Photoplethysmographic fingertip with miokinetic noise attenuation using an accelerometer and adaptive filtering techniques

We describe the development of a photoplethysmographic fingertip with miokinetic noise attenuation using a pediatric cuff, followed by a 3-axis accelerometer (Acc) that records the movements that distort the photoplethysmographic pulse wave (PPG). Both PPG and accelerometer signals are used as inputs to an adaptive filter in configuration as Adaptive Noise Canceller (ANC) using the averaged least squares algorithm (LMS), the type digital filter is of finite impulse response (FIR) transversal structure. The implementation of adaptive filter algorithm is offline in MATLAB. We take 12 registers from PPG with miokinetic noisy with duration of one minute each one. The results show miokinetic noise attenuation of the PPG signal, which could be used when the PPG records are contaminated with sudden movements, the average error between the PPG recovered by the filter and the PPG without noise miokinetic in percentage terms was 11.5%.