Design and Implementation of Hybrid Techniques and DA-based Reconfigurable FIR Filter Design for Noise Removal in EEG Signals on FPGA

Abstract

Virtual Reality (VR) technology assists physically challenged personnel in their daily routine activities. The evolution of technology has enhanced the critical activities of people who use wheelchairs by extracting features through electroencephalogram (EEG) and promoting options for their choice for decision-making on their own. During extraction of EEG, signal artifacts may mislead the decision-making environment. Hence noise has to be removed with help of an FIR filter for accuracy. In this context utilization of finite impulse response (FIR) filters are so vital hence filters are incorporated with the hidden Markov model (HMM) and Gaussian mixture model (GMM) and supervised machine learning architecture of multirate support vector machine (SVM). The proposed EEG-based diagnosis system is a fully automated audio announcement system. The entire environment has been developed by Verilog HDL and MATLAB. Validated on Artix-7 FPGA development board and synthesized with Vivado Design Suite 2018.1. Obtained results exhibit an enhancement of 32% of signal-to-noise ratio (SNR),7% of mean square error (MSE), and 69% of abnormality recognition.