Modulation classification of mixed signals using independent component analysis

Modulation classification (MC) of a signal is one of the major tasks of an intelligent receiver in various civilian and military applications. For mixed signals, it is a big challenge to recognize their modulation types. In this paper, a novel MC algorithm based on independent component analysis (ICA) is proposed. The proposed algorithm can separate mixed signals and identify modulation types for demodulation simultaneously under block fading channel. The design of the algorithm essentially involves two steps, namely, the separation of mixed signals through ICA algorithm and MC for the separated signals. Depending on ICA algorithm chosen in the first step, statistically independent signals are separated in block fading channel. Regarding the second step, higher-order cumulants are chosen as a promising approach for MC of the separated signals. Through the algorithm, the problem of the MC of the mixed signals can be solved and the mixed signals can be demodulated. Furthermore, simulation results are provided to verify the performance and robustness of the algorithm.