Algorithm for Detection with Localization of Multi-targets in Wireless Acoustic Sensor Networks

Abstract

In most multitarget tracking approaches based on joint probabilistic data association (JPDA), it is difficult to apply the solutions to problems (due to the dimensionality curse of heavy complexity) where the number of target varies dramatically. In this paper, we introduce an algorithm for detection of multitargets in wireless acoustic sensor networks (ADMAN); we localize detected targets by particle filtering after ADMAN. The purpose of ADMAN is detecting any number of targets (We know the approximate locations of targets during the detection algorithm.) in the field of interest. The advantage of ADMAN is its ability to cope with varying number of targets in time. ADMAN does not have any restrictions on the varying pattern of the target number