Mobile Node Localization in Underwater Wireless Networks

Localizing mobile nodes in underwater networks is a highly challenging endeavor due to range errors caused by the mobility and uncertainty of sound speed. We propose a novel localization approach, which incorporates time alignment and range bending compensation to meet this challenge. Given the lengthy and varied propagation delays to different anchor nodes, we use a Kalman filter to align different time instants and locations as a mobile node receives timestamps from different anchors during a localization period. Based on Snell's law, the ray tracing theory is applied to compensate for sound speed variations. These two steps minimize the errors caused by mobility and sound speed uncertainty. A penalty convex concave procedure approach is also applied to accurately solve a nonconvex optimization problem to minimize localization errors. Deep sea trial results show that the final localization error for the mobile node is only 1.44 m (with the differential GPS as the true-value reference), marking a substantial improvement over existing state-of-the-art solutions.