POWER LINE HUM CANCELLATION FROM ELECTROCARDIOGRAPHY USING LMS ALGORITHM Jigar H. Shah 1 , Jay M. Joshi 2 , Viranchi C. Pandya 3 Shri S’ad Vidya Mandal Inst. Of Tech., Old NH-08, Bharuch 1,2 , Gujarat A.D.Patel Institute Of Tech., New Vallabh Vidyanagar 3 , Gujarat India jhs707in@yahoo.com 1 , jaymjoshi@yahoo.com 2 , pandyavc2001@yahoo.com 3 ABSTRACT Biomedical Instrumentation is one of the fields where the accuracy of the instruments can not be compromise with the cost. Biomedical instruments have to deal with the signals which have lower amplitude as well low frequency, so removal of the low frequency noise like power line humming noise from these signals is the challenging task. Adaptive signal processing is the one of the techniques, which can be used to process these sensitive signals because adaptive filters change its coefficients with the change into the property of the incoming signals. In this paper, we present problem occurring in convectional signal processing schemes for removal of low frequencies power line noise of 50 Hz from ECG signal and how the adaptive signal processing solve this problem with appropriate simulation results. KEY WORDS Hum, Adaptive Noise Cancellation, LMS 1. Introduction and problem formulation Measurements of biological activity by means of monitoring electrical discharge, as typified by the monitoring of heart patients, parallels the communication problem: A transmitter (the electrical discharge) radiates energy through a propagation path (the body’s tissue) to a receiving antenna (an electrode) positioned to maximize energy reception. Because the electrical discharge involves very small potentials, the received energy is very weak and requires care to prevent degradation of the signal content by added noise or filtering. Probably the strongest source of interference is 50 Hz [12] . Pick up and its harmonics emanating from nearby electrical equipment such as lighting and instrument power supplies. Fig. 1 shows the sensor signal corrupted by additive low frequency power line hum noise. The remote sensor is connected to an amplifier via length of the cable. The amplifier output contains not only the sensor signal, but also 50 Hz component due to stray pick up may be often associated with sensitive high impedance sensor like microphone or electrodes. Such power line hum may prove to be merely bothersome in hi-fidelity sound reproduction system. But at the same time it can be overwhelming when attempting to measurement of very small voltages such as found when trying to obtain a patient’s electrocardiogram (ECG). So, it may produce wrong result. Our goal here is to produce output that closely resembles sensor signal. 2. Limitation of conventional signal processing methods If the spectrum of the sensor signal does not include 50Hz, then the hum component can be removed via notch filtering. But ECG measurement can not tolerate the exclusion of signal energy below 100Hz. So solution to removal of power line hum may be notch FIR filter (Fig. 2), as long as the electric power is steady enough to remain within the notch of the filter. We can measure the power line mains frequency and the same can be used to eliminate power hum component entering the amplifier. Thus, then the amplifier input and output will contain only the sensor signal as shown in fig. 3. The problem with this approach is that the amplitude and phase of 50 Hz must be carefully adjusted to make it accurately cancel the hum at the input of the amplifier. In most cases, amount of 50Hz pick up available at our home, may change in amplitude and phase with time on which we do not having any control. So, as a practical matter then the gain and the phase adjustment can be applied to the reference 50Hz waveform, which must be variable and some automatic techniques should be available to adjust them in real time to assure good cancellation. Fig. 1 Problem formulation Amplified signal Plus Hum Sensor Signal Amplifier Power Line Hum A cos(ωt+θ) + + G Fig. 2 Hum removal by inline filtering Amplified signal Plus Hum Sensor Signal Amplifier Power Line Hum A cos(ωt+θ) + + G Notch Filter