Converting the Nonlinear Least Squares Problem for Localization in Wireless Sensor Networks

Estimating the positions of sensor nodes precisely is one of the major tasks in Wireless Sensor Networks (WSN). Due to the limited hardware resources and the energy capacity it is necessary to develop energy-aware localization algorithms. Here we present a new technique to convert the complex non-linear least squares calculation and distribute the tasks over the network effectively. Our proposed algorithm achieves more than 85% energy savings compared to the "Fine-Grained Localization"algorithm (FGL) by Savvides et al., more than 12% compared to the high-performance "Distributed Least Squares"-algorithm (DLS) by Reichenbach et al. while improving the localization accuracy simultaneously. Moreover our algorithm is more robust against beacon failures than DLS.