Improving localization in wireless sensor networks with an evolutionary algorithm

Abstract

Wireless sensor networks are highly useful for many location-sensitive applications including environmental monitoring, military applications, disaster management, etc. Localization in wireless sensor networks concerns about the precise estimation of node positions given a relatively small portion as anchor nodes with their absolute positions predetermined. Intrinsically, localization is an unconstrained optimization problem based on various distance/path measures. Most of the existing work focus on increasing the accuracy in position estimation typically by using different heuristic-based or mathematical techniques. On the other hand, there were many complex optimization problems successfully tackled by the nature inspired search algorithms including the ant-based or genetic algorithms. In this paper, we propose to adapt an evolutionary approach, namely a microgenetic algorithm, and integrate as a post-optimizer into some existing localization techniques such as the Ad-hoc Positioning System (APS) to further improve their position estimation. Clearly, our proposed MGA is so adaptable that it can easily be integrated into other localization methods. More importantly, the remarkable improvements obtained by the prototype of our proposed evolutionary optimizer on certain anisotropic topologies of our simulation tests prompt for further investigation. © 2006 IEEE.