Abstract

Info

Title:: Photoplethysmography signal smoothing technology based on locally orthogonal weighted polynomial fitting

Author(s):: LI Jinlu1; LAI Zhanyu2; DONG Keyang2; DUAN Yufan3; DAI Zidong3; LIU Yurong2; JIANG Xiaoping3; 1. Nuctech Company Limited, Beijing 100084, China; 2. School of Artificial Intelligence, China University of Mining and Technology(Beijing), Beijing 100083, China; 3. School of Mechanical and Electrical Engineering, China University of Mining and Technology(Beijing), Beijing 100083, China

Keywords:: photoplethysmography; locally orthogonal weighted polynomial fitting method; signal smoothing

Abstract:: Abstract: To address the issue of reduced signal quality of photoplethysmography caused by local fluctuation, an approachcalled locally orthogonal weighted polynomial fitting (LOWPF) is proposed for signal smoothing. After determining thepositions of the fluctuation sequences using the forward-backward difference XOR method, weighted polynomial fitting isapplied to these sequences, and the fitted values are used to replace the fluctuation sequences to achieve signal smoothing. Byconstructing orthogonal basis functions, the condition number of the coefficient matrix is reduced, and the stability of theequation system solution for higher-order fitting is improved. Simulation results demonstrate that the smoothed signal’s XORsmoothness of the proposed method surpasses that of the moving average algorithm and the empirical mode decompositionreconstruction algorithm. The smoothing results on 241 sets of measured PPG signals show that LOWPF achieves anefficiency of smoothness of 89.10%, significantly higher than the 78.05% of empirical mode decomposition and the 59.13%of the 5-point moving average algorithm. LOWPF has promising application prospects for smoothing signals with significantlocal fluctuations.