En-NBOR: An Enhanced Newton-Raphson-Based Optimizer for Underwater Acoustic Signal Reconstruction

Underwater acoustic signals are frequently subject to noise interference during propagation, leading to signal attenuation and distortion, which present significant challenges for accurate reconstruction. This paper proposes an enhanced Newton-Raphson-Based Optimizer (NRBO) for underwater acoustic signal reconstruction, integrating three key innovations to improve accuracy and robustness. First, a Frequency-Domain Adaptive Newton-Raphson Search Rule enables adaptive adjustment of the search step size based on spectral characteristics, allowing detailed exploration in low-frequency ranges rich in signal content while quickly bypassing noisy high-frequency components, thereby enhancing reconstruction accuracy. Second, a Multi-Scale Feature Extraction with NRBO decomposes the acoustic signal into multiple scales, applies denoising and reconstruction independently at each scale, and then merges the results to preserve multi-level signal characteristics, improving overall signal quality. Finally, a Noise-Adaptive Objective Function dynamically adjusts weights according to noise levels in underwater environments, enhancing the algorithm’s denoising capability across diverse noise conditions. Experimental results show that the proposed method consistently outperforms traditional approaches in terms of signal-to-noise ratio (SNR), reconstruction error, and convergence speed, demonstrating significant improvements in accuracy and robustness. This study provides an efficient and adaptable solution for underwater acoustic signal processing, suitable for various underwater environments.