Modulation Classification of Mixed Signals using Fast Independent Component Analysis

摘要

In military and civilian communications, signals are often interfered by hostile jamming or illegal transmission. In these situations, determining the modulation format of mixed signals is a challenging task, which is tackled using a three step algorithm named PFCC (PCA, FICA, Cumulants Based Classification Algorithm) in this paper. In the first step, centering and whitening is conducted using principal component analysis (PCA) to suppress noise. In the second step, mixed signals are separated using fast independent component analysis (FICA), which can transform received signal into components that are maximally independent from each other. In the third step, high-order cumulants (HOC) are calculated to determine the modulation format of each signal. Through extensive simulation, the convergence speed and performance of PFCC are validated. We also notice that the relative power of mixed signals has a big influence on performance.