Detection and localization in a wireless network of randomly distributed sensors

This paper introduces a statistical methodology for target detection and localization in a network of randomly distributed wireless sensor nodes. The aim is to detect and localize (with respect to the nearest sensor) a ´source´ with minimal ´a priori´ knowledge of sensor locations. The novelty of the proposed detection and localization algorithms is that they are viewed as two independent classification problems which are solved using pattern recognition techniques. In addition, the datasets required for each classification problem are created by a unique autonomous training mechanism. The performance of the proposed algorithms is evaluated through Monte Carlo simulations and is demonstrated to be robust in the presence of noise and changes in the propagation environments.