Hierarchical Blind Modulation Classification for Underwater Acoustic Communication Signal via Cyclostationary and Maximal Likelihood Analysis

Abstract

Modulation detection is important to many communication and electronic warfare applications. In this paper, we propose a hierarchical blind modulation classification algorithm for underwater acoustic communication signals. Due to the complex environment encountered by underwater acoustic communication, the cyclostationary features of such signals are significantly different from those of their radio frequency (RF) counterparts. In our previous work, we have applied second order cyclostationary analysis to underwater acoustic communication signals and successfully distinguished BPSK modulation and QPSK (and higher PSK and QAM) modulation types. To further distinguish different PSK and QAM modulations, higher order cumulants such as fourth order second conjugate cumulants are normally exploited in RF communication signals. However, such high order cumulant do not perform well in classifying modulation types of underwater acoustic communication signals, due to the lack of enough data symbols in short length of stable underwater channel. In this paper, we propose a hierarchical blind modulation classification scheme that can classify BPSK, QPSK and 16QAM modulation schemes for underwater acoustic communication signals. Specifically, the second order cyclostationary features are exploited first to classify signal into BPSK and non-BPSK types. Next, a maximal likelihood detection algorithm is performed to further distinguish QPSK from 16QAM. Experimental data collected at sea is used to confirm the effectiveness of the proposed algorithm. The proposed method does not assume any a priori knowledge of the target signal, making it a truly blind modulation classifier.