Cluster-based two-step target localization with incremental cooperation

Target localization is one of the momentous research subjects in wireless sensor networks (WSNs). Several methods have been initiated so far for the aim of target (e.g. sniper) localization using the acoustic signals produced, such as muzzle blast and shock wave in WSNs. One of the preeminent available methods is maximum likelihood estimation (MLE) algorithm, using time difference of arrival (TDOA) of the target signal received at sensor nodes. Although the MLE algorithm is asymptotically optimum and obtains high level of accuracy in comparison with other methods, nevertheless, using MLE has two major challenges. Firstly, the crucial need of this method to begin with a proper initial guess, and secondly, the possibility of not converging to a global minimum. Moreover, employing WSNs constrains the amount of power consumption that is practically possible. In this paper, to overcome the aforementioned obstacles, a two-step algorithm is proposed which in first step, a fast spherical interpolation (SI) method is utilized to prepare an appropriate initial guess for the MLE algorithm. In the second step, a clustering-based network is described to attain less power consumption across the WSN. Furthermore, to increase the probability of convergence, a cooperative incremental cluster-based estimation strategy is proposed. In addition, major issues that can affect the performance of the proposed method are investigated. Simulation results prove the capability of this method and support the claims.