Powerline Interference Reduction in ECG Using Combination of MA Method and IIR Notch

Transcription

1 International Journal of Recent Trends in Engineering, Vol 2, No. 6, November 29 Powerline Interference Reduction in ECG Using Combination of MA Method and IIR Notch Manpreet Kaur, Birmohan Singh 2 Department of EIE, Sant Longowal Institute of Engineering & Technology, Longowal- 486 Sangrur(Pb) India aneja_mpk@yahoo.com 2 Department of CSE, Sant Longowal Institute of Engineering & Technology, Longowal- 486 Sangrur(Pb) India bskomal@yahoo.com Abstract The ECG Signal is a graphical representation of the electromechanical activity of the cardiac system. It is one of the most important physiological parameter, which is being extensively used for knowing the state of cardiac patients. For diagnostic quality ECG recordings, signal acquisition must be noise free. Since ECG signals are only of the order of mv in amplitude, the ECG acquisition is susceptible to many types of noises and power line interference is one of them. Many types of analog and Digital Filters have been proposed/ suggested by the researchers. The present paper has proposed a combination technique of two methods i.e. the moving averages technique and IIR notch to reduce the power line interference. The results have clearly indicated that there is reduction in the power line noise in the ECG signal. The filter proposed in this paper has fewer coefficients and hence lesser computation time. So the above combination is applicable for real time processing. The results have been concluded using Matlab and MIT-BIH database (with 6 Hz powerline interference) Index Terms ECG, Powerline Interference, IIR, Notch, Moving Average I. INTRODUCTION TO ECG SIGNAL The function of the heart is to contract rhythmically and pump blood to the lungs for oxygenation and then pump this oxygenated blood into the general circulation. This perfect rhythm is continuously maintained and signaled by the spread of electrical signals generated by the heart pacemaker, the Sinoatrial (SA) node. Detecting such electrical activity of the heart can help identify many heart disorders. This is the main concept behind using an ECG (Electrocardiogram), tracing the electrical activity of the heart. By measuring and tracing the potential difference between two points on the outer surface of the body we obtain the simplest ECG chart. Two typical measuring points are between left arm and the right arm. By defining two points and setting up conventional positive direction for measuring the voltage what created is called a Lead. The first phase of cardiac muscle activation is the stimulation of the right and left atria by an electrical signal generated from SA node. This phase appears as the P-wave on the ECG chart. The electrical signal generated, originally generated by the SA node, then spreads through the Atrioventricular node (AV) Junction, the bundle of HIS, and the Purkinje fibers to finally reach and stimulate the ventricles. The spread of electrical signal through ventricles causes ventricular contraction. The phase of ventricular contraction appears as the characteristic QRS complex on ECG chart. Finally, with the relaxation of two ventricles, a depolarization signal is generated and appears as the T-wave on ECG chart.[] Fig.: The ECG Signal (Wikipedia) II. THE NOISES IN ECG SIGNAL The common noise sources are Baseline wander and ECG amplitude modulation with respiration, Power line interference, Muscle contraction noise, Electrosurgical noise, Motion artifacts, Noise due to variation of electrode skin contact impedance and Noise generated by electronic devices used in signal processing circuits Power line interference noise is electromagnetic field from the powerline which causes 5/6 Hz sinusoidal interference. This noise causes problem in interpreting low amplitude waveform like ECG. Although modern instrumentation amplifiers/differential amplifiers have high common mode rejection ratio and other measures like shielding and grounding are taken care of while recording ECG, even then the recordings are often contaminated by powerline interference. In order to avoid the wrong identification of the characteristics of ECG signals and its impact on analysis and diagnostic accuracy, it is necessary to remove the power-line interference effectively. III. ECG FILTERING ECG signal has been a major diagnostic tool for the cardiologists and ECG signal provides almost all the information about electrical activity of the heart. So care should be taken while doing the ECG filtering, such that the desired information is not distorted or altered in any way. Although many methods have been suggested by researchers to reduce the Powerline noise yet filtering introduces unacceptable ringing effect, especially after the QRS complexes. 29 ACADEMY PUBLISHER 25

2 International Journal of Recent Trends in Engineering, Vol 2, No. 6, November 29 Many Researchers have worked on the removal of the power line interference in the ECG signals, S.Pooranchandra, N.kumarave[4] have used the wavelet coefficient threshold based hyper shrinkage function to remove power line frequency. Santpal Singh Dhillon and Saswat Chakrabarti[5] have used a simplified lattice based adaptive IIR Notch filter to remove power line interference. Mahesh S. Chavan, R.A. Aggarwala, M.D.Uplane[6] have used Digital FIR Filters based on Rectangular window for the powerline noise reduction. ChavdarLevkov, Georgy Mihov, Ratcho Ivanov, Ivan Daskalov, Ivaylo Christov, and Ivan Dotsinsky,[7] have used the subtraction procedure to remove the powerline noise Ziarani AK, KonardA [8] used a nonlinear adaptive method to remove noise. LIN Yue-Der, YU HEN HU[9] have proposed a PLI detector that employs an optimal linear discriminant analysis (LDA). G. Mihov, Iv Dotsinsky, Ts Georgieva[] have proposed subtraction procedure for powerline interference removing from ECG which is extended to almost all possible cases of sampling rate and interference frequency variation. IV. COMBINATION METHOD : THE PROPOSED TECHNIQUE A major source of interference in ECG signals is the 5/6Hz power line frequency. The 6Hz frequency component can be removed by using IIR notch filter. A very simple approach to filter powerline interference is to create a notch filter with zeros on the unit circle in the z-domain at the specific frequencies to be rejected If f o is the interference frequency, the angles of the (complex conjugate) zeros required will be ± f o /f s (2π); the radius of the zeros will be unity. The sharpness of the notch may be improved by placing a few poles near or symmetrically around the zeros and inside the unit circle. But the increased transient response time results in a ringing artifact. The general, more sophisticated filters can be constructed by having a narrower notch but it is not possible to design an IIR notch to remove the noise without causing ringing.[2] So a combination technique has been proposed /suggested to filter out 5/6 Hz power line interference in this paper. In this proposed method the moving average algorithm is applied to the ECG signal first and afterwards the IIR notch is applied. A. Moving Average Method Moving averages algorithm is applied in cases where a given waveform is cluttered with noise, where a mean needs to be extracted from a periodic signal or where a slowly drifting baseline needs to be eliminated from a higher frequency signal. A moving average filter smoothes data by replacing each data point with the average of the neighboring data points defined within the span. The algorithm accomplishes a moving average by taking two or more number of data points from the waveform data, adding them, dividing their sum by the total number of data points added, replacing the first data point of the waveform with the average just computed and repeating the steps with the second, third and so on data points until the end of the data is reached. The result is a second or generated waveform consisting of the averaged data and having the same number of points as the original waveform. The moving average of a waveform can be calculated by: Where a is averaged value, n is data point position, s is smoothing factor, and y is actual data point value. The procedure used by the algorithm to determine the amount of filtering involves the use of smothing factor. For the present work the appropriate smoothing factor has been determined by dividing one waveform period with the channel s sample interval.[3] B. The IIR Notch Characteristics The filter is designed taking care that lesser ringing effect is introduced and the desired information is not altered in any way. The filter characteristics have been shown using FDATool design box, in Fig..2 to.8 Magnitude (db) Magnitude Response (db) Fig.2: Magnitude response Phase (degrees) Phase Response Fig.3: Phase response Amplitude Step Response Time (seconds) 29 ACADEMY PUBLISHER 26

3 International Journal of Recent Trends in Engineering, Vol 2, No. 6, November 29 Fig.4: Step response 2 Original ECG.2 Impulse Response Amplitude Time (seconds) Fig.5: Impulse response Fig.8: The original ECG signal (Sample) 25 Original ECG Pole/Zero Plot Imaginary Part Real Part Fig.9: The original ECG signal (Sample2) PSD of Original ECG Fig.6: Pole Zero Plot Round-off Noise Power Spectrum :.679 Pow er/frequency (db/hz): Power (db) Fig.: PSD of original signal (Sample) Fig.7: Round of noise power spectrum 8 PSD of Original ECG V. THE RESULTS The results are obtained by taking the data of sample signals available in BIT-MIH database and are shown below : :.679 Pow er/frequency (db/hz): Fig.: PSD of original signal (Sample2) Two sample signals taken from MIT-BIH database have been plotted and their Power spectral density has been found at 6 Hz and the value has been shown on the graph itself. IIR Notch is applied now on the original ECG signal.(fig.8,.9) to get the signal as in Fig.2, ACADEMY PUBLISHER 27

4 International Journal of Recent Trends in Engineering, Vol 2, No. 6, November 29 2 ECG with IIR Notch 4 PSD of ECG after removing baseline drift :.679 Pow er/frequency (db/hz): Fig.2: ECG signal with IIR notch (Sample) 25 ECG with IIR Notch Fig.6: PSD of ECG after removing Baseline drift (Sample ) 2 5 PSD of ECG after removing baseline drift : Pow er/frequency (db/hz): Fig.3: ECG signal with IIR notch (Sample2) 2 ECG after removing Baseline drift Fig.7: PSD of ECG after removing Baseline drift (Sample 2) The above figures show the removal of the baseline drift of original ECG signal without disturbing the waveform information. The power spectral density shown for both the samples after baseline drift has been removed, indicated a very small change in the power at 6 Hz when compared with the power spectral density of original ECG signal. Now IIR Notch filter is applied to this smoothed signal, the following results were obtained: Fig.4: ECG after removing Baseline drift (Sample ) 2 ECG after removing Baseline drift and then applying IIR Notch 3 ECG after removing Baseline drift Fig.5: ECG after removing Baseline drift (Sample 2) When IIR notch was applied on the sample signals, it was observed that ringing effect is introduced at the starting as well as at the QRS complexes. Now Moving Average method is applied on the original ECG signal and its PSD has been plotted Fig.8: ECG after removing Baseline drift and then applying Notch (Sample ) 29 ACADEMY PUBLISHER 28

5 International Journal of Recent Trends in Engineering, Vol 2, No. 6, November ECG after removing Baseline drift and then applying IIR Notch is reduced significantly. The filter proposed in this paper has fewer coefficients and hence lesser computation time. So the above combination is more suitable and efficient for real time processing of ECG signals. REFERENCES Fig.9: ECG after removing Baseline drift and then applying Notch (Sample 2) PSD of ECG after removing baseline drift and applying Notch :.679 Pow er/frequency (db/hz): Fig.2 : PSD of ECG after removing Baseline drift, then applying Notch(Sample ) PSD of ECG after removing baseline drift and applying Notch :.679 Pow er/frequency (db/hz): Fig.2 : PSD of ECG after removing Baseline drift, then applying Notch (Sample 2) The above ECG signals shown for both the samples have clearly indicated lesser ringing effect as compared to the ECG signals obtained after applying IIR notch filter directly on the original signal. The power spectrums have shown a change from dB to dB for signal Sample and from dB to db for signal Sample 2. [] Wikipedia. [2] Rangarej M Rangayyan, Biomedical Signal Analysis A case study approach, Wiley Interscience. [3] DATAQ Instruments. [4] S.Pooranchandra, N.Kumaravel, A novel method for elimination of power line frequency in ECG signal using hyper shrinkage function, Digital Signal Processing, Volume8, Issue 2, March 28, pp [5] Santpal Singh Dhillon, Saswat Chakrabarti, Power Line Interference removal From Electrocardiogram Using A Simplified Lattice Based Adaptive IIR Notch Filter, Proceedings of the 23 rd Annual EMBS International conference, October 25-28, Istanbul, Turkey, 2, pp [6] Mahesh S. Chavan, R.A.Aggarwala, M.D.Uplane, Interference reduction in ECG using digital FIR filters based on Rectangular window WSEAS Transactions on Signal Processing, Issue 5, Volume 4, May 28, pp [7] Chavdar Levkov, Georgy Mihov, Ratcho Ivanov, Ivan Daskalov, Ivaylo Christov, and Ivan Dotsinsky, Removal of power line interference from the ECG : a subtraction Procedure, Biomed Eng Online, 25;4;5, published online 25 August 23 doi:.86/ x-4-5. [8] Ziarani AK, Konard A, A nonlinear adaptive method of elimination of power line interference in ECG signals, IEEE Trans BiomedEng; 49(6), Jun 22, pp [9] LIN Yue-Der, YU HEN HU, Power-Line Interference Detection and Suppression in ECG Signal Processing, IEEE Transactions on Biomedical Engineering ISSN , 28, vol. 55, pp [] G. Mihov, Iv Dotsinsky, Ts Georgieva, Subtraction procedure for powerline interference removing from ECG: improvement for non-multiple sampling, Journal of Medical Engineering &, Technology, Volume 29, Issue 5, September 25, pp [] Ju-Won-Lee, Gun-Ki Lee, Design of an Adaptive Filter with a Dynamic Structure for ECG signal Processing, International Journal of Control, Automation and Systems, vol.3, no., March 25, pp [2] Mahesh S. Chavan, R.A.Aggarwala, M.D.Uplane, Design and Implementation of Digital FIR Equiripple Notch Filter on ECG Signal for removal of powerline Interference, WSEAS Transactions on Signal Processing, pp [3] Digital Filters Plus, User s Guide, Version 2.4, Numerix Ltd., Leicestershire, LE 67 4QQ UK CONCLUSIONS The IIR notch applied directly on the ECG sample has shown more ringing effect as discussed earlier in this paper that the QRS complex influence the output of the filter that is QRS complex often poses a large unwanted impulse to the filter. When the same filter is applied to the ECG signal which has been smoothed first i.e. having the baseline wander removed, lesser ringing effect has been observed. Also there is lesser variation of the PQRST Waveform. The power at 6Hz 29 ACADEMY PUBLISHER 29