Real-Time PPG Signal Quality Assessment System for Improving Battery Life and False Alarms

Abstract

This brief presents a new real-time photoplethysmogram (PPG) signal quality assessment system for significantly reducing an overall energy consumption and reducing frequent false alarms. The proposed algorithm is based on the hierarchical decision rules by combining with simple features, such as absolute amplitude, threshold crossing rate, and autocorrelation function features. The algorithm is validated on both MATLAB software and Arduino Due with 32-bit ARM core microcontroller by using the standard PPG datasets and our real-time PPG datasets. The algorithm achieved a sensitivity of 99.29%, specificity of 95.31% and accuracy of 97.76% with 754 false positive segments and 182 false negative segments of a total of 16 068 unacceptable and 25 772 acceptable segments, respectively. The algorithm provides a promising false alarm reduction rate of 95.31% on different kinds of artifacts and noise. The proposed quality-aware monitoring system can save transmission energy consumption from 3.43% to 95.1% for transmission of noisy 5–60 s PPG signal.