Underwater Node Localization Using Modified Hough Transform and Least Square Method

Underwater wireless sensor networks (UWSNs) attract increasing interests for many applications. The positions of sensor nodes are meaningful information for data collection and fusion. Since the anchor nodes are the references for the whole network localization, the estimation of node positions determines the accuracy of localization. Conventionally, the anchor node is localized with the survey of a radius approximately equal to the depth with collecting round trip acoustic ranges and GPS outputs using nonlinear least squares estimation approach. Modified Hough transform (MHT) was proposed recently for acoustic localization of subsea transponders. The approach utilized the range measurements and GPS data to achieve an intersection search with the advantage of not requiring an initial estimation of the transponder position and providing high precision of estimation. The main idea of MHT is an accumulating process: each point of the survey ship can draw a circle with the radius equal to slant range projection in the horizontal plane. With the whole trace of accumulating, the circles will intersect to one point in the plane. Thus the peak of accumulating defines the estimation of transponder position. Shallow water and deepsea experiments were conducted and the actual data was processing in this paper to compare the performance of node localization with different methods. The results show that the node position can be estimated with the precision to centimetres in shallow water and meters in deep sea. In this way, MHT algorithm can be widely used and have great value in ocean engineering.