Optimized Implementation of ECG Signal Noise Cancelation Using Fir and IIR Filter Techniques Based On FPGA

Abstract

The procedure of measuring the heart's electrical activity over time by putting electrodes on the human body places is known as an electrocardiogram (ECG). The long-term (ECG) patterns acquired in the intensive care unit are frequently distorted by different types of noises. These varieties of noises included with ECG signal such as power source interference, artifacts, muscular contraction, equipment noise, and electrode-contact noise may be present in the collected ECG signal. Which can cause several false alarms, including the incorrect diagnosis of atrial fibrillation. These noise types must be removed through preprocessing of the ECG data recording. This study proposes the ECG signal filtered design for elimination most of all these types of noise using a variety of techniques, including finite impulse response (FIR) and infinite impulse response (IIR). The least mean square adaptive filter and band pass filter is employ to improve the filter response and performance. The performance metrics include competition time, mean square error (MSE) and signal to noise ratio (SNR) are employed in order to determine which approach is best for the obtained noiseless ECG signal. The current work, applies the simulated ECG signal combined with a Gaussian white noise (GWN) and other type of noise as the input signal for filter testing and evaluating. The different implementation stricture methods are explored for filtering implementation process. The direct form-II second order section is chosen construction method as it more effective in term of hardware size reduction. Then, comparisons are conducted based on time, hardware use, and filter order requirement. There are several window techniques available to optimize the FIR filter design, the present paper utilize the Kaiser window for this design because it is one of the popular methods. Although there are alternative approaches for filtering the ECG signal through the IIR filters, the current investigation applies the Butterworth approach for IIR filters optimization for the system implementation, because it is one of the common procedures. The optimal filter order, construction, SNR, MSE, hardware consumption and processing speed for each the filtering technique are discovered. This investigation include the filter response in time and frequency domain for several filtering orders and techniques. The Matlab software is adopted in this work for identifying the best optimal filter among all these methods. In addition, the hardware evaluating is used to assessments the filter noise cancelation using several hardware implementation approachs. The Xilinx platform and field programmable gate arrays (FPGA) Board are employed for this task because the ability and flexibility of the FPGA for reconfiguration. Moreover, all these hardware filters are tested and examined in real time with artificial ECG signal generator mixed with disturbances including Electromyography (EMG) signals of the muscle surround the electrode and GWN.