Algorithm of blind source separation based on maximum pseudo-SNR

A novel instantaneous linear mixture blind source separation algorithm is proposed in view of that it is difficult to separate all kinds of signals mixed seriously in EW environment. The pseudo-signal to noise ratio function constructed by the covariance matrix of source signals and noises is built as an objective function, which is based on the theory that signal to noise ratio (SNR) is maximal when statistically independent source signals are completely separated. And the source signals are replaced by the processing of estimate signals with the unit symmetrical moving weight vector. Then the objective function is expressed by the generalized eigenvalues (GE) problem, and an unmixing matrix is achieved by solving the GE problem without any iterative. Compared with the typical information-theoretical approaches, this new optimal blind separation algorithm is a global with low computational complexity. Computer simulation demonstrates that this algorithm can separate the radar signals and communication signals effectively in the complicated environment.