Adaptive Frequency-Domain Sparsity Control Based on Kullback–Leibler Divergence for Tonal Enhancement in Passive Sonar

Abstract

Tonal components embedded in the radiated noise of underwater targets constitute critical spectral features in passive acoustic measurement systems. To improve the reliability of tonal measurement under colored noise and low-signal-to-noise-ratio (SNR) conditions, this article proposes a robust Kullback–Leibler divergence-guided sparsity-control adaptive line enhancer (KL-SC-ALE). By exploiting the intrinsic frequency-domain sparsity of tonal components, the proposed KL-SC-ALE suppresses wideband background noise. To adapt the sparsity control to varying noise conditions, a KL-divergence criterion is introduced to regulate the sparsity strength by guiding the estimated distribution of the frequency-domain weights toward a prescribed sparse prior. A stabilization mechanism is further incorporated to ensure stable adaptation. Simulation and sea-trial results demonstrate improved stability and reliable tonal measurement in colored-noise and low-SNR environments.