Motion Artifacts Detection from Ambulatory ECG using Wavelet Transform Sachin T. Darji 1 , Rahul K. Kher 2 1, 2 G. H. Patel College of Engineering and Technology, V.V.Nagar, Gujarat, India Email: sachin_ec453@yahoo.co.in, rahul2777@yahoo.com Abstract—Ambulatory ECG signal analysis for detection of various motion artifacts using discrete wavelet transform is addressed in this paper. Ambulatory ECG monitoring provides electrical activity of the heart when a person does normal routine activities. The recorded ECG signal consists of cardiac signal along with motion artifacts introduced due to person’s body movements during routine activities. Detection of motion artifacts due to different physical activities might help in diagnosis. We have used BIOPAC MP 36 system for acquiring ECG signal. The ECG signal was recorded for 300 seconds with sampling frequency of 500 Hz while the person performs various body movements like up and down movement of left hand, up and down movement of right hand, waist twisting movement while standing and change from sitting down on chair to standing up movement. We have employed discrete wavelet transform method to determine the wavelet coefficients that correspond to approximate cardiac signal. The signal decomposition was performed with levels 5 and 7. Different wavelets like bior, db, sym were used for decomposition. The synthesized cardiac signal was then subtracted from the originally recorded ECG signal with motion artifacts to get estimated motion artifact signal. Various statistical parameters like mean, standard deviation, variance and median were calculated for the estimated motion artifact signals. Index Terms—Ambulatory ECG, Body Movements, Motion artifact, DWT. I. INTRODUCTION MBULATORY ECG signal monitoring is gaining popularity due to long term monitoring and convenience they offer. With Ambulatory ECG device the ECG signal can be recorded while the person can perform daily routine activities [4]. The major challenge with ambulatory ECG monitoring is that the cardiac signal gets altered due to motion artifacts resulting due to body movements [2]. Also, the motion artifact signal has spectral overlap with cardiac signal in 1-10 Hz which corresponds to ECG features like P wave and T wave. While attempting to remove motion artifact completely, it also affects the cardiac features [3]. In [1], study of motion artifacts resulting from body movements is presented to improve the accuracy of cardiac monitoring. Researchers have developed various techniques to get information regarding physical activity [1] [3]. The wavelet based methods are widely used in pre-processing, denoising and analysis of ECG signals [4]. In [6], DWT method is used to extract ECG parameters. Principal component analysis is used in [2] to estimate the motion artifact component. The organization of the paper is as follows. Different body movement activities and data acquisition is described in section II. Use of discrete wavelet transform for estimation of motion artifacts is described in section III. Experimentation procedure and obtained results are presented in section IV. The conclusions are marked in section V. II. DATA ACQUISITION-HARDWARE We have used BIOPAC MP 36 system that integrates hardware and software for data acquisition process. The specifications used in this study are as follows: bandwidth-0.05 Hz to 35 Hz, sampling frequency- 500 Hz, recording duration- 300 seconds. The Lead-I configuration is chosen for data acquisition. The following body movement activities were performed in this experiment [2]: 1) left arm up and down movements, 2) right arm up and down movements, 3) twisting of waist left-right-left while standing, 4) change from sitting on a chair to standing up and vice versa. Two healthy male subjects were asked to perform above activities at a normal pace. III. WAVELET TRANSFORM An ECG signal corrupted with motion artifacts is a non- stationary signal which cannot be analyzed effectively with Fourier transform (FT). FT gives very good frequency localization, but no time localization. Short term Fourier transform (STFT) may be used for analysis of such signals. In this technique the signal is divided into small parts by windowing and the Fourier transform of those signal segments are taken to obtain STFT. The STFT provides time localization as well as frequency localization but due to the fixed window length performance is limited and there is a lower limit on time and frequency resolutions. Taking these shortcomings into consideration as well as the nature and possible characterization to be performed on the signal, Wavelet Transforms have been used in this study [5]. The discrete wavelet transform (DWT) is found to yield a fast computation of wavelet transform. It is easy to implement and reduces the computation time and required resources. In DWT, a time-scale representation of the digital signal is obtained using digital filtering techniques. The signal to be A