A new multi-human detection algorithm using an IR-UWB radar system

Abstract

This paper presents a multi-human detection algorithm using an impulse radio ultra-wideband (IR-UWB) radar system. The algorithm operates in a similar manner to the conventional clean algorithm that has mainly been used for estimating a channel impulse response (CIR) by computing deconvolution from the clutter subtracted received signal. Unlike the clean algorithm, the algorithm attempts to divide the clutter subtracted received signal into a number of coherent clusters so that the detection success rate of the multi-humans can be improved without a priori knowledge of the number of existing humans. The algorithm performs quite well in dense multipath environments that take place almost always in practice. The validity of the algorithm is demonstrated by successfully tracking the distances of randomly walking people using a commercial IR-UWB radar system. The detection error rate is also discussed based on the normalized histogram obtained from the experimentally measured data.