Respiration Signal Monitoring of Multiple Individuals via BSS and UWB-SIMO Radar

Abstract

In recent years, multi-person respiration detection has been more interest, especially in medical environments and emergency situations where several people need to be monitored at the same time. However, it is still challenging to get accurate vital signs from people who are very close together, either at the same distance or in the same angular cell, because the reflected signals overlap. To overcome this limitation, this study employs a Blind Source Separation (BSS) approach using the Fast Independent Component Analysis (FastICA) algorithm to separate multiple targets situated in close proximity or sharing similar angles. The Fast Fourier Transform (FFT) is then used to estimate the respiratory frequencies. Simulation results show that the proposed solution is capable and efficient. They also show that it can successfully separate individual respiratory signals in complex situations with more than one person.