Multi-Human Detection Algorithm based on an Impulse Radio Ultra-Wideband Radar System

Abstract

In this paper, we propose a multi-human detection algorithm based on impulse radio ultra-wideband radar system. With our proposed algorithm, the multi-human detection can be performed by repeatedly performing clustering and detecting processes. More specifically, the system detects an effective peak of the first single cluster which is composed of peaks adjacent to each other and then repeat this process until effective peaks of clusters caused by multiple people are successfully detected sequentially. As our performance metrics, we take into account the performance analysis in terms of the error probability based on the results of the statistical analysis. More specifically, we first cross-verify that the empirical result theoretically follows the Log-normal distribution by comparing the theoretical and empirical results obtained through laboratory experiments. Then, we statistically analyze the received signals under the Log-normal distribution assumption. After that, this statistical result is adopted to the performance analysis of the error probability in terms of the total error probability. Note that the performance of our proposed algorithm is affected by the threshold value. Based on it, the optimal threshold is analyzed and we provide the sample guidelines for optimally adjusting the threshold value under given various environment factors. Finally, some selected experimental results are presented to show the validity of our proposed algorithm by comparing the performance between the proposed algorithm and the conventional algorithm.