Modulation Recognition of Underwater Acoustic Communication Signals Based on Joint Feature Extraction

Abstract

Recently, the research on modulation identification of underwater acoustic (UWA) communication signals has attracted much attention. It is a challenge to identify the modulation of underwater acoustic communication signals in the environment of low signal-to-noise ratio (SNR), multipath interference, frequency and phase drift. In this paper, a joint feature extraction algorithm is proposed which combines the method of cyclic spectrum with high-order cumulant to perform the modulation recognition of UWA communication. In the algorithm, the higher-order cumulant features of the communication signals and their counterparts after the square transformation are used firstly for the intra-class modulation recognition of phase shift keying (PSK) and frequency shift keying (FSK), and the inter-class modulation recognition of PSK. Then the cross-section parameter of spectral correlation function of signals are used for inter-class recognition of FSK modulation. Simulation results show that BPSK, QPSK, 2FSK and 4FSK underwater acoustic communication signals can be identified at rate better than 90% by using the joint modulation recognition algorithm with a signal-to-noise ratio of -2dB.