RECOVERING VITAL PHYSIOLOGICAL SIGNALS FROM AMBULATORY DEVICES

摘要

Sensors that are attached to a patient for monitoring the variousvital physiological parameters, under ambulatory conditions,often acquire signals with very low signal-to-noiseratio. One characteristic that is observed is that the acquireddata has the signal and the noise exhibiting overlappingspectral characteristics. Drawing a meaningful analysis,from these signals, can be quite a challenge when theclassical linear filtering techniques are unable to separatethe signal from the noise. In addition, the signals are nonstationaryin nature. So, any method that enhances thesesignals should safeguard the salient points within the seriesand avoid smoothing of the feature rich segments. In thiswork, we propose a method to recover the vital physiologicalsignals from the observation. As the problem is notwell posed, it is redefined by introducing some smoothnessconstraints on the solution. We introduce different waysof modeling the problem, and show how the physiologicalsignal can be recovered with minimal loss in information.A numerical case example is taken from Electrocardiogramsignals to demonstrate the proposed approach, while comparisonis made with some popular techniques from the literature.