Application of artificial bee colony algorithm in blind source separation of chaotic signals

Chaotic signal is nonlinear and non-Gaussian signal. Features of non-periodic and wide band spectrum make it quite difficult to separate mixture of chaotic signals blindly. To address this issue, a new blind source separation method based on artificial bee colony optimizer is proposed. Additionally, a parameterized representation of orthogonal matrices through principal rotation is adopted to reduce the complexity of the optimization problem. Simulation results show that the proposed method can accurately separate the mixture of chaotic signals within a few tens of iterations, and its overall performance is better than traditional independent component analysis approaches.