Multi-Target Sonar Localization Using an Optimal Mass Transport Framework

This article addresses the problem of multi-target localization from acoustic time delay measurements in a 3-D underwater environment, taking into account the depth-dependent sound speed of the underwater wave propagation. Specifically, we formulate the problem as an optimal mass transport (OMT) optimization over a grid of potential candidate positions. By exploiting an iterative refinement approach allowing for off-grid effects, the method allows for robust and accurate positioning of unlabeled underwater target echoes. Using numerical experiments, we illustrate the preferable performance of the presented method as compared to current approaches.