Precise Underwater Localization Based On Ocean Current Information

Precise localization is crucial for missions such as safety monitoring,mine hunting, glider navigation, marine mammals tracking, etc. To obtainmeaningful data from autonomous underwater vehicles (AUV) in anunderwater sensor network, we must have precise underwater localizationand tracking algorithms.When navigating in the underwater environment, the gliders areaffected by the ocean currents. Ocean models that are able to provideaccurate and real-time prediction of ocean currents will improve theperformance of glider navigation. In this paper, we also propose a novelapproach to compute a model for ocean currents at higher resolution thanexisting approaches. By focusing on a small area and incorporatingmeasurements from multiple gliders, we are able to perform real-timecomputation of the model, which can be used to improve performance ofunderwater glider navigation in the ocean. Our model uses a lowerresolution, larger scale model to initialize the computation. We have alsodemonstrated incorporating data streams from other observation systemssuch as Wave Radar (WERA) system. Glider navigation performance usingthe proposed ocean currents model is demonstrated in a simulated flow fieldbased on data collected off the coast of Georgia.In most range-based underwater localization methods, measurementsof time of arrivals (TOA) or time difference of arrivals (TDoAs) areconverted into ranges and later incorporated into multi-lateration algorithms.Existing localization algorithms assume the sound speed is constant and thepropagation path is a straight line. However, sound speed varies withtemperature, pressure, and salinity, causing the bend of underwater sound propagation. Therefore, even if perfect underwater time synchronizationand TOA or TDoA measurements can be obtained, the localizationperformance is still degraded by the wrong assumption model.We also propose a novel precise underwater localization (PUL) methodthat addresses the issue of sound speed variation and provides accuratelocalization and tracking results in unknown underwater environment.Effective sound speed is calculated after gathering the pairwise TDoAs,which is incorporated into the localization algorithm. We implementKalman filter in the localization process to improve localizationperformance and provide reliable localization. According to the simulationresults,50%less localization error can be achieved by precise soundpropagation model compared with the one with less accurate propagationmodel.