An Improved Sparsity-Driven MMCC Adaptive Line Spectrum Enhancement Algorithm to Better Ship’s Shaft-Frequency Electric Field Detection

The ship’s electric field signal generated by corrosion is an important source for underwater target detection, of which the shaft-frequency electric field is regarded as a key component and characterized by a distinct line spectrum. Due to the noisy marine environment, it is hard to extract the shaft-frequency at low signal-to-noise ratios, making the enhancement methods necessary. Traditional line enhancement methods rely on LMS-based adaptive line enhancement (ALE), the effectiveness of which decreases in non-Gaussian backgrounds. To address this, an adaptive line enhancement approach based on Maximum Mixed Correntropy Criterion(MMCC) is proposed to improve applicability under non-Gaussian conditions. To further improve the output signal-to-noise ratio, a sparsity-driven MMCC adaptive line enhancement method is introduced, which adopts an inverse tangent function as the sparse penalty function. Both simulated results and sea experimental results show the effectiveness of the proposed algorithm in the improvement on the signal-to-noise ratio gain of the fundamental frequency line spectrum and of the weak harmonic signals. In terms of fundamental frequency enhancement, the proposed method is comparable to the existing sparsity-driven algorithm MMCC-ZA-ALE, achieving 18.6 dB under non-Gaussian background conditions. In harmonic enhancement, the proposed method can improve by up to 7.4 dB, significantly outperforming MMCC-ZA-ALE.