A ballistocardiogram (BCG), a measurement of cardiogenic
whole body movements, is a technique that enables non-invasive
cardiovascular monitoring. A main challenge of the BCG
signal is that its morphology and amplitude are sensitive to the posture and / or position of the subject during the recording period. The effects of posture on the BCG measured from a subject standing on a weighing scale have been investigated in the literature, but the effects of
body posture and / or position on BCG signals measured from a subject
lying in a
bed have not been quantified. A contemplated method for
bed-based BCG recordings includes (1) creating templates for standing BCG signals obtained from subjects in a prior study, and (2) quantifying the distance between these templates and BCG waveforms obtained in different
body positions on the
bed for a new set of subjects. In addition, an
array processing technique is presented, which includes the application of
Gaussian weights on a joint probability density function (PDF) and the similarity
score called the q-value to assess those PDFs. The
Gaussian curve weights the joint probability according to the reference value obtained from the previous inter-beat-interval (IBI) estimations. The PDFs are selected and combined according to their reliability measured by q-values. This
array processing significantly reduces the FIR
estimation error by comparing the performance of selective channel combinations to the existing multi-channel
algorithm.