Mismatch-Robust Underwater Acoustic Localization Using A Differentiable Modular Forward Model
This addresses localization accuracy issues in underwater acoustics for applications like marine navigation, but it is incremental as it builds on existing neural network and optimization methods.
The paper tackles underwater acoustic localization under environmental mismatch by using a pre-trained neural network for wave propagation in a gradient-based optimization framework, achieving effective source location estimation with conditions for robustness.
In this paper, we study the underwater acoustic localization in the presence of environmental mismatch. Especially, we exploit a pre-trained neural network for the acoustic wave propagation in a gradient-based optimization framework to estimate the source location. To alleviate the effect of mismatch between the training data and the test data, we simultaneously optimize over the network weights at the inference time, and provide conditions under which this method is effective. Moreover, we introduce a physics-inspired modularity in the forward model that enables us to learn the path lengths of the multipath structure in an end-to-end training manner without access to the specific path labels. We investigate the validity of the assumptions in a simple yet illustrative environment model.