CN105997064B - A kind of discrimination method for human body lower limbs surface electromyogram signal - Google Patents
A kind of discrimination method for human body lower limbs surface electromyogram signal Download PDFInfo
- Publication number
- CN105997064B CN105997064B CN201610326948.5A CN201610326948A CN105997064B CN 105997064 B CN105997064 B CN 105997064B CN 201610326948 A CN201610326948 A CN 201610326948A CN 105997064 B CN105997064 B CN 105997064B
- Authority
- CN
- China
- Prior art keywords
- signal
- noise
- filtering
- frequency coefficient
- formula
- 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.)
- Expired - Fee Related
Links
Classifications
-
- 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
Landscapes
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The present invention relates to a kind of discrimination methods for human body lower limbs surface electromyogram signal.The present invention mainly comprises the following steps：Surface EMG signal of the corresponding muscle masses of acquisition in real time under active actions stimulation；The EMG signal of acquisition is pre-processed, the EMG signal after eliminating artefact signal is obtained；The EMG signal of acquisition is decomposed using discrete small wave converting method, low frequency coefficient vector sum high frequency coefficient row are obtained after decomposition；Using when, the filtering method that combines of frequency domain, the Wavelet Component of acquisition is subjected to singular value decomposition, and obtained singular value constitutive characteristic matrix will be decomposed；Feature samples are trained using support vector machines, and generate support vector machine classifier for carrying out Classification and Identification to blind sample.Beneficial effects of the present invention are, relative to traditional technology, the present invention proposes a kind of new lower limb EMG signal preprocess method.
Description
Technical field
The present invention is designed for the surface electromyogram signal (Electromyographic in the bioelectrical signals of human body lower limbs
Signal, EMG) artefact eliminate, feature extraction and identification technique field.
Background technology
Data are shown according to statistics, and China formally marches toward aging society at the beginning of 21 century, and aging process is more than it
He is national, it is contemplated that the year two thousand twenty China elderly population can reach 2.48 hundred million, and the year two thousand fifty is up to 400,000,000.The elderly and extremity disabled persons
It is being expanded rapidly in population structure, the distinguishing feature of above-mentioned crowd is that its daily behavior activity is both needed to provide and assists help.
Paralysis is to lead to one of the most common reason of above-mentioned crowd's loss of athletic ability, and especially the lower part of the body is paralysed, it is related to limbs, body
Dry part or complete function is lost.Clinical expert is it is believed that limb motion rehabilitation is considered as one effective at present
Solution, it is required that impacted lower limb actively assists in positive exercise.However, for limb function, there are obstacles
Crowd, be generally difficult to smoothly complete such as standing, squat down, walk lower limb and move.Therefore, grinding by EMG signal
Study carefully the biofeedback mechanism that can help to explore kinesitherapy nerve and musculature, prediction and perception limb motion situation, assessment are old
The muscle activity ability of year people, disabled person and sub-health population, for the health for developing suitable for the elderly and physically disabled
Multiple lower extremity movement ancillary equipment provides theoretical foundation and application foundation.
Surface EMG signal is a kind of physiology letter of analysis human body and the relevant lower limb athletic performance of number of storage tanks produced per day
Source is ceased, it is non-when which is the neuromuscular system activity recorded from muscle surface by electrode guiding, amplification
Steady One-dimension Time Series bioelectrical signals, it can reflect muscle strength and the motion feature of people.The motion intention of people is usual
Stimulate contraction and the diastole of muscle cell to realize by nervous excitation, due to contraction of muscle in different limb motions pattern not
Together, cause the feature of corresponding surface electromyogram signal that also there is difference, the surface EMG signal energy generated under the intention control of people
Reflected well limb motion or motion characteristic assess human body motion intention.The realization that human motion is intended to mainly passes through limb
Body is completed, and since lower limb EMG signal is increasingly complex relative to upper limb, while by the noise jamming of bigger, therefore is transported to human body
Dynamic research is concentrated mainly on upper limb EMG signal, and is needed to the research of lower limb EMG signal and its identification further perfect.Mesh
Preceding common method is all based on traditional classification or clustering algorithm in the identification of EMG signal, such as support vector machines, nerve
Network (Neural Network Algorithm, referred to as " NNA "), linear discriminant analysis (Linear Discriminant
Analysis, referred to as " LDA ") etc..Wherein, LDA algorithm can identify certain single action, can also add multiple actions
Upper label is identified as special one kind.In the research of lower limb EMG signal, 2016, John A.Spanias et al.
Using LDA algorithm, the other types only classified with EMG signal and return to EMG signal and instrument sensor are had studied
The method classified together of data；2014, AJ Young et al. used the method pair of Sensor Time History
EMG signal is classified, but this method only accounts for the time span of signal in whole process.
Invention content
The purpose of the present invention provides a kind of identification for human body lower limbs surface electromyogram signal aiming at the above problem
Method.
By experimental analysis, it is found that lower limb EMG signal scale is usually very faint, simultaneously because hardware limitations and limbs
The reasons such as mobile, it is also very easy to be interfered by power frequency, baseline drift and white Gaussian noise, therefore test the collected lower limb of institute
EMG signal artefact is serious, and it is unpractical that feature extraction and Classification and Identification are directly carried out based on initial data.Traditional artefact is eliminated
Method is that (the main component integrated distribution of lower limb EMG signal is in 20~500Hz frequency ranges to original signal progress bandpass filtering treatment
On).Notch filter and low-pass filtering are carried out to Hz noise, baseline drift respectively in the present invention, according to work proposed by the present invention
The frequency interference noise factor and baseline drift noise factor respectively assess lower limb EMG signal, when the energy of noise exceeds threshold
When value, using trap and low-pass filtering, otherwise using filtering front signal.And for white Gaussian noise, lower limb EMG letters are calculated first
Number zero passage points, by zero passage points to lower limb EMG time-domain signals carry out interval division, then by the white Gaussian noise factor by
The secondary noise to each section is assessed, when the energy of noise exceeds threshold value, using filter result, before otherwise using filtering
Signal.Under the premise of on surface, EMG signal has non-stationary, non-linear behavior, the present invention is by wavelet transform
(Discrete Wavelet Transform, DWT) is applied in the feature extraction of lower limb EMG signal, it is by time domain, frequency domain
Analysis is combined the information for being included to the time of surface EMG signal and frequency and analyzes.And traditional artefact is eliminated (such as band
Bandpass filter), feature extraction (such as Fourier transform, time and frequency domain analysis) method is only independently in time domain or frequency domain
Data are analyzed, and EMG signal is considered as steady or short-term stationarity signal and is handled, therefore traditional method can not be accurate
Ground obtains the EMG physiological reaction features of human body lower limbs limb motion action.Singular value decomposition (Singular Value
Decomposition, SVD) it is a kind of effective algebraic characteristic extracting method, since singular value features are describing signal numerically
It is more stable, and with the critical natures such as transposition invariance, invariable rotary shape, shift invariant, therefore singular value features can be with
A kind of effective algebraic characteristic description as signal.Finally, the present invention is by wavelet transform (DWT) and singular value decomposition
(SVD) be combined, based on new lower limb EMG signal preprocess method, it is further proposed that when, the filtering method that combines of frequency domain, root
According to eigenmatrix obtained above, using support vector machines (Support Vector Machine, SVM) method to signal into
Row Classification and Identification.
The technical scheme is that：A kind of discrimination method for human body lower limbs surface electromyogram signal, which is characterized in that
Include the following steps：
A. disposable electromyographic electrode is glued and is put to lower hind limb musculature epidermis, the corresponding muscle masses of acquisition in real time are moved in activity
Surface EMG signal under stimulating；
B. the EMG signal acquired in step a is pre-processed, obtains the EMG signal after eliminating artefact signal；It is described pre-
Processing method includes Hz noise filtering, baseline drift filtering and white Gaussian noise filtering；In this step, Hz noise filtering,
The sequence of baseline drift filtering and white Gaussian noise filtering can carry out arbitrary arrangement；
C. the step b EMG signals obtained are decomposed using discrete small wave converting method, low frequency coefficient is obtained after decomposition
Vectorial cA1 and high frequency coefficient vector cD1；Low frequency coefficient vector cA1 is decomposed using the method for wavelet transform, is obtained
Low frequency coefficient vector cA2 and high frequency coefficient cD2；The method for continuing to reuse wavelet transform carries out low frequency coefficient vector
It decomposes, until until obtaining low frequency coefficient vector cA5 and 5 high frequency coefficient row cD1, cD2, cD3, cD4, cD5；
D. use when, frequency domain combine filtering method, by the Wavelet Component obtained in step c carry out singular value decomposition, and
The singular value constitutive characteristic matrix that decomposition is obtained；
E. using the eigenmatrix obtained in step d as sample, feature samples are trained using support vector machines, and raw
It is used to carry out Classification and Identification to blind sample at support vector machine classifier.
2, a kind of discrimination method for human body lower limbs surface electromyogram signal according to claim 1, feature exist
In pretreated specific method described in step b includes：
B1. Hz noise filters；Specific method is：The EMG signal c (t) collected in step a is used as and is originally inputted
Signal CPLI(t) notch filter is carried out, it is a (t) to obtain filter result, defines Hz noise factor εPLI, then εPLIIt can be by as follows
Formula 1 calculates：
Wherein, var is signal variance operator, and the variance for calculating time series passes through Hz noise factor εPLITo filter
Wave result is modified shown in following formula 2：
Wherein, sPLI(t) it is the final result for filtering out Hz noise noise, formula 2 shows the energy if Hz noise noise
Accounting is more than original energy 10% is measured, then uses the filter result of notch filter；
B2. baseline drift filters；Specific method is：The signal s that step b1 is obtainedPLI(t) it is used as original input signal cBW
(t) low-pass filtering is carried out, filter result is d (t), and the definition baseline drift factor is εBW, then εBWIt can be calculated by following formula 3：
According to baseline drift factor εBWThe baseline drift noise d (t) obtained to filtering is modified to obtain b (t), expresses
Formula is as shown in formula 4：
Finally, the signal s after removal baseline drift noise is obtainedBW(t) shown in following formula 5：
sBW(t)=cBW(t)-b (t) (formula 5)；
B3. white Gaussian noise filters；Specific method is：The signal s that step b2 is obtainedBW(t) it is used as original input signal
cWGN(t) it is filtered；To time-domain signal cWGN(t) it is based on zero crossing and carries out interval division, then by thresholding method to letter
Number cWGN(t) white Gaussian noise filtering is carried out；Specially：Assuming that including j zero crossing z in signali(i=1,2 ..., j), then
For section zi< tj< zi+1Interior signal has：
Wherein, T is the threshold value of signal c (t), tmFor the extreme point in section；Threshold value T is true by following formula 7 and formula 8
It is fixed：
σ=median (s |cwGN(t)|:T=1,2 ..., L)/0.6745 (formula 8)
Herein, σ is input signal cWGN(t) noise level；L is signal cWGN(t) length, in particular for discrete letter
Number L is discrete point points；Median is signal median operator, the median for obtaining time series.
Beneficial effects of the present invention are, relative to traditional technology, the present invention proposes a kind of new lower limb EMG signal and locates in advance
Reason method, while based on the lower limb EMG signal feature after artefact elimination, proposing to combine DWT and SVD first, lower limb typical case is transported
The EMG signal of dynamic action such as walk support, walking walking, squat down and progress feature extraction of standing, finally by SVM to being carried
The eigenmatrix gone out carries out classification identification.
Description of the drawings
Fig. 1 is the flow chart of the present invention；
Fig. 2 is the logic diagram for carrying out 5 layers of wavelet decomposition process to signal using wavelet function.
Specific implementation mode
Below in conjunction with the accompanying drawings, detailed description of the present invention technical solution：
As shown in Figure 1, the present invention is used for the discrimination method of human body lower limbs surface electromyogram signal, mainly include the following steps that：
Step 1) lower limb movement raw EMG signal obtains：
Disposable electromyographic electrode is glued and is put to lower hind limb musculature epidermis, the corresponding muscle masses of acquisition in real time are in active actions
Surface EMG signal c (t) under stimulation；
Step 2) original signal pre-processes (artefact removing method)：
Artefact elimination is carried out to the raw EMG signal that collects comprising to Hz noise noise, white Gaussian noise,
Baseline drift noise filters out.The no fixed sequencing of filtering of three kinds of noises in this step presses Hz noise here
The filtering of noise, the filtering of white Gaussian noise, baseline drift filtering this be sequentially described；Secondly, upper herein in narration
In " denoising " it is identical as " filtering to noise " meaning.
A) denoising of Hz noise：The surface EMG signal that will be collected, c (t), as the original of this denoising process
Input signal, cPLI(t), notch filter is carried out, obtains filter result, a (t), while defining the Hz noise factor, εPLI, such as public
Shown in formula (1)：
Wherein, var is signal variance operator, the variance for calculating time series.Pass through Hz noise factor εPLITo filter
Wave result is modified as follows：
Wherein, sPLI(t) it is the final result for filtering out Hz noise noise.Formula (2) shows if Hz noise noise
Energy accounting is more than the 10% of original energy, then uses the filter result of notch filter.
B) denoising of baseline drift：It will carry out the signal obtained after Hz noise filtering, sPLI(t), as this denoising
The original input signal of journey, cBW(t), low-pass filtering is carried out, filter result is d (t).Similarly, the baseline drift factor is defined,
εBW, as shown in formula (3)：
According to the baseline drift factor, εBW, to the baseline drift noise that filtering obtains, d (t) is modified to obtain, b (t),
Shown in its expression formula such as formula (4)：
Finally, the signal s after removal baseline drift noise is obtainedBW(t) as follows：
sBW(t)=cBW(t)-b(t) (5)
C) denoising of white Gaussian noise：It will carry out the signal s obtained after baseline drift noise filteringBW(t) it is gone as this
The original input signal c for process of making an uproarWGN(t) it is filtered.To time-domain signal cWGN(t) it is based on zero crossing and carries out interval division, then
By thresholding method to signal cWGN(t) white Gaussian noise filtering is carried out.It is assumed that including j zero crossing z in signali(i=
1,2 ..., j), then for section zi< tj< zi+1Interior signal has：
Wherein, T is the threshold value of signal c (t), tmFor the extreme point in section.Threshold value T is determined by formula (7) and (8)：
σ=median (s |cWGN(t)|:T=1,2 ..., L)/0.6745 (8)
Herein, σ is input signal cWGN(t) noise level；L is signal cWGN(t) length, in particular for discrete letter
Number L is discrete point points；Median is signal median operator, the median for obtaining time series.
Step 3) wavelet transform (DWT)：
Wavelet transformation is that time-domain analysis and frequency-domain analysis are combined to a kind of new analysis method to be formed, and reflection is table
The variation that facial muscle electric signal is presented in two dimensions of time and frequency, therefore this method is for both the above method
Should theoretically have certain advantage, can make full use of the information that surface electromyogram signal is included.Wavelet analysis is a kind of window
The Time-Frequency Localization signal analysis method that the size of mouth is fixed, shape is variable has higher frequency discrimination in low frequency part
Rate and lower temporal resolution have higher temporal resolution and lower frequency resolution in high frequency section.
The signal S that a given length is N, discrete wavelet transformation (DWT) at most can be signal decomposition at log2N number of frequency
Rate grade.First step decomposition starts from signal S, and decomposition coefficient consists of two parts after decomposition：Low frequency coefficient vector cA1 and high frequency system
Number vector cD1, also referred to as approximation (Approximation) ingredient, the latter are also referred to as details (Detail) ingredient for the former.To
Amount cA1 is obtained with low pass resolution filter by convolution algorithm by signal S, and vectorial cD1 is to be decomposed to filter by signal S and high pass
What wave device was obtained by convolution algorithm.In next step is decomposed, low frequency coefficient cA1 is divided into two parts with same method, i.e.,
Signal S above is replaced with cA1, the low frequency coefficient cA2 and high frequency coefficient cD2 of return scale 2 after decomposition, while scale 1
High frequency coefficient cD1 is remained unchanged, and so on continues to decompose.Finally obtain a low frequency coefficient row cA5 and five high frequency coefficients
CD1, cD2, cD3, cD4, cD5 are arranged, as shown in Figure 2.
The component that decomposition based on the EMG signal to lower extremity movement obtains, we are carrying out bandpass filtering treatment to it, are making
CD1, cD2, cD3, the frequency range of cD4, cA5 be respectively 256-512Hz, 128-256Hz, 64-128Hz, 32-64Hz and
16-32Hz。
Step 4) singular value decomposition (SVD)：
According to the method in step 3, the EMG signal of lower extremity movement can be decomposed by DWT methods, from step 3
In obtained a low frequency coefficient row cA5 and five high frequency coefficient row cD1, cD2, cD3, cD4, cD5, chooses 5 data and carry out structure
At matrix expression, such as select：CD2, cD3, cD4, cD5 and cA5 constitute matrix, and the expression matrix for obtaining 5 × I is as follows, Mij=
[d2i；d3i；d4i；d5i；a5i], i=1,2 ..., I, wherein I be lower extremity movement sample action number, d2i, d3i, d4i, d5i and
A5i corresponds to the coefficient that DWT decomposition obtains will use SVD here in order to obtain the characteristic value that can be used in svm classifier
To MijDecompose and obtains singular value.
For M, N ∈ Cm×n, make equation U if there is m rank unitary matrice U and n rank unitary matrice VTMV=N is set up, then it is assumed that
Matrix M and N Unitary Equivalent.IfThen matrix MTThe characteristic value of M should meet following relationship：
λ1≥λ2≥…≥λr≥λr+1=...=λn=0 (9)
Then it can obtain, the singular value of matrix M is
IfThen certainly exist m rank unitary matrice U and n rank unitary matrice V so that matrix M, matrix U and
Matrix V meets formula (10)：
Σ=diag (σ1, σ2... σr), wherein σi(i=1,2 ..., r are the non-zero singular values of matrix M；Formula (10)
It is equivalent to formula (11)：
Here,
U=MMT (12)
V=MTM (13)
Wherein, formula (11) is also referred to as the singular value decomposition (SVD) of matrix M.
Step 5) svm classifier：
In order to classify to lower extremity movement, these matrixes M for being decomposed by DWFijSingular value, be used as lower limb
The time and frequency domain characteristics matrix of EMG signal.For chosen data segment, matrix MijIt is determined by following formula：
Wherein, i is the singular value number of the data of each experiment (Trial), and j is EMG number of samples.
Vapnik et al. creates support vector machines (SVM) method.The basic thought of SVM is to map the data into higher-dimension sky
Between in, and find out the hyperplane with maximum decision edge.The principle of minimization risk that SVM formula are characterized has been demonstrated to be better than
Other traditional empirical principles of minimization risk.In order to obtain the good real-time grading device of a robustness, in the present invention,
It will be obtained in terms of time domain or frequency domain two for the characteristic value needed for svm classifier, or the analysis method being combined using time-frequency domain.
Assuming that there are the training vector x of l sampleiWith corresponding tag along sort yi, (x1, y1) ..., (x1, y1)∈RD
× { -1,1 }, then SVM is substantially to solve for quadratic programming problem：
yi(w·xi+b)≥1-ζi, ζi> 0 (i=1,2 ..., 1) (15)
Wherein, C is constant, ζiIt is slack variable, w is weight vectors, and b is deviation.For example x, discriminant function
As follows：
Wherein, NsIt is the number of supporting vector, αiIt is positive Lagrange multiplier, G (x, xi) it is kernel function.It is most heavy in SVM
The parameter of regularity wanted also is determined by ten times of cross-validation process of training set.
This patent uses linear kernel function K processing feature matrixes Mij, the classification function f (x) of support vector machines can describe
For：
N is number of samples, xjIt is j-th of sample, yjThe SVM outputs of j-th sample, K be converted for data it is linear
Kernel function, αjIt is the Lagrange multiplier of primal-dual optimization problem.
Sample data is then intersected by ten times first by normalized (zero-mean+unit norm deviation processing)
Verification technique assessment classification identification effect.The present invention can be such that the recognition accuracy of human body lower limb athletic performance obtains effectively
It improves.
Claims (1)
1. a kind of discrimination method for human body lower limbs surface electromyogram signal, which is characterized in that include the following steps：
A. disposable electromyographic electrode is glued and is put to lower hind limb musculature epidermis, the corresponding muscle masses of acquisition in real time are pierced in active actions
Surface EMG signal under swashing；
B. the EMG signal acquired in step a is pre-processed, obtains the EMG signal after eliminating artefact signal；The pretreatment
Method includes Hz noise filtering, baseline drift filtering and white Gaussian noise filtering；Specific method includes：
B1. Hz noise filters；Specific method is：It regard the EMG signal c (t) collected in step a as original input signal
CPLI(t) notch filter is carried out, it is a (t) to obtain filter result, defines Hz noise factor εPLI, then εPLIFollowing formula can be passed through
1 calculates：
Wherein, var is signal variance operator, and the variance for calculating time series passes through Hz noise factor εPLIFiltering is tied
Fruit is modified shown in following formula 2：
Wherein, sPLI(t) it is the final result for filtering out Hz noise noise, formula 2 shows if the energy of Hz noise noise accounts for
Than 10% more than original energy, then the filter result of notch filter is used；
B2. baseline drift filters；Specific method is：The signal s that step b1 is obtainedPLI(t) it is used as original input signal cBW(t)
Low-pass filtering is carried out, filter result is d (t), and the definition baseline drift factor is εBW, then εBWIt can be calculated by following formula 3：
According to baseline drift factor εBWThe baseline drift noise d (t) obtained to filtering is modified to obtain b (t), and expression formula is such as
Shown in formula 4：
Finally, the signal s after removal baseline drift noise is obtainedBW(t) shown in following formula 5：
sBW(t)=cBW(t)-b (t) (formula 5)；
B3. white Gaussian noise filters；Specific method is：The signal s that step b2 is obtainedBW(t) it is used as original input signal cWGN
(t) it is filtered；To time-domain signal cWGN(t) it is based on zero crossing and carries out interval division, then by thresholding method to signal
cWGN(t) white Gaussian noise filtering is carried out；Specially：Assuming that including j zero crossing z in signali(i=1,2 ..., j), then it is right
In section zi< tj< zi+1Interior signal has：
Wherein, T is the threshold value of signal c (t), tmFor the extreme point in section；Threshold value T is determined by following formula 7 and formula 8：
σ=median (s |cWGN(t)|：T=1,2 ..., L)/0.6745 (formula 8)
Herein, σ is input signal cWGN(t) noise level；L is signal cWGN(t) length, in particular for discrete signal L
As discrete point is counted；Median is signal median operator, the median for obtaining time series；
C. the step b EMG signals obtained are decomposed using discrete small wave converting method, low frequency coefficient vector is obtained after decomposition
CA1 and high frequency coefficient vector cD1；Low frequency coefficient vector cA1 is decomposed using the method for wavelet transform, obtains low frequency
Coefficient vector cA2 and high frequency coefficient cD2；The method for continuing to reuse wavelet transform divides low frequency coefficient vector
Solution, until until obtaining low frequency coefficient vector cA5 and 5 high frequency coefficient row cD1, cD2, cD3, cD4, cD5；
D. use when, frequency domain combine filtering method, by the Wavelet Component obtained in step c carry out singular value decomposition, and will point
The singular value constitutive characteristic matrix that solution obtains；
E. using the eigenmatrix obtained in step d as sample, feature samples are trained using support vector machines, and generate branch
Vector machine classifier is held for carrying out Classification and Identification to blind sample.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610326948.5A CN105997064B (en) | 2016-05-17 | 2016-05-17 | A kind of discrimination method for human body lower limbs surface electromyogram signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610326948.5A CN105997064B (en) | 2016-05-17 | 2016-05-17 | A kind of discrimination method for human body lower limbs surface electromyogram signal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105997064A CN105997064A (en) | 2016-10-12 |
| CN105997064B true CN105997064B (en) | 2018-10-23 |
Family
ID=57098243
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610326948.5A Expired - Fee Related CN105997064B (en) | 2016-05-17 | 2016-05-17 | A kind of discrimination method for human body lower limbs surface electromyogram signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105997064B (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107126303A (en) * | 2017-02-15 | 2017-09-05 | 上海术理智能科技有限公司 | A kind of upper and lower extremities exercising support method based on mobile phone A PP |
| CN107440716A (en) * | 2017-07-26 | 2017-12-08 | 电子科技大学 | Human body lower limbs athletic performance classification discrimination method based on single channel electromyographic signal |
| CN108143414B (en) * | 2018-01-11 | 2021-05-18 | 福州大学 | Online method for adaptively filtering surface electromyographic signal electrical stimulation artifact |
| CN108196677A (en) * | 2018-01-15 | 2018-06-22 | 上海交通大学 | A kind of gesture identification wrist strap |
| CN108890621A (en) * | 2018-07-10 | 2018-11-27 | 安春亭 | A kind of wearable smart machine control method that auxiliary is carried |
| CN109062401B (en) * | 2018-07-11 | 2020-12-15 | 北京理工大学 | Real-time gesture recognition system based on electromyographic signals |
| CN109344788A (en) * | 2018-10-15 | 2019-02-15 | 徐妍 | A kind of wearable smart machine control method that auxiliary is carried |
| CN109446957B (en) * | 2018-10-18 | 2021-01-01 | 广州云从人工智能技术有限公司 | EMG signal identification method |
| CA3124751A1 (en) * | 2018-12-26 | 2020-07-02 | Analytics For Life Inc. | Method and system for automated quantification of signal quality |
| CN109567799A (en) * | 2018-12-26 | 2019-04-05 | 杭州电子科技大学 | EMG Feature Extraction based on smooth small echo coherence |
| CN109567798A (en) * | 2018-12-26 | 2019-04-05 | 杭州电子科技大学 | Daily behavior recognition methods based on myoelectricity small echo coherence and support vector machines |
| CN110151177A (en) * | 2019-05-28 | 2019-08-23 | 长春理工大学 | Drop foot detection device and detection method based on surface electromyogram signal |
| CN112037812B (en) * | 2020-09-01 | 2021-06-15 | 深圳爱卓软科技有限公司 | Audio processing method |
| CN112733721B (en) * | 2021-01-12 | 2022-03-15 | 浙江工业大学 | Surface electromyographic signal classification method based on capsule network |
| CN116981401A (en) * | 2021-03-19 | 2023-10-31 | 深圳市韶音科技有限公司 | Motion monitoring method and system |
| CN116593831B (en) * | 2023-07-19 | 2023-11-07 | 西安交通大学 | Cable defect positioning method, device and medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN2439840Y (en) * | 1999-10-15 | 2001-07-25 | 北京卡迪欧医疗设备有限责任公司 | CDI detector for patient in high danger of sudden death from heart disease |
| WO2004066832A3 (en) * | 2003-01-31 | 2004-12-29 | Enel Produzione Spa | System and method of processing electromyographic signals for the diagnosis of parkinson’s disease |
| CN101859377A (en) * | 2010-06-08 | 2010-10-13 | 杭州电子科技大学 | Electromyographic signal classification method based on multi-kernel support vector machine |
| CN102073881A (en) * | 2011-01-17 | 2011-05-25 | 武汉理工大学 | Denoising, feature extraction and pattern recognition method for human body surface electromyography signals |
| CN102169690A (en) * | 2011-04-08 | 2011-08-31 | 哈尔滨理工大学 | Voice signal recognition system and method based on surface myoelectric signal |
| CN102426651A (en) * | 2011-08-25 | 2012-04-25 | 武汉理工大学 | Human body forearm surface electromyogram signal acquisition and pattern recognition system |
| CN104935292A (en) * | 2014-03-17 | 2015-09-23 | 西南科技大学 | Source number estimation-based surface electromyogram signal adaptive filtering method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080015452A1 (en) * | 2006-06-30 | 2008-01-17 | Ricci Carlos A | Method of processing electrocardiogram waveform |
-
2016
- 2016-05-17 CN CN201610326948.5A patent/CN105997064B/en not_active Expired - Fee Related
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN2439840Y (en) * | 1999-10-15 | 2001-07-25 | 北京卡迪欧医疗设备有限责任公司 | CDI detector for patient in high danger of sudden death from heart disease |
| WO2004066832A3 (en) * | 2003-01-31 | 2004-12-29 | Enel Produzione Spa | System and method of processing electromyographic signals for the diagnosis of parkinson’s disease |
| CN101859377A (en) * | 2010-06-08 | 2010-10-13 | 杭州电子科技大学 | Electromyographic signal classification method based on multi-kernel support vector machine |
| CN102073881A (en) * | 2011-01-17 | 2011-05-25 | 武汉理工大学 | Denoising, feature extraction and pattern recognition method for human body surface electromyography signals |
| CN102169690A (en) * | 2011-04-08 | 2011-08-31 | 哈尔滨理工大学 | Voice signal recognition system and method based on surface myoelectric signal |
| CN102426651A (en) * | 2011-08-25 | 2012-04-25 | 武汉理工大学 | Human body forearm surface electromyogram signal acquisition and pattern recognition system |
| CN104935292A (en) * | 2014-03-17 | 2015-09-23 | 西南科技大学 | Source number estimation-based surface electromyogram signal adaptive filtering method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105997064A (en) | 2016-10-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105997064B (en) | A kind of discrimination method for human body lower limbs surface electromyogram signal | |
| Abbaspour et al. | Evaluation of surface EMG-based recognition algorithms for decoding hand movements | |
| CN105956624B (en) | Mental imagery brain electricity classification method based on empty time-frequency optimization feature rarefaction representation | |
| CN110238863A (en) | Based on brain electricity-electromyography signal lower limb rehabilitation robot control method and system | |
| Abdullah et al. | Surface EMG signal classification by using WPD and ensemble tree classifiers | |
| CN107945817A (en) | Heart and lung sounds signal sorting technique, detection method, device, medium and computer equipment | |
| Wołczowski et al. | Electromyography and mechanomyography signal recognition: Experimental analysis using multi-way array decomposition methods | |
| AlOmari et al. | Analysis of extracted forearm sEMG signal using LDA, QDA, K-NN classification algorithms | |
| CN110013248A (en) | Brain electricity tensor mode identification technology and brain-machine interaction rehabilitation system | |
| CN108681685A (en) | A kind of body work intension recognizing method based on human body surface myoelectric signal | |
| CN108268844A (en) | Movement recognition method and device based on surface electromyogram signal | |
| Aly et al. | Bio-signal based motion control system using deep learning models: A deep learning approach for motion classification using EEG and EMG signal fusion | |
| CN115640827B (en) | Intelligent closed-loop feedback network method and system for processing electrical stimulation data | |
| Pancholi et al. | Intelligent upper-limb prosthetic control (iULP) with novel feature extraction method for pattern recognition using EMG | |
| Narayan | Direct comparison of SVM and LR classifier for SEMG signal classification using TFD features | |
| Radman et al. | Generalized sequential forward selection method for channel selection in EEG signals for classification of left or right hand movement in BCI | |
| Wang et al. | Motor imagination eeg recognition algorithm based on dwt, CSP and extreme learning machine | |
| CN113128384A (en) | Brain-computer interface software key technical method of stroke rehabilitation system based on deep learning | |
| Abibullaev et al. | Epileptic spike detection using continuous wavelet transforms and artificial neural networks | |
| Yu et al. | The research of sEMG movement pattern classification based on multiple fused wavelet function | |
| Hussain et al. | Amputee walking mode recognition based on mel frequency cepstral coefficients using surface electromyography sensor | |
| Ji et al. | Stationary wavelet-based two-directional two-dimensional principal component analysis for EMG signal classification | |
| KR102630840B1 (en) | EMG signal-based recognition information extraction system and EMG signal-based recognition information extraction method using the same | |
| Vijayvargiya et al. | Implementation of machine learning algorithms for automated human gait activity recognition using sEMG signals | |
| Kocer | Classifying myopathy and neuropathy neuromuscular diseases using artificial neural networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20181023 Termination date: 20210517 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |