Ratiometric Artefact Reduction in Low Power, Discrete-Time, Reflective Photoplethysmography

摘要

This paper investigates the feasibility of a ratio metric approach to compensating for ambient light and motion artefacts in a reflective photoplethysmography (PPG) sensor suitable for wearable applications. A low-power, discrete-time pulse-oximeter development platform is used to capture both infra-red (IR) and red photoplethysmograms, as well as the ambient light level, such that the data used for analysis has noise levels representative of what a true body sensor network (BSN) device would experience. The performance of a ratiometric artefact reduction technique is tested with both simulated multiplicative noise sources, as well as real motion artefacts to best determine where the limitations of the system lie.