CC BY
8STUDY OF NOISE AND INTERFERENCE OF SURFACE ELECTROMYOGRAPHY SIGNAL _AND WAVELET DENOISING
Vladimir Bonilla
Postgraduate student, Litvin Anatoliy V. Candidate of Science, assistant professor, Lukyanov Evgeniy A Candidate of Science, assistant professor, Deplov Dmitriy A.
Master student, Don state technical university, г. Ростов-на-Дону.
ИССЛЕДОВАНИЕ ШУМОВ И ПОМЕХ ПОВЕРХНОСТНОГО ЭЛЕКТРОМИОГРАФИЧЕСКОГО СИГНАЛА С ЦЕЛЬЮ ИХ УДАЛЕНИЯ ВЕЙВЛЕТАМИ
Бонилья Владимир. Ф., Аспирант
Литвин Анатолий Витальевич, кандидат техн. наук, доцент. Лукьянов Евгений Анатольевич, кандидат техн. наук, доцент.
Деплов Дмитрий Алексеевич. магистрант, Донской государственный технический университет, г. Ростов-на-Дону
ABSTRACT
The aim of this study was investigate noises and interferences which disturb the surface electromyography signal (sEMG). It was shown that the noises and interferences are caused by various sources. Sources of interference and noise can be divided into internal and external. The internal noise are caused by the electrodes, EMG signals of other muscles; noise associated with the functioning of other organs such as the heart or stomach. The external noses are due to electrical environment the most prominent of which is the direct interference of the power hum, produced by the incorrect grounding of other devices and electro motors. The block diagram of the noise sources was developed and with accordance with the diagram EMG signal was simulated. Denosing of simulated EMG signal was fulfilled by different wavelets and compare with digital filtering. The smallest error was observed in the case when using wavelet db4 of level 6.
Keywords—Surface EMG, noise and interference, 50 Hz power line interference, wavelet denoising, digital filtering. АННОТАЦИЯ
Исследовались шумы и помехи поверхностных элекромиографических сигналов (пЭМГ). Было установлено, что шумы и помехи вызываются различные источники, которые подразделяются на внутренние и внешние. Внутренние источники вызываются поляризацией электродов, ЭМГ сигналами скелетных мышц, шумами связанными с работой других органов таких как желудок, сердце. Внешние помехи вызываются окружающей электромагнитной средой. К вешним помехам и шумам относятся сетевая помеха, помехи от работающего электрооборудования. Для исследования методов удаления помех и шумов были созданы модели ЭМГ сигналов. Удаление шумов и помех выполнялось с помощью вейвлет преобразования. Наименьшая погрешность наблюдалась при использовании вейвлета db4 6 -того уровня.
Ключевые слова - Поверхностная ЭМГ, шумы и помехи, удаление помех вейвлетами
Introduction. The skeletal muscle is a biological tissue able to transform chemical energy to mechanical energy. The smallest functional unit describing the neural control of the muscular contraction process is called a Motor Unit.
In biomedical signals studies the motor unit action potentials (MUAPs) of all active motor units under the electrodes are observed as a bipolar signal with a symmetric distribution of positive and negative amplitudes (mean value equals to zero) forming an interference pattern [1]. The signal obtained from surface electrodes is called a surface electromyography signal (sEMG) and is often used to evaluate the relative level of muscle activity during movement.
An unfiltered and unprocessed signal consistent with the superposed MUAPs is called a raw sEMG signal. The raw sEMG signal is a complex signal having interfering components and therefore its analysis is rather difficult [2].
sEMG is established as an evaluation tool for applied research, physiotherapy, rehabilitation, sports training and interactions of the human body with industrial products and work conditions. It has been proven that amputees who have lost their hand are able to generate signals in the forearm
muscles that are similar to those generated by healthy subjects. EMG signal is widely used to control various electromechanical devices such as prostheses [3], exoskeletons [4] and other mechatronic systems. EMG signals obtained from implanted electrodes from motor fibers of a peripheral nerve are used to control bioelectric prostheses but for controlling other mechatronic devices parameters of sEMG are more suitable. [5]. Thus, sEMG signals can prove extremely useful in restoring some of the lost motor functionalities in amputees.
Bioelectric control of mechatronic devices is carried out in conditions of high interference, for example, industrial electromagnetic noises, which requires the development of methods for the removal of a variety of noises from the raw sEMG. To remove noise from the raw sEMG signals band-pass and lattice filters are usually used, however, such filtering distorts the signal, and removes useful information from the sEMG signal [6]. Analysis of known scientific works [7,8] shows that the most efficient method for removal of artifacts from biomedical signals is the wavelet decomposition of the
signal, followed by the correction of the decay and the recovery of the signal by removing noises and interferences.
The choice of the mother wavelet (continuous or discrete) is not a trivial task. Analysis of published scientific research shows that for the analysis and filtering of biomedical signals, different mother wavelets are applied such as the derivatives of the Gaussian function for the continuous wavelet analysis of EMG signal, a family of the orthogonal wavelets of Daubechies [7] and etc.
According to the sampling theorem, the continuous signal S(t) whose spectrum contains no frequencies higher than fm, is completely determined by a discrete sequence of their instantaneous values {Si}, i = 0,1,..., N -1, measured at time intervals of At:
At =1/ 2 fm, fg =1/ At, (1)
where At and fg - time interval, c. and the sampling frequency, Hz. Thus, the signal sampled at an interval of At can be determined by the following expression
S-1 (2)
Sd(t) = {Si}=^S(iAt)S(t-iAt),
where 5 (t) the Dirac delta function
It is known that any signal S(t) for which the condition/tt2[S(t)]2dt < ro is true, may be represented by an orthogonal system of functions ^(t):
S(t) =Co^oOO + - + Cn^n(t) (3)
= 1
СпФпОО
The wavelet coefficients Cn can be calculated using an iterative procedure known as the fast wavelet transform FWT [7].
The fast wavelet transform is a mathematical algorithm designed to turn a waveform or signal in the time domain into a sequence of coefficients based on an orthogonal basis of small finite waves, or wavelets. The transform can be easily extended to multidimensional signals, such as images, where the time domain is replaced by the space domain.
When the order n =1 the WAVE-wavelet having the raw moment equal to zero is received. When n =2 we get MHAT-wavelet, called "Mexican hat, it provides better resolution than the WAVE wavelet. The Haar wavelet is also the simplest possible wavelet. The technical disadvantage of the Haar wavelet is that it is not continuous, and therefore
not differentiable. The drawbacks of Haar wavelet are asymmetrical form and ripples that results in endless alternations of the "petals" decreasing in the frequency domain or spectral leakage being proportional to 1/w.
Among the complex wavelet transform the Morlet wavelet is most commonly used. This wavelet is well localized both in time and in frequency domain. The characteristic parameter W0 allows changing the selectivity of the basis.
Most of the wavelet functions have no analytical description in a single formula, and are iterative expressions and for this reason are easily calculated by computers. Functions of Daubechies are an example of such wavelets
Noise and interference. When registering biomedical signals with the main signal the noise and interference of various natures are simultaneously recorded. The interference and noise also include distortion of the useful signal by various destabilizing factors while being measured, such as the effects of lightning discharges, interference from operating industrial equipment, etc. Sources of interference and noise are divided into internal and external. The internal noise or physiological noise are the noise of the electrodes, EMG noise other muscles; noise associated with the functioning of other organs such as the heart or stomach. Special care must be taken in very noisy electrical environment, categorized as external noise, the most prominent of which is the direct interference of the power hum, typically produced by the incorrect grounding of other external devices. The block diagram of the noise sources are shown in Fig. 1
The aim of the study is to identify interference and noise present in the sEMG signal and to fulfill comparative analysis of applying wavelet transformations for the removal of artifacts and various interferences from a raw sEMG signal.
Experimental Procedures. At the first stage of research the interference and noise were recorded and analyzed in accordance with the diagram shown in fig.1. The registration channel of sEMG signal interference and noise includes surface electrodes, bio amplifier, band pass filter with cutoff frequencies of 10 and 250 Hz and ADC. The Ag/AgCl electrodes, with a diameter of 5 mm, were attached to the surface of the right M. Biceps Brachii along the line of muscle fibers, at a distance of 2 cm from each other. The noise was amplified and passed through Matlab using the data acquisition system NI USB-6212. The sampling rate of the noise signals was set at 1 kHz and a 16-bit ADC was used. During the recording of the noise and interference, right hand with fixed electrodes on the biceps remained motionless.
Hamp(t) inA/DOO
Fig. 1. Block diagram showing the principal noise sources in electromyography.
Fig. 2 shows the total noise and interference induced by the 0.55 kW three-phase asynchronous electromotor at a distance of 0.5 - 1 m, EMG of skeletal muscles due to rotations of the trunk and the ECG signal.
After determining the noise and interference, the EMG signals models were created. Base EMG signal was taken from
the Physionet Database Examples of Electromyograms Record (emgdb). This signal was used as the original signal, on the basis of which the synthesis of EMG signals with varying degrees of noise was carried out to assess the quality of different filtering methods and DWT denoising (fig. 3).
s >
e, в «
1000 3000 2000 1000 0
0.2 В 0.1
Ё
<
-0.2
0.1
ff 0.3
S -- ll.ï
S- Ê
E 0.1
о £ 0
time, sec
50
100
150
2D О
250
fiecueitty. Hi b)
20
frecuency, Hi d)
30
40
¿te
го зо
frequency, Hz I)
Fig. 2. Noise induced on the electrodes installed on the biceps of the right hand: a) the total interference from the AC electromotor; EMG skeletal muscles during rotations of the trunk around its axis ECG; b) the spectral power and the total interference; c) noise induced EMG of the skeletal muscles during rotations of the trunk and ECG; d) the power spectrum of noise consisting of EMG and ECG; e) and f) - hindrance caused by the ECG, and its power spectrum
When noise is added to an EMG signal, it can be represented by the summation of the original signal (sEMG ) and the total noise ( n^ ), as shown in equation (1).
Discrete wavelet based denoising. The discrete wavelet transform (DWT) is an effective method to represent and analyze signals, featuring sharp transients. It splits the signal into its "low resolution" parts and a series of details at different resolutions. One of the common applications of the DWT is denoising which receives considerable attention in the
removal of noise in biomedical signals [9, 10]. The denoising algorithm based on DWT has three steps:
s£ — sEMG + n£
(4)
Step 1. The DWT transforms of the EMG signal; Step 2. Defying thresholds of wavelet coefficients; Step 3. Reconstruction of the denoised EMG signal by using inverse wavelet transforms of the thresholded wavelet coefficients.
Fig. 3. EMG model signal, a) the EMG signal from the database; b) synthetic noised EMG signal.
After the wavelet decomposition, a proper threshold must be defined to denoise the signal. The following four classical thresholds are generally used: universal, SURE, minimax and hybrid.
The noise from the synthetic EMG signal was removed using the DWT from the MATLAB Wavelet toolbox. The following mother wavelets were analyzed to denoise the
Standard deviation of E
synthetic EMG signal: Daubechies db4, db6, db8; simlet sym6 sym7 and; the biorthogonal wavelets bior 3.6 - bior 3.8. The choice of the wavelet is determined by the shape of the sEMG signals. The results of denoising the synthetic EMG signal is shown in the table. The smallest error was observed in the case when the mother wavelet db4 of level 6 was used.
Table
3 signal after denoising_
Mother wavelets
DF db4 db6 db8 sym6 sym7 bior 3.7 bior 3.9 bior 6.8
Standard deviation, mV
0.556 0.318 0.337 0.331 0.346 0.321 0.326 0.335 0.321
In the above table: DF - digital filtering: the Butterworth 4th order band-pass filters with cutoff frequencies of 10 and 250 Hz and notch filters -50 and 150 Hz.
Fig. 4 shows the results of denoising the synthetic noised EMG signal using Db4 wavelet at different threshold values.
The effectiveness of the wavelet denoising was demonstrated by processing the real-life raw sEMG signals
obtained from the right M. Biceps Brachii during the right elbow reflection [11]
The recorded sEMG signals were composed of the 50 Hz noise, the ECG signal noise and the EMG signals of the skeletal muscles. Fig. 5 shows the raw and processed signals, and the spectral characteristics of the signals after digital filtering and wavelet denoising.
0 12 3 4 5 6 7 8
time, sec a)
LI
i 1
il v VWjtL:* ■„AvMv.fl1 ■».JWAu
frecuency, Hz
b)
Fig. 4. The denoising of the synthetic EMG signal using Db4 wavelet
time, sec frecuency, Hz
c) d)
Fig. 5 Digital filtering(a, b) and wavelet denoising of the real-life raw sEMG signals (c, d).
References
1. Рангайян, Р.М. Анализ биомедицинских сигналов. Практический подход.- М.: ФИЗМАТЛИТ, 2007. -440 с.
2. Konrad, P. The ABC of EMG A Practical Introduction to Kinesiological Electromyography.- Version 1.4, March 2006.- Noraxon INC. - режим доступа: http://www. noraxon.com/docs/education/abc-of-emg.pdf.
3. Сафин, Д. Современные Системы управления протезами. Конструкции электродов и усилителей биосигналов // ЭЛЕКТРОНИКА: Наука, Технология, Бизнес 4/2009. -режим доступа: http://www. electronics.ru/journal/article/219. (Дата обращения 26.11.2014)
4. Zeeshan O. Khokhar, Zhen G Xiao, Menon C. Surface EMG pattern recognition for real-time control of a wrist exoskeleton// BioMedical Engineering On Line 2010, 9:41. - режим доступа: http://www. biomedical-engineering-online.com/content/9/1/41.
5. Shenoy P. Online Electromyographic Control of a Robotic Prosthesis.- режим доступа: http://homes. cs.washington.edu/~rao/emg-08.pdf
6. Slim Ya. Noise Removal from Surface Respiratory EMG Signal // World Academy of Science, Engineering and Technology Vol:2 2008-02-28
7. Воробьев В.И., Грибунин В.Г. Теория и практика вейвлет-преобразования. - СПб.: Изд-во ВУС, 1999. - 208 с.
8. Bigliassi, M., Fourier and Wavelet Spectral Analysis of EMG Signals in 1-km Cycling Time-Trial. Fnd [et all]// Applied Mathematics, 5, 1878-1886. - режим доступа: http://www.scirp.org/journal/am
9. Phinyomark A., Limsakul C., Phukpattaranont P. A comparative study of wavelet denoising for multifunction myoelectric control// International Conference on Computer and Automation Engineering, ICCAE, pp. 21-25, 2009.
10. Gao J., Sultan H., Hu J., Tung W. W. Denoising nonlinear time series by adaptive filtering and wavelet shrinkage: a comparison// Signal Processing Letters, IEEE, Vol. 17, pp. 237-240, 2010.
11. Бонилья В.Ф., Лукьянов Е.А, Литвин А.В., Деплов Д. А. Влияние кинематических параметров движения локтя на электромиографический сигнал двуглавой мышцы плеча // Вестник ДГТУ №4 - 2014, с. 48-67.
ЭЛЕКТРОМЕХАНИЧЕСКИЙ КОМПЕНСАТОР РЕАКТИВНОЙ МОЩНОСТИ
Черепанов Дмитрий Александрович
Магистр, Петрозаводский Государственный Университет, г.Петрозаводск
Тихомиров Александр Андреевич Канд. физ.-мат. наук, доцент, Петрозаводский Государственный Университет, г.Петрозаводск
Изотов Юрий Анатольевич Бакалавр, Петрозаводский Государственный Университет, г.Петрозаводск
Соболев Никита Владимирович Бакалавр, Петрозаводский Государственный Университет, г.Петрозаводск
ELECTROMECHANICAL REACTIVE POWER COMPENSATOR
Cherepanov Dmitriy, Master's degree, of Petrozavodsk State University, Petrozavodsk, Tikhomirov Alexander, PhD, assistant professor, of Petrozavodsk State University, Petrozavodsk Izotov Yuriy, Bachelor's degree, of Petrozavodsk State University, Petrozavodsk Sobolev Nikita, Bachelor's degree, of Petrozavodsk State University, Petrozavodsk АННОТАЦИЯ
Данная статья посвящена проблеме компенсации реактивной мощности. В ней рассмотрена возможность компенсации реактивной мощности на основе электромеханического эффекта. Результатом работы стало получение данных, подтверждающих переход устройства в режим компенсации электрической энергии. ABSTRACT
This article deals with the problem of reactive power compensation. The article contains the information about a new model of reactive power compensator that works with the help of it's moving part. The result of the work was obtaining data that confirms switch-over of the device in electrical power compensation mode.
Ключевые слова: компенсация реактивной мощности, компенсатор, компенсация, электромеханический компенсатор.
Keywords: reactive power compensation, compensator, compensation, electromechanical compensator.
Электрические компенсаторы реактивной мощности широко применяются для уменьшения потерь электроэнергии [1, с 358]. В данной работе рассмотрен компенсатор реактивной мощности на основе электромеханического эффекта [2].
Конструкция электромеханического компенсатора реактивной мощности, представлена на рисунке 1. Рассматриваемый компенсатор состоит из: движущихся рамок с током, подключаемых к сети в которой компенсируется реактивная мощность (1), обмотки намагничивания (2) и сердечника (3).
