Accurate Localization Under Correlated Errors for Underwater Acoustic Sensor Networks

Localization is an indispensable part for most underwater wireless sensor networks (UWSNs). In these networks, acoustic ranging signals are typically used. Because of the depth dependent sound speed, acoustic ray bends in water. This presents a unique problem making conventional localization algorithms assuming straight line distances between target and anchor nodes inaccurate. In this paper, authors present a Maximum Likelihood estimation scheme for accurate location estimation in UWSNs where there is correlation of range measurement errors. The method takes into account ray bending by making use of ray tracing, with sound speed profile as input. The performance of the scheme was studied by means of Monte-Carlo simulation and the results are reported. The Cramer-Rao lower bound of the scheme is also derived and reported.