Robustness Analysis of Frequency-Difference Methods for Shallow-Water Source Localization Under Water Depth Mismatch

This study investigates the robustness of underwater source localization methods, specifically focusing on frequency-difference-based approaches: Matched Autoproduct Processing (MAP), Phase-Only Matched Autoproduct Processing (POMAP), and their cross-frequency coherent counterparts, CMAP and POCMAP. The research examines how these methods perform under varying water depth mismatch conditions. Through numerical simulations conducted in a shallow-water environment, the performance of these methods is systematically evaluated against the conventional Matched Field Processing (MFP). The results indicate that frequency-difference methods significantly outperform conventional MFP, resulting in lower localization errors in both range and depth. In particular, MAP and CMAP achieved a reduction in range root mean square error (RMSE) of approximately 70% relative to conventional MFP. Additionally, the frequency-difference methods achieved over a 56.9% success rate, while conventional MFP only demonstrated a 15.7% success rate when there was a 10% mismatch in water depth. These results demonstrate the potential of frequency-difference methods to provide more reliable source localization in scenarios with uncertain environmental parameters, thereby enhancing the robustness and practicality of underwater acoustic applications.