CN113598759A - Lower limb action recognition method and system based on myoelectric feature optimization - Google Patents

Lower limb action recognition method and system based on myoelectric feature optimization Download PDF

Info

Publication number
CN113598759A
CN113598759A CN202111066419.3A CN202111066419A CN113598759A CN 113598759 A CN113598759 A CN 113598759A CN 202111066419 A CN202111066419 A CN 202111066419A CN 113598759 A CN113598759 A CN 113598759A
Authority
CN
China
Prior art keywords
lower limb
optimization
electromyographic
signal
characteristic
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202111066419.3A
Other languages
Chinese (zh)
Other versions
CN113598759B (en
Inventor
曹佃国
王加帅
武玉强
张中才
王金强
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Qufu Normal University
Original Assignee
Qufu Normal University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Qufu Normal University filed Critical Qufu Normal University
Priority to CN202111066419.3A priority Critical patent/CN113598759B/en
Publication of CN113598759A publication Critical patent/CN113598759A/en
Application granted granted Critical
Publication of CN113598759B publication Critical patent/CN113598759B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
    • A61B5/1123—Discriminating type of movement, e.g. walking or running
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
    • A61B5/1126—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb using a particular sensing technique
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/389—Electromyography [EMG]
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/389—Electromyography [EMG]
    • A61B5/397—Analysis of electromyograms
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/725—Details of waveform analysis using specific filters therefor, e.g. Kalman or adaptive filters
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7253—Details of waveform analysis characterised by using transforms
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
    • A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Public Health (AREA)
  • Surgery (AREA)
  • Veterinary Medicine (AREA)
  • General Health & Medical Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Physiology (AREA)
  • Artificial Intelligence (AREA)
  • Signal Processing (AREA)
  • Psychiatry (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Dentistry (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Evolutionary Computation (AREA)
  • Fuzzy Systems (AREA)
  • Mathematical Physics (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)

Abstract

The invention provides a lower limb action recognition method and system based on myoelectric feature optimization, wherein the method comprises the following steps: acquiring a plurality of electromyographic signals corresponding to the lower limb actions in real time; extracting main characteristic segment signals of the plurality of electromyographic signals respectively; respectively extracting the characteristics of each main characteristic section signal, and performing characteristic optimization according to the correlation degree between the electromyographic signal energy and the action; fusing the optimized features to obtain a feature vector; and based on the feature vector, adopting a pre-trained lower limb action recognition model to recognize the lower limb action. According to the invention, through introducing the weighting characteristic optimization based on the muscle and motion correlation degree, the characteristic redundancy is reduced, meanwhile, as much useful information as possible is reserved, and the identification efficiency and the accuracy are improved.

Description

Lower limb action recognition method and system based on myoelectric feature optimization
Technical Field
The invention belongs to the technical field of motion recognition, and particularly relates to a lower limb motion recognition method and system based on myoelectric feature optimization.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
The lower limb action recognition technology based on the surface electromyogram signal has the characteristics of safety, real time and convenience, and is widely applied to the fields of sports medicine, biomedicine, rehabilitation engineering, intelligent artificial limbs and the like.
The lower limb motion recognition technology generally comprises a plurality of stages of feature extraction, processing and feature-based classification. For example, chinese patent CN110238863B proposes a lower limb rehabilitation robot control method and system based on electroencephalogram-electromyogram signals, which performs 3 kinds of simple lower limb motion recognition by utilizing electroencephalogram-electromyogram coherence analysis. Chinese patent CN107092861B proposes a lower limb action recognition method based on pressure and acceleration sensors, after extracting the required features, adopting dimension reduction processing to reduce the 48-dimensional features to 5-dimensional, and recognizing 6 simple lower limb actions through a simple SVM classifier. Shixin and the like of Chongqing university automation research institute provide a lower limb movement feature extraction and classification method based on surface electromyographic signals, and 5 simple lower limb actions are identified by sEMG offline through simple feature fusion.
In the prior art including the implementation method, in the feature processing process of the electromyographic signals corresponding to the lower limb actions, part of researchers perform dimension reduction processing on the features, so that part of effective features can be lost while feature redundancy is reduced, although the data volume can be reduced and the recognition speed can be improved, the discrimination of the actions is reduced, and the recognition rate is reduced; some researchers do not process the extracted features, and input the extracted features into a classifier for recognition after simple feature fusion, so that although all effective signals are reserved, the recognition speed is low, the delay is high, the timeliness is poor, and the method has no practical popularization and application value. In the process of training a model by a classifier, part of personnel use the most original support vector machine to carry out simple classification without a parameter optimization process, the recognition rate cannot be guaranteed, the parameter optimization is carried out by a genetic algorithm, better parameters can be obtained, the recognition rate is improved to a certain extent, but a traditional genetic algorithm selection operator is easy to fall into a local optimal solution, the adaptability is poor, and the method is not suitable for actual motion recognition.
Disclosure of Invention
In order to overcome the defects of low identification rate, poor adaptability and low practical application value of the classification method, the invention provides a lower limb action identification method and system based on electromyographic signal characteristic optimization, introduces a weighting characteristic optimization algorithm based on muscle and motion correlation, retains as much useful information as possible while reducing characteristic redundancy, and improves identification rate while improving identification speed.
In order to achieve the above object, one or more embodiments of the present invention provide the following technical solutions:
a lower limb action recognition method based on myoelectric feature optimization comprises the following steps:
acquiring a plurality of electromyographic signals corresponding to the lower limb actions in real time;
extracting main characteristic segment signals of the plurality of electromyographic signals respectively;
respectively extracting the characteristics of each main characteristic section signal, and performing characteristic optimization according to the correlation degree between the electromyographic signal energy and the action;
and fusing the optimized features to obtain a feature vector, and based on the feature vector, adopting a pre-trained lower limb action recognition model to recognize the lower limb action.
Further, after acquiring a plurality of electromyographic signals of the lower limb, denoising the plurality of electromyographic signals: the wavelet transformation threshold denoising method and the digital filtering threshold denoising method remove baseline drift, high-frequency noise and low-frequency noise, and the trap is used for removing 50Hz power frequency interference.
Further, the main characteristic segment signal extraction adopts a frame energy method:
framing the electromyographic signals, and calculating the total energy of the signals in each frame;
if the total energy in a certain frame is greater than a set threshold and the total energy in the following set number of frames is greater than the set threshold, taking the frame as the initial frame of an action segment;
extracting the electromyographic signals with set time from the initial frame to obtain main characteristic segment signals.
Further, the characteristic optimization according to the correlation degree between the electromyographic signal energy and the action comprises the following steps:
extracting the characteristics of the main characteristic segment signals;
acquiring a signal energy value of a channel corresponding to each muscle in the main characteristic segment signal;
calculating the weight of the channel in the main characteristic section signal according to the proportion of the signal energy value of each channel in the accumulated sum of the signal energy of all the channels;
and performing weighted optimization on the characteristics of each channel signal of the main characteristic section signal according to the weight to obtain optimized characteristics.
Further, the lower limb action recognition model is obtained by training based on a support vector machine classifier.
Further, when the support vector machine classifier is used for training, a genetic algorithm based on the combination of championship and sequencing is adopted for parameter optimization.
Further, the parameter optimization process includes:
(1) determining the initial population number, and calculating the fitness of individuals in the population based on a feature vector obtained by feature extraction, optimization and fusion according to training data;
(2) sequencing the individuals in the population from small to large in sequence according to the fitness;
(3) dividing the sorted individuals into a plurality of grades in sequence;
(4) according to the principle that more superior individuals are selected and less inferior individuals are selected, the diversity of the population is considered as much as possible under the condition that all the superior individuals exist, the multiple grades of individuals are selected according to set probabilities respectively, and then individual interpolation and cross mutation are carried out to obtain a new population;
(5) and (3) calculating the fitness of the population, if the fitness does not reach a target value, repeating the steps (2) - (5) until the fitness reaches the target value, and applying the punishment factor and the numerical value of the nuclear parameter to the support vector machine.
One or more embodiments provide a lower limb movement recognition system based on electromyographic signal feature optimization, including:
the signal acquisition module is used for acquiring a plurality of electromyographic signals corresponding to the lower limb actions in real time;
the signal preprocessing module is used for respectively extracting main characteristic section signals of the plurality of electromyographic signals;
the characteristic extraction module is used for respectively extracting the characteristics of each main characteristic section signal;
the characteristic optimization module is used for carrying out characteristic optimization on the extracted characteristics according to the correlation degree between the electromyographic signal energy and the action;
the feature fusion module is used for fusing the optimized features to obtain a feature vector;
and the action recognition module is used for recognizing the lower limb actions by adopting a pre-trained lower limb action recognition model based on the characteristic vector.
One or more embodiments provide an electronic device, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to implement the lower limb movement recognition method based on electromyographic signal characteristic optimization.
One or more embodiments provide a computer-readable storage medium storing a computer program which, when executed by a processor, implements the lower limb motion recognition method based on electromyogram signal feature optimization.
The above one or more technical solutions have the following beneficial effects:
the method comprises the steps of providing a weighting characteristic optimization method based on muscle and motion correlation, weighting each channel according to the proportion of different channel energy values in each electromyographic signal, and enabling the electromyographic signals of the same muscle to have stronger discrimination on different actions; by carrying out weighted optimization on the features, the redundancy of the same features under different actions can be effectively reduced, the distinguishing degree of the features is improved, the optimal feature vector containing all effective information is obtained, the characteristic of improving the identification rate of complex actions is achieved, and the complex 6 lower limb actions can be identified by utilizing a single signal type surface myoelectric signal.
The genetic algorithm is improved by combining championship and sequencing, the problem that the selection operator is easy to fall into a local optimal solution is solved, a global optimal solution with strong adaptability is obtained, and useful information is kept more comprehensively, so that the genetic algorithm has better adaptability and is convenient for practical application. By improving two steps in the identification process, the timeliness and the accuracy of lower limb action identification are improved, the adaptability of a classification model is enhanced, online verification is carried out on the method through online action identification, a high online action identification rate is obtained, and the method has a high practical application value.
The main characteristic section signals are extracted through a frame energy method, the data volume is reduced, the characteristic extraction time is shortened, the characteristic precision is improved, and the real-time performance of the system is guaranteed.
The signal is preprocessed and denoised by using a composite wavelet denoising method and a wave trap, so that the high-frequency noise, the baseline drift and the 50Hz power frequency interference of the electromyographic signal can be effectively removed.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the invention and together with the description serve to explain the invention and not to limit the invention.
FIG. 1 is a flowchart of a lower limb movement recognition method based on electromyographic signal characteristic optimization in an embodiment of the invention;
FIG. 2 is a schematic diagram illustrating noise reduction of a surface electromyogram signal according to an embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating detection of an active segment of an electrical signal of a surface muscle for right leg lift according to an embodiment of the present invention;
FIG. 4 is a schematic diagram illustrating extraction of main characteristic segments of the electrical signals of the surface muscles of the right leg lift operation according to the embodiment of the present invention;
FIG. 5 is a schematic diagram illustrating comparison between wavelet packet coefficient energy characteristics of 8 groups of muscles before and after optimization under 6 actions in the embodiment of the present invention;
FIG. 6 is a diagram illustrating a convergence curve of population fitness after improving a genetic algorithm according to an embodiment of the present invention.
Detailed Description
It is to be understood that the following detailed description is exemplary and is intended to provide further explanation of the invention as claimed. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
Example one
The embodiment discloses a lower limb action recognition method based on electromyographic signal characteristic optimization, which can accurately recognize 6 lower limb actions of lifting/putting a right leg, lifting/putting a left leg, sitting to standing and standing to sitting by using 8-channel sEMG signals. The method comprises an offline action recognition classification model training stage and an online action recognition stage, wherein the offline action recognition classification model training stage mainly comprises the steps of signal acquisition, preprocessing, feature extraction, weighted feature optimization based on muscle and motion correlation, feature fusion and classification model training as shown in figure 1; the online action recognition stage comprises signal acquisition, preprocessing, feature extraction, weighted feature optimization based on muscle and motion correlation, feature fusion and the like, and the fused features are input into an offline action recognition classification model to recognize the action of the lower limbs in real time.
Off-line classification model training stage
Step 1: acquiring myoelectric signals of a plurality of muscles corresponding to the lower limb movement.
In this embodiment, surface electromyographic signals of 8 muscles are collected when 4 subjects perform 6 lower limb movements: left/right thigh rectus muscle, left/right thigh biceps, left/right leg tibialis anterior muscle, left/right leg gastrocnemius muscle.
Step 2: and preprocessing the electromyographic signals and extracting main characteristic segment signals.
The surface electromyogram signal sEMG preprocessing comprises two parts of signal noise reduction and main characteristic segment signal extraction. The sEMG signal is a non-stable and non-linear weak electric signal and is often accompanied by noise and interference, so that the quality of data is greatly improved by removing the noise and the interference, and a higher action recognition rate is obtained. The sEMG after noise and interference removal is pure and smooth, but many useless resting signals exist in the sEMG, so that the difficulty of signal processing is increased, and the timeliness is reduced. Therefore, extracting the main feature segment signal which has a small data amount and contains most effective signals in one action from the redundant sEMG signals can improve the timeliness of action identification. In the embodiment, a wavelet transformation threshold denoising method and a digital filtering threshold denoising method are adopted to remove baseline drift, high-frequency noise and low-frequency noise of surface electromyographic signals, a trap is used to remove 50Hz power frequency interference, and a framing energy method is used to extract main characteristic segment signals.
The step 2 specifically comprises:
step 2.1: a wavelet transformation threshold denoising method and a digital filtering threshold denoising method are adopted to remove baseline drift, high-frequency noise and low-frequency noise, a trap is used to remove 50Hz power frequency interference, and a smooth pure signal with a low signal-to-noise ratio is obtained through two steps, wherein the denoising effect is shown in FIG. 2.
Step 2.2: performing main characteristic segment signal extraction based on a framing energy method on the noise-reduced signal:
the main feature segment signal extracted in this embodiment is: and (3) in the range of 0.5s to 2.5s after the starting point of the action activity segment, all the sEMG signals used for the subsequent research are 2s signals. When the main characteristic section signal is extracted, the starting point of the action needs to be detected, and the starting point of the signal can be effectively detected by a signal framing energy method.
The method comprises the following specific steps of extracting a main characteristic segment signal based on a framing energy method:
(1) selecting proper frame length and frame shift, framing the signal:
(M-1)×I+L＝N (1)
wherein M is the total frame number of the signal, I is the incremental frame step length, L is the frame length, i.e. the signal length of each frame, and N is the total length of the signal.
(2) Calculating the total energy En (i) of each frame signal:
Figure BDA0003258571980000071
(3) calculating an adaptive threshold th based on the signal energy at the stationary stand:
Figure BDA0003258571980000072
(4) if En in a frame is greater than th, and in the following 3 frames is greater than th, the frame is the start frame of the signal action segment.
(5) Intercepting 2s data 0.5s after the initial point of the signal action segment to obtain a main characteristic segment signal:
Figure BDA0003258571980000073
wherein SN is the detected starting point, FS is the starting point frame number, FsAnd if the sampling frequency is MSN, the sampling point 0.5s after the initial point is MSN, and MEN is the sampling point 2s after the MSN, the signals from MSN to MEN are main characteristic segment signals.
The method has the advantages that the main characteristic section signals are extracted by adopting a signal framing energy method, the length of the signals needing to be processed is shortened, the data volume is reduced, the timeliness of action identification is improved, the monitoring of the starting point is shown in figure 3, the self-adaptive threshold value is obtained by calculating the energy sum of the rest state, and the starting point and the ending point of the right leg lifting action are judged through the threshold value. As shown in fig. 4, the surface electromyogram signal is divided into a start segment signal, a main feature segment signal, and an end segment signal, and the main feature segment signal is extracted for feature extraction and optimization.
And step 3: and (3) carrying out feature extraction and feature optimization on the main feature segment signals:
in the embodiment, 4 features such as time domain features, time frequency domain features and the like are extracted, and the complexity and the stability of the electromyographic signals are considered, so that the difference of the extracted features under different actions is small, the features are considered to be subjected to weighting optimization based on muscle and motion correlation, the discrimination of the same features under different actions is improved, the data redundancy is reduced, the feature discrimination is increased, and the improvement of the action recognition rate is facilitated.
The step 3 specifically includes:
step 3.1: the features extracted in this example are shown in table 1.
Table 1sEMG motion recognition correlation physiological response characteristics
Figure BDA0003258571980000081
The 4 features in table 1 were calculated as follows:
Figure BDA0003258571980000082
Figure BDA0003258571980000083
Figure BDA0003258571980000084
Figure BDA0003258571980000091
wherein x isiThe amplitude of the ith sampling point of the surface electromyogram signal is shown, M is the total number of windows after the signal is framed, n is the window length, f (x) is equal to 1 when x is larger than or equal to th, otherwise f (x) is equal to 0, SjWavelet packet coefficient for jth window。
Step 3.2: based on the weighted feature optimization of the muscle and motion correlation degree, calculating the signal energy value of a channel corresponding to each muscle when the lower limb performs different actions, establishing an energy meter and obtaining the contribution degree; calculating the correlation degree of the muscles and the movement according to the contribution degree in the energy table, and establishing a correlation table of the muscles and the movement to obtain a correlation coefficient; and according to the correlation coefficient, giving different weights to each channel to complete the optimization of each characteristic. The specific implementation steps are as follows:
(1) recording the extracted main characteristic segment signal as X, and taking an absolute value to obtain absX;
(2) calculating the signal energy value of each channel, namely the amplitude sum to obtain sumabs XiWhere i is 1, 2.. and n is the serial number of the channel, this embodiment uses 8 channels, so n is 8, and establishes an energy meter, see table 2;
(3) and summing the energies of all channels to obtain submebsX, wherein the weight calculation mode of each channel is as follows:
Figure BDA0003258571980000092
wherein, CiEstablishing a characteristic optimization coefficient table for the optimization coefficient of the ith channel, wherein i is the serial number of the channel, and n is the total number of the channels, and the table is shown in a table 3;
(4) extracted features and feature optimization coefficients CiMultiplying to obtain the optimized characteristics.
Figure BDA0003258571980000093
Wherein featureiRespectively represent the extracted semGMAV、sEMGRMS、sEMGWAAnd sEMGEC1Feature, FeatureiIs the optimized characteristic. After the characteristics are subjected to weighted optimization, the redundancy of the characteristics is reduced among different channels under the same action; the discrimination of features is enhanced between the same channels under different actions. Wavelet packet for right leg lifting actionCoefficient energy characteristic sEMGEC1Comparing the before and after optimization with fig. 5, it can be seen that the original wavelet packet coefficient energy characteristic sEMGEC1The redundancy rate before optimization is high and difficult to distinguish, and WFO-EC1 characteristics with low redundancy rate and high distinguishing degree are obtained after weighted characteristic optimization of muscle and motion correlation degree.
TABLE 2 energy meter
Figure BDA0003258571980000101
TABLE 3 muscle-to-exercise correlation
Figure BDA0003258571980000102
And 4, step 4: and (3) improving a genetic algorithm to find the optimal SVM parameter:
in the embodiment, the penalty factor c and the nuclear parameter g are optimized by utilizing the capability of solving the optimal solution by the genetic algorithm, in order to ensure that the solution solved by the genetic algorithm is the global optimal solution, the genetic algorithm is improved by utilizing a method combining championship competition and sequencing, a grade elimination system is established, individuals in a population are sequenced according to the fitness, and are divided into four grades of poor, medium, good and excellent, and when next generation selection is carried out, the populations of the four grades are selected according to a certain proportion, so that the proportion of the excellent individuals of the population is high, and the diversity of the population is maintained.
The specific implementation process of the step 4 is as follows:
(1) determining initial population number, fusing feature vector XpCalculating the fitness of individuals in the population as training data of a genetic algorithm, and recording a matrix recording all parameters and the fitness of the population as oldhop;
(2) sequencing individuals in the population from small to large in sequence according to the fitness value, and recording the sequence as a matrix sortpop;
(3) the well-ordered individuals are sequentially divided into four grades of poor, medium, good and excellent, and are sequentially represented as Cbad,Cmid,Cwell,CgoodAnd C ═ Cbad+Cmid+Cwell+CgoodAnd C is the total population number.
(4) According to the principle that the superior individuals are selected more and the inferior individuals are selected less, the diversity of the population is considered as much as possible under the condition that all the superior individuals exist, and the poor, medium, good and excellent four-grade individuals are selected according to the probabilities of P, P + delta, P +2 delta and P +3 delta (0< P <1 and 0< delta < 1).
(5) And recombining the individuals selected according to the proportion to obtain a new population newcrop, wherein the individuals contained in the population are recorded as:
Cnew＝[Cbad×P]+[Cmid×(P+δ)]+[Cwell×(P+2δ)]+[Cgood×(P+3δ)] (11)
(6) in step (4), a part of the population is discarded, so that the population matrix is incomplete, and interpolation is required in newport. The interpolation principle adopts a principle of superior-inferior selection, if C-C is less than or equal to CgoodThen from CgoodRandomly selecting C-C individuals to be added into a new population newport if C-C is more than or equal to CgoodThen C will begoodAll individuals in (a) are interpolated into a new population and C is selectedwellC-C-C ofgoodAnd (4) interpolating the random individuals into the newpos to enable the population number of the newpos to be C, and performing cross variation in the newpos to obtain more excellent offspring.
(7) And (3) calculating the fitness of the population by adopting a 5-fold cross-validation method, if the fitness does not reach a target value, repeating the steps (2) - (6) until the fitness reaches the target value, and recording the values of the penalty factor c and the nuclear parameter g at the moment to obtain bestc and bestg.
(8) Bestc and bestg are applied to a support vector machine to obtain a training network of an improved genetic algorithm-support vector machine (IGA-SVM), an optimal classification model ClassifyModel is trained through a training set, and finally a test set is input into the ClassifyModel to obtain the optimal action recognition rate.
In the embodiment, the average convergence rate and the average fitness of the population are obviously higher than those of the genetic algorithm before optimization by improving the genetic algorithm, as shown in fig. 6, the improved genetic algorithm has high convergence rate and high fitness, and the improved genetic algorithm is used for optimizing the support vector machine to obtain the IGA-SVM.
And 5: and (3) preprocessing the surface electromyogram signals obtained in the step (1) to extract main characteristic section signals, extracting and optimizing the characteristics in the step (3), improving the genetic algorithm and optimizing a classifier in the step (4), completing the offline classification test of 6 lower limb actions, realizing high recognition rate identification of the lower limb actions, and showing classification results in a table (4).
TABLE 4 test classification results
Figure BDA0003258571980000121
As can be seen from the table, the application of the feature optimization and the improved genetic algorithm-support vector machine classifier enables the average recognition rate of 6 lower limb actions to reach 94.75%.
In the embodiment, the sensors with 8 channels are used for collecting 8 paths of sEMG signals, the lower limb actions in 6 daily lives are successfully identified at a high identification rate of 94.75%, and the identification rate of the lower limb actions is greatly improved.
(II) on-line action recognition stage
The action recognition method can be applied to a rehabilitation platform of the lower limb exoskeleton robot, and drives the patient to do rehabilitation movement by recognizing the lower limb action of the demonstrator.
Training a recognition model of 6 actions of the lower limb through an off-line action recognition stage, and applying the model to an on-line action recognition stage.
Specifically, the recognition models of 6 actions of the lower limbs in the off-line action recognition stage are packed into a dynamic link library through Matlab, new data collected in real time are input into the models, and specific action instructions are returned.
The acquisition and processing of the real-time collected new data comprises the following specific steps:
step 1: firstly, Csharp is used for programming upper computer software and is connected with a Delsys Trigno server so as to obtain a numerical signal of the electromyographic sensor.
Step 2: selecting 8 channels corresponding to the offline action recognition, and placing the channels at 8 muscles of the lower limb: left/right thigh rectus muscle, left/right thigh biceps, left/right leg tibialis anterior muscle, left/right leg gastrocnemius muscle, and collecting surface electromyographic signals.
And step 3: preprocessing electromyographic signals, denoising the signals by a method combining wavelet transform threshold denoising, digital filtering threshold denoising and a trap removing 50Hz power frequency interference, and extracting main characteristic segment signals by a framing energy method.
And 4, step 4: and performing feature extraction on the extracted main feature segment signals, performing weighted feature optimization based on muscle and motion correlation, extracting 4 features such as time domain features and time-frequency domain features in the same way as the feature extraction in the offline action identification stage, and performing weighted optimization on the features to obtain all effective features with low data redundancy and high feature discrimination.
And 5: packing the recognition models of 6 actions of the lower limbs in the off-line action recognition stage into a dynamic link library through Matlab, and implanting the dynamic link library into upper computer software written by Csharp; and 4 kinds of characteristics mentioned in the step 4 are fused into a characteristic vector and input into the action recognition model to finish online action recognition, the average recognition rate of the online action recognition reaches 90.6 percent through verification, and specific classification results are shown in a table 5, wherein actions 1-6 respectively represent right leg lifting, right leg placing, left leg lifting, left leg placing, sitting and standing, and sitting and standing.
TABLE 5 Online action recognition test Classification results
Figure BDA0003258571980000131
The actions recognizable in the online action recognition stage are limited to the 6 types of lower limb actions in the offline action recognition stage, and the greater the number of actions recognized in the offline action recognition, the greater the number of online action recognitions, which is feasible.
Example two
The embodiment aims to provide a lower limb action recognition system based on electromyographic signal characteristic optimization.
A lower limb action recognition system based on electromyographic signal characteristic optimization comprises:
the signal acquisition module is used for acquiring a plurality of electromyographic signals corresponding to the lower limb actions in real time;
the signal preprocessing module is used for respectively extracting main characteristic section signals of the plurality of electromyographic signals;
the characteristic extraction module is used for respectively extracting the characteristics of each main characteristic section signal;
the characteristic optimization module is used for carrying out characteristic optimization on the extracted characteristics according to the correlation degree between the electromyographic signal energy and the action;
the feature fusion module is used for fusing the optimized features to obtain a feature vector;
and the action recognition module is used for recognizing the lower limb actions by adopting a pre-trained lower limb action recognition model based on the characteristic vector.
EXAMPLE III
The embodiment aims at providing an electronic device.
An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the following steps when executing the program:
acquiring a plurality of electromyographic signals corresponding to the lower limb actions in real time;
extracting main characteristic segment signals of the plurality of electromyographic signals respectively;
respectively extracting and optimizing the characteristics of each main characteristic segment signal, and fusing a plurality of optimized characteristics to obtain a characteristic vector;
and based on the feature vector, adopting a pre-trained lower limb action recognition model to recognize the lower limb action.
Example four
An object of the present embodiment is to provide a computer-readable storage medium.
A computer-readable storage medium having stored thereon a computer program, upper computer software, and the like, the program being executed by a processor to perform the steps of:
acquiring a plurality of electromyographic signals corresponding to the lower limb actions in real time;
extracting main characteristic segment signals of the plurality of electromyographic signals respectively;
respectively extracting and optimizing the characteristics of each main characteristic segment signal, and fusing a plurality of optimized characteristics to obtain a characteristic vector;
and based on the feature vector, adopting a pre-trained lower limb action recognition model to recognize the lower limb action.
The steps involved in the second to fourth embodiments all correspond to the first embodiment, and the detailed description thereof can be found in the relevant description of the first embodiment.
One or more technical schemes provide a lower limb action recognition method and system based on electromyographic signal characteristic optimization, a weighting characteristic optimization algorithm based on muscle and motion correlation is provided, the characteristic redundancy is reduced, meanwhile, as much useful information as possible is reserved, the recognition rate is improved while the recognition speed is improved, and the popularization and application of online recognition are facilitated; the genetic algorithm is improved by combining championship and sequencing, the problem that the selection operator is easy to fall into a local optimal solution is solved, a global optimal solution with strong adaptability is obtained, and useful information is kept more comprehensively, so that the genetic algorithm has better adaptability and is convenient for practical application. By improving two steps in the identification process, the timeliness and the accuracy of lower limb action identification are improved, the adaptability of a classification model is enhanced, online verification is carried out on the method through online action identification, a high online action identification rate is obtained, and the method has a high practical application value.
Those skilled in the art will appreciate that the modules or steps of the present invention described above can be implemented using general purpose computer means, or alternatively, they can be implemented using program code that is executable by computing means, such that they are stored in memory means for execution by the computing means, or they are separately fabricated into individual integrated circuit modules, or multiple modules or steps of them are fabricated into a single integrated circuit module. The present invention is not limited to any specific combination of hardware and software.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications, improvements and changes may be made without inventive efforts based on the technical solutions of the present invention.

Claims (10)

1. A lower limb action recognition method based on myoelectric feature optimization is characterized by comprising the following steps:
acquiring a plurality of electromyographic signals corresponding to the lower limb actions in real time;
extracting main characteristic segment signals of the plurality of electromyographic signals respectively;
respectively extracting the characteristics of each main characteristic section signal, and performing characteristic optimization according to the correlation degree between the electromyographic signal energy and the action;
and fusing the optimized features to obtain a feature vector, and based on the feature vector, adopting a pre-trained lower limb action recognition model to recognize the lower limb action.
2. The method for recognizing the lower limb movement based on the electromyographic feature optimization according to claim 1, wherein after acquiring a plurality of electromyographic signals of the lower limb, denoising the plurality of electromyographic signals: and removing baseline drift, high-frequency noise and low-frequency noise by adopting a wavelet transform threshold denoising method and a digital filtering threshold denoising method, and removing 50Hz power frequency interference by using a wave trap.
3. The lower limb movement recognition method based on electromyographic feature optimization according to claim 1, wherein the main feature segment signal extraction adopts a frame energy method:
framing the electromyographic signals, and calculating the total energy of the signals in each frame;
if the total energy in a certain frame is greater than a set threshold and the total energy in the following set number of frames is greater than the set threshold, taking the frame as the initial frame of an action segment;
extracting the electromyographic signals with set time from the initial frame to obtain main characteristic segment signals.
4. The method for recognizing the lower limb movement based on electromyographic characteristic optimization according to claim 1, wherein the characteristic optimization according to the correlation degree between the electromyographic signal energy and the movement comprises the following steps:
extracting the characteristics of the main characteristic segment signals;
acquiring a signal energy value of a channel corresponding to each muscle in the main characteristic segment signal;
calculating the weight of the channel in the main characteristic section signal according to the proportion of the signal energy value of each channel in the accumulated sum of the signal energy of all the channels;
and performing weighted optimization on the characteristics of each channel signal of the main characteristic section signal according to the weight to obtain optimized characteristics.
5. The method for recognizing the lower limb actions based on the electromyographic feature optimization of claim 1, wherein the lower limb action recognition model is obtained by training based on a support vector machine classifier.
6. The method for recognizing the lower limb movement based on the electromyographic feature optimization as claimed in claim 5, wherein a genetic algorithm based on the combination of championship and sequencing is adopted for parameter optimization during training based on the support vector machine classifier.
7. The method for recognizing the lower limb actions based on the electromyographic feature optimization as claimed in claim 6, wherein the parameter optimization process comprises the following steps:
(1) determining the initial population number, and calculating the fitness of individuals in the population based on a feature vector obtained by feature extraction, optimization and fusion according to training data;
(2) sequencing the individuals in the population from small to large in sequence according to the fitness;
(3) dividing the sorted individuals into a plurality of grades in sequence;
(4) according to the principle that more superior individuals are selected and less inferior individuals are selected, the diversity of the population is considered as much as possible under the condition that all the superior individuals exist, the multiple grades of individuals are selected according to set probabilities respectively, and then individual interpolation and cross mutation are carried out to obtain a new population;
(5) and (3) calculating the fitness of the population, if the fitness does not reach a target value, repeating the steps (2) - (5) until the fitness reaches the target value, and applying the punishment factor and the numerical value of the nuclear parameter to the support vector machine.
8. A lower limb action recognition system based on electromyographic signal characteristic optimization is characterized by comprising:
the signal acquisition module is used for acquiring a plurality of electromyographic signals corresponding to the lower limb actions in real time;
the signal preprocessing module is used for respectively extracting main characteristic section signals of the plurality of electromyographic signals;
the characteristic extraction module is used for respectively extracting the characteristics of each main characteristic section signal;
the characteristic optimization module is used for carrying out characteristic optimization on the extracted characteristics according to the correlation degree between the electromyographic signal energy and the action;
the feature fusion module is used for fusing the optimized features to obtain a feature vector;
and the action recognition module is used for recognizing the lower limb actions by adopting a pre-trained lower limb action recognition model based on the characteristic vector.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the method for identifying lower limb movements based on electromyographic signal characteristic optimization according to any one of claims 1 to 7 when executing the program.
10. A computer-readable storage medium storing a computer program, wherein the program is executed by a processor to implement the method for recognizing a lower limb movement based on electromyogram signal feature optimization according to any one of claims 1 to 7.
CN202111066419.3A 2021-09-13 2021-09-13 Myoelectricity feature optimization-based lower limb action recognition method and system Active CN113598759B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111066419.3A CN113598759B (en) 2021-09-13 2021-09-13 Myoelectricity feature optimization-based lower limb action recognition method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111066419.3A CN113598759B (en) 2021-09-13 2021-09-13 Myoelectricity feature optimization-based lower limb action recognition method and system

Publications (2)

Publication Number Publication Date
CN113598759A true CN113598759A (en) 2021-11-05
CN113598759B CN113598759B (en) 2023-09-22

Family

ID=78310376

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111066419.3A Active CN113598759B (en) 2021-09-13 2021-09-13 Myoelectricity feature optimization-based lower limb action recognition method and system

Country Status (1)

Country Link
CN (1) CN113598759B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115034273A (en) * 2021-12-27 2022-09-09 驻马店市中心医院 Myoelectricity biofeedback equipment and system based on pattern recognition
CN116019467A (en) * 2023-03-28 2023-04-28 潍坊医学院附属医院 Feature extraction method of facial nerve electromyographic signals
CN117331047A (en) * 2023-12-01 2024-01-02 德心智能科技(常州)有限公司 Human behavior data analysis method and system based on millimeter wave radar

Citations (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102764167A (en) * 2012-06-12 2012-11-07 天津大学 Myoelectric prosthesis control source lead optimization method based on correlation coefficients
CN103577639A (en) * 2013-10-31 2014-02-12 浙江大学 Multi-objective optimization method for synthesis of rotary detection sound field
CN104537382A (en) * 2015-01-12 2015-04-22 杭州电子科技大学 Electromyographic signal gait recognition method for optimizing support vector machine based on genetic algorithm
CN105389442A (en) * 2015-12-09 2016-03-09 天津大学 Reverse design method for coupling genetic algorithm, neural network and numerical simulation
CN105550437A (en) * 2015-12-09 2016-05-04 天津大学 Reverse design method for indoor environment based on genetic algorithm
CN105595990A (en) * 2016-01-27 2016-05-25 浙江铭众科技有限公司 Intelligent terminal device for evaluating and distinguishing quality of electrocardiosignal
KR101711302B1 (en) * 2015-10-26 2017-03-02 한양대학교 산학협력단 Discriminative Weight Training for Dual-Microphone based Voice Activity Detection and Method thereof
CN107255785A (en) * 2017-04-28 2017-10-17 南京邮电大学 Based on the analog-circuit fault diagnosis method for improving mRMR
CN109948465A (en) * 2019-02-28 2019-06-28 浙江工业大学 A kind of surface electromyogram signal classification method based on multiple target
CN110974212A (en) * 2019-11-07 2020-04-10 曲阜师范大学 Electrocardio and myoelectric characteristic fused rehabilitation training motion state monitoring method and system
CN112036357A (en) * 2020-09-09 2020-12-04 曲阜师范大学 Upper limb action recognition method and system based on surface electromyogram signal
CN112971800A (en) * 2021-01-27 2021-06-18 东南大学 Deep learning-based electrocardio premature beat signal processing method, module and device
CN113229831A (en) * 2021-05-10 2021-08-10 燕山大学 Movement function monitoring and management method based on myoelectricity and myooxygen signals
WO2021161901A1 (en) * 2020-02-13 2021-08-19 富士フイルム株式会社 Feature value selection method, feature value selection program, multiclass classification method, multiclass classification program, feature value selection device, multiclass classification device, and feature value set

Patent Citations (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102764167A (en) * 2012-06-12 2012-11-07 天津大学 Myoelectric prosthesis control source lead optimization method based on correlation coefficients
CN103577639A (en) * 2013-10-31 2014-02-12 浙江大学 Multi-objective optimization method for synthesis of rotary detection sound field
CN104537382A (en) * 2015-01-12 2015-04-22 杭州电子科技大学 Electromyographic signal gait recognition method for optimizing support vector machine based on genetic algorithm
KR101711302B1 (en) * 2015-10-26 2017-03-02 한양대학교 산학협력단 Discriminative Weight Training for Dual-Microphone based Voice Activity Detection and Method thereof
CN105389442A (en) * 2015-12-09 2016-03-09 天津大学 Reverse design method for coupling genetic algorithm, neural network and numerical simulation
CN105550437A (en) * 2015-12-09 2016-05-04 天津大学 Reverse design method for indoor environment based on genetic algorithm
CN105595990A (en) * 2016-01-27 2016-05-25 浙江铭众科技有限公司 Intelligent terminal device for evaluating and distinguishing quality of electrocardiosignal
CN107255785A (en) * 2017-04-28 2017-10-17 南京邮电大学 Based on the analog-circuit fault diagnosis method for improving mRMR
CN109948465A (en) * 2019-02-28 2019-06-28 浙江工业大学 A kind of surface electromyogram signal classification method based on multiple target
CN110974212A (en) * 2019-11-07 2020-04-10 曲阜师范大学 Electrocardio and myoelectric characteristic fused rehabilitation training motion state monitoring method and system
WO2021161901A1 (en) * 2020-02-13 2021-08-19 富士フイルム株式会社 Feature value selection method, feature value selection program, multiclass classification method, multiclass classification program, feature value selection device, multiclass classification device, and feature value set
CN112036357A (en) * 2020-09-09 2020-12-04 曲阜师范大学 Upper limb action recognition method and system based on surface electromyogram signal
CN112971800A (en) * 2021-01-27 2021-06-18 东南大学 Deep learning-based electrocardio premature beat signal processing method, module and device
CN113229831A (en) * 2021-05-10 2021-08-10 燕山大学 Movement function monitoring and management method based on myoelectricity and myooxygen signals

Non-Patent Citations (7)

* Cited by examiner, † Cited by third party
Title
DIANCONG ZHANG: "Bimodal Anxiety State Assessment Based on Electromyography and Electroencephalogram", 2020 INTERNATIONAL CONFERENCE ON SENSING, MEASUREMENT & DATA ANALYTICS IN THE ERA OF ARTIFICIAL INTELLIGENCE (ICSMD) *
LANGE, G: "Classification of Electroencephalogram Data from Hand Grasp and Release Movements for BCI Controlled Prosthesis", 3RD INTERNATIONAL CONFERENCE ON SYSTEM-INTEGRATED INTELLIGENCE: NEW CHALLENGES FOR PRODUCT AND PRODUCTION ENGINEERING *
SHAMSIN, M: "Supervised and unsupervised learning in processing myographic patterns", BIGDATA CONFERENCE (FORMERLY INTERNATIONAL CONFERENCE ON BIG DATA AND ITS APPLICATIONS), vol. 1117 *
李奂良: "基于表面肌电信号的手部动作识别与交互", 中国优秀硕士学位论文全文数据库 (基础科学辑), no. 8 *
陈彦钊: "基于sEMG信号的上肢自主康复训练动作模式识别", 中国博士学位论文全文数据库 (医药卫生科技辑), no. 1 *
高清: "基于肌电信号的手势动作识别", no. 06, pages 080 - 6 *
黄虎，谢洪波: "基于蚁群算法的表面肌电信号特征选择", vol. 31, no. 02, pages 164 - 169 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115034273A (en) * 2021-12-27 2022-09-09 驻马店市中心医院 Myoelectricity biofeedback equipment and system based on pattern recognition
CN115034273B (en) * 2021-12-27 2023-09-01 驻马店市中心医院 Myoelectricity biofeedback equipment and system based on pattern recognition
CN116019467A (en) * 2023-03-28 2023-04-28 潍坊医学院附属医院 Feature extraction method of facial nerve electromyographic signals
CN117331047A (en) * 2023-12-01 2024-01-02 德心智能科技(常州)有限公司 Human behavior data analysis method and system based on millimeter wave radar

Also Published As

Publication number Publication date
CN113598759B (en) 2023-09-22

Similar Documents

Publication Publication Date Title
CN113598759B (en) Myoelectricity feature optimization-based lower limb action recognition method and system
US6660042B1 (en) Methods of biomimetic finger control by filtering of distributed forelimib pressures
CN110974212A (en) Electrocardio and myoelectric characteristic fused rehabilitation training motion state monitoring method and system
CN113065526B (en) Electroencephalogram signal classification method based on improved depth residual error grouping convolution network
CN108874149B (en) Method for continuously estimating human body joint angle based on surface electromyogram signal
CN101766509B (en) Real-time control method for artificial limb based on single-point acquiring muscle signals
CN102169690A (en) Voice signal recognition system and method based on surface myoelectric signal
CN102073881A (en) Denoising, feature extraction and pattern recognition method for human body surface electromyography signals
CN113128552A (en) Electroencephalogram emotion recognition method based on depth separable causal graph convolution network
CN112603758A (en) Gesture recognition method based on sEMG and IMU information fusion
CN111643321B (en) Exoskeleton joint angle prediction method and system based on sEMG signals
CN112633195A (en) Myocardial infarction identification and classification method based on frequency domain features and deep learning
CN108958474A (en) A kind of action recognition multi-sensor data fusion method based on Error weight
CN113116361A (en) Sleep staging method based on single-lead electroencephalogram
CN107280659A (en) Electrocardiosignal processing method and system
CN116225222A (en) Brain-computer interaction intention recognition method and system based on lightweight gradient lifting decision tree
CN112036357B (en) Upper limb action recognition method and system based on surface electromyographic signals
CN111950460B (en) Muscle strength self-adaptive stroke patient hand rehabilitation training action recognition method
CN112405539B (en) Robot natural control method based on electromyographic signals and electroencephalogram error potentials
CN114387668A (en) Classification method based on multi-level neuromuscular coupling characteristic information fusion
CN114569143A (en) Myoelectric gesture recognition method based on attention mechanism and multi-feature fusion
Lu et al. Channel-distribution hybrid deep learning for sEMG-based gesture recognition
CN111973388B (en) Hand rehabilitation robot control method based on sEMG
CN111973184B (en) Model training data optimization method for nonideal sEMG signals
Jarrah et al. A comparative analysis on the impact of linear and non-linear filtering techniques on EMG signal quality of transhumeral amputees

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant