Ratiometric Artifact Reduction in Low Power Reflective Photoplethysmography

This paper presents effective signal-processing techniques for the compensation of motion artifacts and ambient light offsets in a reflective photoplethysmography sensor suitable for wearable applications. A ratiometric comparison of infrared (IR) and red absorption characteristics cancels out noise that is multiplicative in nature and amplitude modulation of pulsatile absorption signals enables rejection of additive noise. A low-power, discrete-time pulse-oximeter platform is used to capture IR and red photoplethysmograms so that the data used for analysis have noise levels representative of what a true body sensor network device would experience. The proposed artifact rejection algorithm is designed for real-time implementation with a low-power microcontroller while being robust enough to compensate for varying levels in ambient light as well as reducing the effects of motion-induced artifacts. The performance of the system is illustrated by its ability to extract a typical plethysmogram heart-rate waveform since the sensor is subjected to a range of physical disturbances.