced to the analog-to-digital conversion module;

[0056] The analog-to-digital conversion module converts the analog electromyographic signal of the received electromyographic data into a discrete digital electromyographic signal, and sends the discrete digital electromyographic signal of the electromyographic data to the wireless transmission module;

[0057] The wireless transmission module sends the discrete digital electromyographic signals of the received electromyographic data to the data processing module of the host computer for processing.

[0058] Preferably, the host computer further includes a data storage module and a waveform display module, wherein:

[0059] The data processing module is also used to send the discrete digital electromyographic signal of the received electromyographic data to the data storage module for storage, and send the waveform of the discrete digital electromyographic signal of the electromyographic data to the display module for waveform display.

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] The wearable scoliosis rehabilitation monitoring method and system obtains the electromyographic data of the patient during the rehabilitation training process, and obtains the rehabilitation movement amplitude characteristics, rehabilitation movement frequency characteristics, rehabilitation movement wavelet coefficient characteristics and rehabilitation movement activation amplitude characteristics in the electromyographic data, and inputs each feature into a feature selector for screening, and then inputs it into a classifier to obtain the judgment result of the degree of completion of the rehabilitation training, so as to monitor the process of the patient's rehabilitation training, and then assist the patient in optimizing the rehabilitation training movements;

[0062] At the same time, when the judgment result is low, the electromyography-activation coupling model is established to calculate whether the muscle activation corresponding to each moment of the electromyography activation segment is sufficient, and the information of insufficient muscle activation at the corresponding moment is generated and fed back to the patient, so as to assist patients and doctors to optimize the rehabilitation training movements and improve the completion of the movements. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a schematic diagram of the flow of the wearable scoliosis rehabilitation monitoring method and system of the present invention;

[0064] Figure 2 This is a module block diagram of the wearable scoliosis rehabilitation monitoring system of the present invention. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0066] In one embodiment, Figure 1 As shown, a wearable scoliosis rehabilitation monitoring method is provided, comprising:

[0067] Step S1, obtaining the electromyographic data of the bilateral muscles of the back of the trunk of the patient during the rehabilitation training process (the electromyographic data can be collected by an electromyographic acquisition device (sensor), such as attaching the electromyographic acquisition device to the bilateral muscles of the back of the trunk of the patient to collect data, and the collected electromyographic data is the data at each moment in the rehabilitation training process, and the data at each moment constitutes the electromyographic data), and limiting the signal frequency of the electromyographic data, and eliminating power frequency interference and electrocardiographic noise, to obtain an electromyographic filter signal, and then the electromyographic filter signal is processed by full-wave rectification and low-pass filtering in sequence to obtain an electromyographic activation signal, as follows:

[0068] Rehabilitation training can adopt the Schross method (the core of this method is to improve spinal curvature through three-dimensional corrective exercises and breathing control), SEAS method (a scientific scoliosis exercise method that emphasizes self-correction and functional exercise), Lyon Method (also known as the Lyon method), BSPTS method, etc.

[0069] A bandpass filter is used to limit the signal frequency of the electromyographic data to a preset frequency band (in this embodiment, the preset frequency band is 20 to 450 Hz), and a comb notch filter is used to eliminate power frequency interference. Then, an adaptive filter is used to filter out the ECG noise to obtain an electromyographic filter signal, which is then processed through full-wave rectification and low-pass filtering to obtain an electromyographic activation signal.

[0070] Step S2: pre-process the electromyographic filter signal and the electromyographic activation signal to obtain electromyographic signal segments and electromyographic activation segments respectively, as follows:

[0071] The extended Teager-Kaiser energy operator is used to detect the start time of the electromyographic filter signal and the electromyographic activation signal respectively, and the start and end time nodes of the muscle contraction are obtained (the extended Teager-Kaiser energy operator detects the start time of the signal, that is, detects the start and end time nodes of the muscle contraction, and then obtains the muscle contraction time segment and the relaxation time segment), and the segments of the active correction period and the continuous maintenance period of the electromyographic filter signal are segmented as electromyographic signal segments, and the segments of the active correction period and the continuous maintenance period of the electromyographic activation signal are segmented as electromyographic activation segments, wherein the active correction period refers to the time segment when the patient maintains the action posture of the rehabilitation training and the muscle contracts, and the continuous maintenance period refers to the time segment when the patient maintains the action posture of the rehabilitation training and the muscle relaxes; wherein the rehabilitation training is composed of multiple actions, and the action postures in the active correction period and the continuous maintenance period can be any action posture (wherein the extended Teager-Kaiser energy operator (extended Teager-Kaiser energy operator, ETKEO) to detect the start time of the signal and obtain the start and end time nodes of muscle contraction, which belongs to the prior art content. For details, see the document TigriniA, MengarelliA, CardarelliS,et al.Improving EMG signal change point detectionfor low SNR by using extended Teager-Kaiser energy operator[J].IEEETransactions on Medical Robotics and Bionics,2020,2(4):661-669.);

[0072] The extended Teague-Kaiser energy operator is calculated by the electromyographic activation signal, and the calculation formula is as follows:

[0073] Ψ (d,m,s) [n] = det(X n );

[0074] in,

[0075]

[0076] Among them, X n represents a matrix of d×d dimensions, x[n] represents the value of the nth sampling point of the electromyographic activation signal, s represents the step size (i.e., the sampling interval selected every number of sampling points in time), m represents the starting offset (i.e., the starting offset of the electromyographic activation signal in time), Ψ (d,m,s) [n] represents the extended Teague-Kaiser energy operator, det represents the function for calculating the determinant of the matrix, and d represents X n The order of .

[0077] Step S3, analyzing the electromyographic signal segments by time domain waveform analysis and sample entropy analysis to obtain rehabilitation movement amplitude characteristics, analyzing the electromyographic signal segments by power spectrum analysis to obtain rehabilitation movement frequency characteristics, and analyzing the electromyographic signal segments by wavelet transform to obtain rehabilitation movement wavelet coefficient characteristics;

[0078] The rehabilitation action frequency characteristics include: peak frequency, median frequency, average power frequency, low frequency ratio and the change rate of each frequency of the electromyographic signal segment;

[0079] The characteristics of the wavelet coefficients of rehabilitation movements include: the absolute mean and root mean square value of the wavelet coefficients of rehabilitation movements in the EMG signal segment;

[0080] The rehabilitation movement amplitude characteristics include: the root mean square value, absolute mean, number of zero crossings, signal segment length, number of slope sign changes, Willison amplitude and rehabilitation movement signal entropy of the electromyographic signal segment, wherein the root mean square value, absolute mean, number of zero crossings, signal segment length, number of slope sign changes and Willison amplitude of the electromyographic signal segment are obtained through time domain waveform analysis, and the rehabilitation movement signal entropy is obtained through sample entropy analysis, and the calculation formula of the rehabilitation movement signal entropy is as follows:

[0081]

[0082] in,

[0083]

[0084] d(X i ,X j )=max k＝0,...,m-1(x(i+k)-x(j+k));

[0085] X i =[x(i),x(i+1),…,x(i+m-1)];

[0086] Among them, SampEn(m,r,N) represents the rehabilitation action signal entropy, B m (r) represents the ratio of signal vectors of length m that meet the similarity condition, B m+1 (r) represents the ratio of signal vectors of length m+1 that meet the similarity condition, I(d(X i ,X j )<r) represents an indicator function, which is used to determine whether the distance between two signal vectors is less than a threshold r (wherein the threshold r is set according to actual needs), X i The vector representing the i-th signal segment, X j represents the vector of the jth signal segment, k is the index used to locate within the signal segment, and is used to calculate the difference between the two vectors at each corresponding point;

[0087] The peak frequency, median frequency, average power frequency and low frequency ratio in the rehabilitation movement amplitude characteristics and rehabilitation movement frequency characteristics are obtained from a single EMG signal segment;

[0088] The change rate of each frequency in the rehabilitation movement frequency characteristics is obtained from multiple electromyographic signal segments.

[0089] Step S4, analyzing the electromyographic activation segment through time domain waveform to obtain activation amplitude characteristics of rehabilitation movements;

[0090] Among them, the activation amplitude characteristics of rehabilitation movements include: the mean, variance and root mean square value of the electromyographic activation segment;

[0091] The activation amplitude characteristics of rehabilitation movements are obtained from single EMG activation segments.

[0092] Step S5, inputting the rehabilitation movement amplitude feature, the rehabilitation movement frequency feature, the rehabilitation movement wavelet coefficient feature and the rehabilitation movement activation amplitude feature into the feature selector to obtain the filtered features, and then inputting the filtered features into the classifier at the same time to obtain the discrimination result of the completion degree of the rehabilitation training (specifically, the output is a value of 0 or 1, for example, 0 can be preset to indicate a low completion degree and 1 to indicate a high completion degree);

[0093] Among them, the feature selector adopts a two-level structure, the first level adopts the elastic network regularization algorithm, and the second level adopts the sequential floating feature selection algorithm; the feature selector is used to screen each feature to reduce the complexity;

[0094] The classifier adopts the support vector machine algorithm (the elastic network regularization algorithm, the elastic network regularization algorithm and the support vector machine algorithm are all existing technologies, and the specific process will not be described again).

[0095] Step S6, when the discrimination result is low (when the discrimination result is high, no processing is performed), a myoelectric-activation coupling model is established for the bilateral muscles on the back of the patient's trunk (the degree of muscle contraction is calculated through the myoelectric-activation coupling model, the muscle activation problem of the bilateral muscle positions on the back of the patient's trunk is identified, and whether the muscle activation at the corresponding position is sufficient), and the model is as follows:

[0096]

[0097] in,

[0098] u j (t) = αe j (td)-β1u j (t-1)-β2u j (t-2);

[0099] Among them, a j (t) represents the muscle activation degree at time t, e j (t) represents the EMG activation segment at time t, u j (t) represents the EMG activation segment after muscle activation processing, A represents the nonlinear shape factor, which is generally constrained to be -3<A<0, α represents the gain coefficient, β1 and β2 both represent the recursive coefficient, and d represents the EMG delay;

[0100] The obtained a j (t) is compared with the first preset value and the second preset value. j (t) is less than the first preset value or a j (t) is less than a second preset value, information about insufficient muscle activation at the corresponding moment is generated and fed back to the patient (to assist the patient and the doctor in optimizing the movements of rehabilitation training and improving the completion of the movements), and the patient needs to increase the muscle force at the corresponding moment to solve the insufficient muscle activation;

[0101] Among them, the first preset value The muscle activation degree obtained by inputting the electromyographic activation segment corresponding to the sustained maintenance period into the electromyographic-activation degree coupling model, the second preset value The muscle activation degree is obtained by inputting the EMG activation segment corresponding to the standing posture and muscle relaxation into the EMG-activation degree coupling model.

[0102] In another embodiment, Figure 1-Figure 2As shown in the figure, based on a wearable scoliosis rehabilitation monitoring method, a wearable scoliosis rehabilitation monitoring system is proposed, including an electromyography acquisition device and a host computer, the host computer includes a data processing module and a rehabilitation action guidance feedback module (the host computer adopts a serial port configuration of 460800bps, 1 stop bit, no parity check bit, and 8 data bits), wherein:

[0103] An electromyographic acquisition device is used to obtain electromyographic data of bilateral muscles on the back of the patient's trunk during rehabilitation training (the electromyographic acquisition device is attached to the bilateral muscles on the back of the patient's trunk by pasting, thereby achieving wearability), wherein the electromyographic acquisition device is attached to the bilateral muscles on the back of the patient's trunk at a preset distance from the midline of the spine (this preset distance is 3 to 10 cm), and collects electromyographic data, and sends the electromyographic data to the data processing module of the host computer; wherein, facing scoliosis patients with different types of scoliosis, the attachment position of the electromyographic acquisition device is different. If it is a C-type curvature, the electromyographic acquisition device should be attached to the muscle group area with the largest Cobb angle. If it is an S-type curvature, the electromyographic acquisition device should be attached to the muscle group areas with the largest Cobb angles on both sides of the S-shaped spine. C-type curvature refers to the curvature of the spine that deviates to one side, forming a "C" shaped curve; S-type curvature refers to at least two spinal curvatures that deviate from the midline of the spine, forming an "S" shaped curve;

[0104] The electromyography acquisition device includes an analog front-end module, an analog-to-digital conversion module and a wireless transmission module, wherein:

[0105] The analog front-end module is used to obtain the electromyographic data of the bilateral muscles of the back of the patient's trunk during rehabilitation training, and perform low-pass filtering and amplification processing on the electromyographic data signals in sequence. It has a high common-mode rejection ratio, provides a common-mode bias voltage, reduces the common-mode noise at the input end, and sends the electromyographic data after reducing the common-mode noise to the analog-to-digital conversion module;

[0106] The analog-to-digital conversion module converts the analog electromyographic signal of the received electromyographic data into a discrete digital electromyographic signal, and sends the discrete digital electromyographic signal of the electromyographic data to the wireless transmission module;

[0107] The wireless sending module sends the discrete digital electromyographic signals of the received electromyographic data to the data processing module of the host computer for processing (wherein the wireless sending module packages the discrete digital electromyographic signals of the received electromyographic data and then transmits them to the data processing module of the host computer via wireless, and the data processing module unpacks the received packages to obtain the discrete digital electromyographic signals of the electromyographic data).

[0108] In this embodiment, the structure of the packet is shown in Table 1. Each packet is 216 bytes long, of which the electromyographic data occupies 210 bytes; the packet sequence occupies 2 bytes, which is used to indicate the order of this data packet; the address occupies 1 byte, which is used to distinguish which electromyographic acquisition device the data comes from; the indicator occupies 3 bytes, which is used to indicate the end of a packet. Since each electromyographic data occupies 1.5 bytes, 2 electromyographic data are combined into a 3-byte combination for transmission. For example, 7560f3 in Table 1 can be disassembled into 075f36, which is 117, -202 after unpacking according to the signed number, and Table 1 is as follows:

[0109] Table 1

[0110]

[0111] The data processing module is used to limit the frequency of the received electromyographic data signal (i.e., discrete digital electromyographic signal) and eliminate power frequency interference and electrocardiographic noise to obtain an electromyographic filter signal, and then process the electromyographic filter signal through full-wave rectification and low-pass filtering in sequence to obtain an electromyographic activation signal;

[0112] The data processing module is also used to pre-process the electromyographic filter signal and the electromyographic activation signal to obtain the electromyographic signal segment and the electromyographic activation segment respectively;

[0113] The data processing module is also used to obtain the rehabilitation movement amplitude characteristics by time domain waveform analysis and sample entropy analysis of the electromyographic signal segments, obtain the rehabilitation movement frequency characteristics by power spectrum analysis of the electromyographic signal segments, and obtain the rehabilitation movement wavelet coefficient characteristics by wavelet transform analysis of the electromyographic signal segments;

[0114] The data processing module is also used to obtain activation amplitude characteristics of rehabilitation movements by analyzing the myoelectric activation segments through time domain waveforms;

[0115] The data processing module is also used to input the rehabilitation movement amplitude feature, rehabilitation movement frequency feature, rehabilitation movement wavelet coefficient feature and rehabilitation movement activation amplitude feature into the feature selector to obtain the filtered features, and then input the filtered features into the classifier at the same time to obtain the judgment result of the completion degree of rehabilitation training, and send the judgment result to the rehabilitation movement guidance feedback module;

[0116] The rehabilitation action guidance feedback module is used to establish an electromyographic-activation coupling model for the bilateral muscles of the patient's trunk back when the discrimination result is low (no processing is required when the discrimination result is high), and the model is as follows:

[0117]

[0118] in,

[0119] u j (t) = αej (td)-β1u j (t-1)-β2u j (t-2);

[0120] Among them, a j (t) represents the muscle activation degree at time t, e j (t) represents the EMG activation segment at time t, u j (t) represents the EMG activation segment after muscle activation processing, A represents the nonlinear shape factor, which is generally constrained to be -3<A<0, α represents the gain coefficient, β1 and β2 both represent the recursive coefficient, and d represents the EMG delay;

[0121] The rehabilitation action guidance feedback module is also used to obtain a j (t) is compared with the first preset value and the second preset value. j (t) is less than the first preset value or a j (t) is less than the second preset value, information about insufficient muscle activation at the corresponding moment is generated as feedback to the patient; otherwise, no information is generated as feedback to the patient.

[0122] In this embodiment, the host computer also includes a data storage module and a waveform display module, wherein:

[0123] The data processing module is also used to send the discrete digital electromyographic signal of the received electromyographic data to the data storage module for storage, and send the waveform of the discrete digital electromyographic signal of the electromyographic data to the display module for waveform display.

[0124] The specific definition of a wearable scoliosis rehabilitation monitoring system can be found in the above definition of a wearable scoliosis rehabilitation monitoring method, which will not be repeated here. Each module in the above wearable scoliosis rehabilitation monitoring system can be implemented in whole or in part by software, hardware and a combination thereof.

[0125] The wearable scoliosis rehabilitation monitoring method and system obtain the electromyographic data of the patient during the rehabilitation training process, and obtain the rehabilitation movement amplitude characteristics, rehabilitation movement frequency characteristics, rehabilitation movement wavelet coefficient characteristics and rehabilitation movement activation amplitude characteristics in the electromyographic data, and input each feature into a feature selector for screening, and then input it into a classifier to obtain a judgment result of the degree of completion of the rehabilitation training, thereby monitoring the patient's rehabilitation training process and assisting the patient in optimizing the rehabilitation training movements; at the same time, when the judgment result is low, a myoelectric-activation degree coupling model is established to calculate whether the muscle activation degree corresponding to each moment of the electromyographic activation segment is sufficient, and information on insufficient muscle activation degree at the corresponding moment is generated and fed back to the patient, so as to assist the patient and the doctor in optimizing the rehabilitation training movements and improving the degree of completion of the movements.

[0126] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0127] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A wearable scoliosis rehabilitation monitoring method, characterized in that: The wearable scoliosis rehabilitation monitoring method comprises: The electromyographic data of the bilateral muscles of the back of the trunk of the patient during rehabilitation training are obtained, and the signal frequency of the electromyographic data is limited, and the power frequency interference and electrocardiographic noise are eliminated to obtain the electromyographic filter signal, and then the electromyographic filter signal is processed by full-wave rectification and low-pass filtering in sequence to obtain the electromyographic activation signal; Preprocessing the electromyographic filter signal and the electromyographic activation signal respectively obtains the electromyographic signal segment and the electromyographic activation segment in turn; The EMG signal segments are analyzed by time domain waveform analysis and sample entropy analysis to obtain the amplitude characteristics of the rehabilitation movement, the EMG signal segments are analyzed by power spectrum analysis to obtain the frequency characteristics of the rehabilitation movement, and the EMG signal segments are analyzed by wavelet transform to obtain the wavelet coefficient characteristics of the rehabilitation movement; The activation amplitude characteristics of rehabilitation movements are obtained by analyzing the EMG activation segments through time domain waveform analysis. Input the rehabilitation movement amplitude feature, rehabilitation movement frequency feature, rehabilitation movement wavelet coefficient feature and rehabilitation movement activation amplitude feature into the feature selector to obtain the filtered features, and then input the filtered features into the classifier at the same time to obtain the judgment result of the completion degree of rehabilitation training; When the judgment result is low, the myoelectricity-activation coupling model is established for the bilateral muscles of the patient's back trunk, and the model is as follows: in, u j (t)=αe j (t-d)-β1u j (t-1)-β2u j (t-2); Among them, a j (t) represents the muscle activation degree at time t, e j (t) represents the EMG activation segment at time t, u j (t) represents the EMG activation segment after muscle activation processing, A represents the nonlinear shape factor, which is generally constrained to be -3<A<0, α represents the gain coefficient, β1 and β2 both represent the recursive coefficient, and d represents the EMG delay; The obtained a j (t) is compared with the first preset value and the second preset value. j (t) is less than the first preset value or a j (t) is less than a second preset value, information about insufficient muscle activation at the corresponding moment is generated and fed back to the patient.

2. The wearable scoliosis rehabilitation monitoring method according to claim 1, characterized in that: The method obtains electromyographic data of the bilateral muscles of the back of the trunk of the patient during rehabilitation training, limits the signal frequency of the electromyographic data, eliminates power frequency interference and electrocardiographic noise, obtains an electromyographic filter signal, and then processes the electromyographic filter signal by full-wave rectification and low-pass filtering in sequence to obtain an electromyographic activation signal, including: A bandpass filter is used to limit the signal frequency of the electromyographic data to a preset frequency band, and a comb notch filter is used to eliminate power frequency interference. Then, an adaptive filter is used to filter out the ECG noise to obtain the electromyographic filter signal, which is then processed through full-wave rectification and low-pass filtering to obtain the electromyographic activation signal.

3. The wearable scoliosis rehabilitation monitoring method according to claim 1, characterized in that: The preprocessing of the electromyographic filter signal and the electromyographic activation signal to obtain the electromyographic signal segment and the electromyographic activation segment respectively comprises: The extended Teague-Kaiser energy operator is used to detect the starting time of the electromyographic filter signal and the electromyographic activation signal respectively, and the start and end time nodes of the muscle contraction are obtained. The segments of the active correction period and the continuous maintenance period of the electromyographic filter signal are segmented as electromyographic signal segments, and the segments of the active correction period and the continuous maintenance period of the electromyographic activation signal are segmented as electromyographic activation segments. The active correction period refers to the time segment when the patient maintains the posture of rehabilitation training and the muscles contract, and the continuous maintenance period refers to the time segment when the patient maintains the posture of rehabilitation training and the muscles relax.

4. The wearable scoliosis rehabilitation monitoring method according to claim 3, characterized in that: The extended Teague-Kaiser energy operator is calculated by the electromyographic activation signal, and the calculation formula is as follows: Ψ (d,m,s) [n]=it(X n ); in, Among them, X n represents a matrix of d×d dimensions, x[n] represents the value of the nth sampling point of the electromyographic activation signal, s represents the step size, m represents the starting offset, Ψ (d,m,s) [n] represents the extended Teague-Kaiser energy operator, det represents the function for calculating the determinant of the matrix, and d represents X n The order of .

5. The wearable scoliosis rehabilitation monitoring method according to claim 1, characterized in that: The rehabilitation movement frequency characteristics include: peak frequency, median frequency, average power frequency, low frequency ratio and change rate of each frequency of the electromyographic signal segment; The rehabilitation action wavelet coefficient characteristics include: the absolute mean and root mean square value of the rehabilitation action wavelet coefficients of the electromyographic signal segment; The activation amplitude characteristics of the rehabilitation movement include: the mean, variance and root mean square value of the electromyographic activation segment; The rehabilitation movement amplitude characteristics include: the root mean square value, absolute mean, number of zero crossings, signal segment length, number of slope sign changes, Willison amplitude and rehabilitation movement signal entropy of the electromyographic signal segment, wherein the root mean square value, absolute mean, number of zero crossings, signal segment length, number of slope sign changes and Willison amplitude of the electromyographic signal segment are obtained through time domain waveform analysis, and the rehabilitation movement signal entropy is obtained through sample entropy analysis, and the calculation formula of the rehabilitation movement signal entropy is as follows: in, d(X i ,X j )=max k＝0,...,m-1 (x(i+k)-x(j+k)); X i =[x(i),x(i+1),…,x(i+m-1)]; Among them, SampEn(m,r,N) represents the rehabilitation action signal entropy, B m (r) represents the ratio of signal vectors of length m that meet the similarity condition, B m+1 (r) represents the ratio of signal vectors of length m+1 that meet the similarity condition, I(d(X i ,X j )<r) represents an indicator function, which is used to determine whether the distance between two signal vectors is less than a threshold value r. i The vector representing the i-th signal segment, X j represents the vector of the jth signal segment, k is the index used to locate within the signal segment, and is used to calculate the difference between the two vectors at each corresponding point.

6. The wearable scoliosis rehabilitation monitoring method according to claim 1, characterized in that: The feature selector adopts a two-level structure, the first level adopts an elastic network regularization algorithm, and the second level adopts a sequential floating feature selection algorithm; The classifier adopts support vector machine algorithm.

7. The wearable scoliosis rehabilitation monitoring method according to claim 3, characterized in that: The first preset value The muscle activation degree obtained by inputting the electromyographic activation segment corresponding to the sustained maintenance period into the electromyographic-activation degree coupling model, the second preset value The muscle activation degree is obtained by inputting the EMG activation segment corresponding to the standing posture and muscle relaxation into the EMG-activation degree coupling model.

8. A wearable scoliosis rehabilitation monitoring system, characterized in that: The wearable scoliosis rehabilitation monitoring system includes an electromyography acquisition device and a host computer, wherein the host computer includes a data processing module and a rehabilitation action guidance feedback module, wherein: The electromyographic acquisition device is used to obtain electromyographic data of the bilateral muscles on the back of the patient's trunk during rehabilitation training, wherein the electromyographic acquisition device is attached to the bilateral muscles on the back of the patient's trunk at a preset distance from the midline of the spine, and collects the electromyographic data, and sends the electromyographic data to the data processing module of the host computer; The data processing module is used to limit the signal frequency of the received electromyographic data and eliminate power frequency interference and electrocardiographic noise to obtain an electromyographic filter signal, and then process the electromyographic filter signal through full-wave rectification and low-pass filtering in sequence to obtain an electromyographic activation signal; The data processing module is further used to pre-process the electromyographic filter signal and the electromyographic activation signal to obtain the electromyographic signal segment and the electromyographic activation segment respectively; The data processing module is further used to obtain the rehabilitation movement amplitude characteristics by performing time domain waveform analysis and sample entropy analysis on the electromyographic signal segments, obtain the rehabilitation movement frequency characteristics by performing power spectrum analysis on the electromyographic signal segments, and obtain the rehabilitation movement wavelet coefficient characteristics by performing wavelet transform analysis on the electromyographic signal segments; The data processing module is also used to obtain activation amplitude characteristics of rehabilitation movements by analyzing the myoelectric activation segments through time domain waveforms; The data processing module is further used to input the rehabilitation movement amplitude feature, rehabilitation movement frequency feature, rehabilitation movement wavelet coefficient feature and rehabilitation movement activation amplitude feature into the feature selector to obtain the filtered features, and then input the filtered features into the classifier at the same time to obtain the judgment result of the completion degree of rehabilitation training, and send the judgment result to the rehabilitation movement guidance feedback module; The rehabilitation action guidance feedback module is used to establish an electromyographic-activation coupling model for the bilateral muscles on the back of the patient's trunk when the judgment result is low, and the model is as follows: in, u j (t)=αe j (t-d)-β1u j (t-1)-β2u j (t-2); Among them, a j (t) represents the muscle activation degree at time t, e j (t) represents the EMG activation segment at time t, u j (t) represents the EMG activation segment after muscle activation processing, A represents the nonlinear shape factor, which is generally constrained to be -3<A<0, α represents the gain coefficient, β1 and β2 both represent the recursive coefficient, and d represents the EMG delay; The rehabilitation action guidance feedback module is also used to obtain a j (t) is compared with the first preset value and the second preset value. j (t) is less than the first preset value or a j (t) is less than a second preset value, information about insufficient muscle activation at the corresponding moment is generated and fed back to the patient.

9. The wearable scoliosis rehabilitation monitoring system according to claim 8, characterized in that: The electromyography acquisition device includes an analog front-end module, an analog-to-digital conversion module and a wireless transmission module, wherein: The analog front-end module is used to obtain the electromyographic data of the bilateral muscles of the back of the patient's trunk during rehabilitation training, and perform low-pass filtering and amplification processing on the electromyographic data signals in sequence, and has a high common-mode rejection ratio, provides a common-mode bias voltage, reduces the common-mode noise at the input end, and sends the electromyographic data after the common-mode noise is reduced to the analog-to-digital conversion module; The analog-to-digital conversion module converts the analog electromyographic signal of the received electromyographic data into a discrete digital electromyographic signal, and sends the discrete digital electromyographic signal of the electromyographic data to the wireless transmission module; The wireless transmission module sends the received discrete digital electromyographic signals to the data processing module of the host computer for processing.

10. The wearable scoliosis rehabilitation monitoring system according to claim 9, characterized in that: The host computer also includes a data storage module and a waveform display module, wherein: The data processing module is also used to send the discrete digital electromyographic signal of the received electromyographic data to the data storage module for storage, and send the waveform of the discrete digital electromyographic signal of the electromyographic data to the display module for waveform display.

Images