Matched Field Source Localization as A Multiple Hypothesis Tracking Problem

Matched field processing (MFP) is a classic approach to locate underwater acoustic source where source level is relatively high without environment mismatch. However, when we wish to track a source radiating weak signal, the performance of MFP degrades as we cannot find the right peak because of the multi-modal structure of the output ambiguity plane. In this paper, we propose a method making full use of the ambiguity plane of MFP to track a point target in low signal-to-noise ratio (SNR) region by employing multiple hypothesis tracking (MHT). The simulation results show that the approach proposed in this paper outperforms the traditional MFP in the condition of low SNR.