A novel method of estimating parameter and its application to blind separation of chaotic signals

Abstract

To estimate effectively parameters of nonlinear mapping, a cubature rule is used to approximate the weighted integral of this mapping. In this paper, based on these parameters modeled by a state-space model, a novel parameter estimation is proposed. Blind separation of chaotic signals is a challenging problem. The proposed method is used to solve this problem to achieve the effective reconstruction of chaotic signals. Simulation results indicate that the proposed method has a faster convergence speed and a higher numerical accuracy, and can effectively separate original chaotic signals.