Miokinetic noise attenuation in photoplethysmographic pulse signals by LMS adaptive filtering with adequacy of the input signals

This article describes the adjustments of the preprocessing stage of a digital processing system, used to attenuate the miokinetic movement of photoplethysmographic (PPG) pulse signals by adaptive filtering, in order to improve the response of the filter with respect to the convergence rate and the signal-to-noise ratio, in the least mean squares (LMS) algorithm. Two PPG signals were recorded simultaneously during 3 minutes with two similar recording systems (one placed on the index finger of the right hand and the other on the index finger of the left hand). Both hands were resting during the first minute (in order to get very similar PPG signals) and the left hand did quasi-periodic movements during the last two minutes (these movements were recorded with a 3-axis accel-erometer). After preprocessing the PPG signals (mean removal and normalization), the filtering process was tested with the original signals and also with the preprocessed signals. The filter had a lower average error when using the prepro-cessed signals (9.20%) compared with the use of the original signals processed by the standard LMS (35.75%). The prepro-cessing module improves the performance of the adaptive filtering to attenuate miokinetic noise in PPG signals with a simple adaptive algorithm.