Enhanced LOFAR Detection Processing for Passive SONAR

Abstract

This study surveys methods for improvements in the processing and visualization of Low Frequency Analysis and Recording (LOFAR) in sonar systems. The surveyed method employs a weighted Fast Fourier Transform (FFT) technique to compute LOFAR (Low Frequency Analysis and Recording) spectrum, with the weighting function inversely proportional to the variance of phase estimates for each frequency bin. The approach leverages the observation that stable line spectra display lower variance in phase estimates when compared to broadband signals and ambient noise. Simulation results substantiate the effectiveness of the proposed technique. The technique is applied to the DeepShip underwater dataset available in the public domain, and the performance of the proposed algorithm is recorded.