Motion Artifact Removal and Feature Extraction from PPG Signals Using Efficient Signal Processing Algorithms

The performance of wearable Photoplethysmographic Biosensors can be highly influenced by the motion artifacts. This work proposes a performance enhancement algorithm which can remove the effect of motion artifacts caused by the voluntary movements during various physical activities of an individual. We have developed and implemented three motion artifact removal algorithms namely, ICA-Adaptive Filter Algorithm, Butterworth-ICA-Adaptive Filter Algorithm and Butterworth-Wavelet Transform Algorithm. These three algorithms were analyzed under four fingertip movements like vertical movement, horizontal movement, shivering, and applying pressure. Based on the analysis we found that the Butterworth-Wavelet Transform Algorithm is better in providing high Signal to Noise Ratio (SNR) without compromising any signal characteristics and the algorithm validation was done by extracting Heart Rate (HR) and Peripheral Oxygen Saturation (SpO
₂
) values using Photoplethysmographic (PPG) signals obtained from available Biosensor. The results are found promising and suggest that the Butterworth-Wavelet Transform Algorithm provides motion artifact-free PPG signal for accurate feature extraction.