Blind Source Separation of Nonlinear Mixtures Based on Functional Link Artificial Neural Networks and Differential Evolution Algorithm

In this paper, a post nonlinear blind sources separation method is proposed. The demixing system of the post nonlinear mixtures is modeled using a functional link artificial neural network whose weights can be determined under the criterion of independence of its outputs. A criterion of independence based on higher order statistics is used to measure the statistical dependence of the outputs of the demixing system, and the differential evolution algorithm is utilized to minimize the criterion. The proposed method takes advantage of less learning parameters, high learning convergence rate of parameters, nonlinear approximation capability of the functional link artificial neural network, and few easily chosen control parameters, global optimization capability of the differential evolution algorithm. Compared to conventional post nonlinear blind sources separation approaches, the proposed approach for post-nonlinear blind source separation is characterized by less computational load, high convergence rate, high accuracy and robustness. Simulation results show that the proposed approach is capable of separating independent sources from their post-nonlinear mixtures.