Cascaded output selection for processing of capacitive electrocardiograms by means of independent component analysis

Innovative measurement systems allow for the contactless recording of vital signs. Thus, applications with medical background for daily life become possible. Aquired signals, however, often cannot compete with their clinically established counterparts. In fact, typical characteristics as small signal amplitudes on the one hand, frequently occuring artefacts and noise on the other hand, introduce the apparent need for sophisticated processing techniques to allow for a reliable function when thinking of contactless measurements. This contribution investigates the possibility of using multichannel capacitive electrocardiogram (cECG) recordings to derive the heart rate for driver monitoring. We propose a processing scheme consisting of a spatio-temporal independent component analysis applied to the cECG together with a newly developed method to select the most appropriate of the output channels by analyzing their frequency characteristics. By an experimental study incorporating 27 healthy subjects we prove the applicability of our method and discuss its advantages compared to existing methods.