Information-theoretic compressive measurement for frequency hopping pattern recognition

In this paper, we propose an information-theoretic compressive measurement scheme for the frequency hopping pattern recognition of frequency-hopping spread spectrum (FHSS) signals. We model the FHSS signals as a Gaussian mixture distribution, where each component is corresponding to one candidate frequency hopping channel. The compressive measurement scheme is then designed to maximize the Shannon mutual information between the compressive measurements and the frequency hopping pattern to be recognized. The approximated mutual information gradient with respect to the compressive measurement matrix is used in a gradient-based approach to search for the optimal compressive measurement scheme. Simulation results demonstrate that the proposed information-theoretic compressive measurement scheme can significantly improve the recognition performance of the frequency hopping pattern than random projections typically used in the classical compressive sensing theory.