Sea Clutter Suppression Based on Complex-Valued Neural Networks Optimized by PSD

Abstract

Sea clutter suppression plays an important role in improving the estimation accuracy of the motion parameters of moving ships. Based on the chaotic characteristics of sea clutter, a novel sea clutter suppression method based on complex-valued neural networks optimized by power spectral density is proposed. The complex-valued neural networks helped reduce the phase prediction error of sea clutter, so that the sea clutter prediction accuracy was significantly improved compared with that of real-valued neural networks. The power spectral density function was added to the loss function of the complex-valued neural networks to optimize the training of the networks, and the prediction accuracy was further improved. The sea clutter could be effectively suppressed by cancellation and the signal-to-clutter ratio of the echo was improved. Experimental results based on measured sea clutter data validated the proposed method.