Object detection and localization using compressed sensing

Abstract

Localization is very significant in Underwater Sensor Network (UWSN) applications. The functionality of the network can face challenges by the force of water current and hostile environmental conditions. This work presents a method for localization using Compressed Sensing (CS). It is implemented without using GPS technology which makes the method reliable. CS is employed in the data acquisition module for transmission and reconstruction of audio signal. It is a dictionary based execution exploiting l1 minimization using Gabor transform. Here, localization using audio is performed using Time Difference of Arrival (TDOA). Moore-Penrose pseudo-inverse is used for matrix operations. An array of audio sensors or hydrophones is assumed while performing this work. The results of simulation indicate that this is an efficient technique for object detection and localization.