Modulation recognition method for underwater acoustic communication signals based on multi-feature fusion

Aiming at the problem that communication signals in complex underwater acoustic environments are easily affected by noise and multipath propagation, which leads to weakened modulation features and decreased recognition accuracy, this paper studies a denoising enhancement and recognition method for underwater acoustic communication modulation signals. The Sparrow Search Algorithm (SSA) is used to adaptively optimize the number of decomposition modes and the penalty factor of Variational Mode Decomposition (VMD), and the optimized VMD is then employed to decompose the original underwater acoustic communication signals. According to the correlation coefficients between Intrinsic Mode Function (IMF) components and the original signal, the IMF components are classified, and wavelet threshold denoising is combined to complete denoising reconstruction. The reconstructed signals are transformed into time-frequency images by Continuous Wavelet Transform (CWT), and then input into the lightweight ShuffleNetV2 network for the recognition of six types of modulation signals. To address the confusion between Binary Phase Shift Keying (BPSK) and Quadrature Phase Shift Keying (QPSK) signals, the second-harmonic feature of the squared power spectrum is extracted, and a Particle Swarm Optimization-Random Forest (PSO-RF) classifier is adopted for secondary discrimination. Simulation results show that the proposed method can effectively suppress noise interference and improve the recognition performance of modulation signals under complex underwater acoustic channel conditions. The overall recognition accuracy of the six types of modulation signals is improved from 75.60% to 91.80%. The proposed method effectively improves the accuracy and stability of communication signal modulation recognition in complex underwater acoustic environments through denoising enhancement, time-frequency feature extraction, and secondary discrimination based on the second-harmonic feature.