Chaos based semi blind identification of nonlinear systems

Summary form only given, as follows. In this paper, we address the problem of parameter estimation of a class of nonlinear system described by a finite order Volterra kernel with a known embedding dimension. In particular, we derive the theoretical lower bound estimation performance of the nonlinear systems driven by chaos signal. Numerical simulations are performed to confirm the theoretical performance obtained. Furthermore, a robust expectation maximization (EM) based estimation algorithm is designed to adaptively estimate the parameters of the nonlinear system. The estimation performance of the proposed algorithm is evaluated using computer simulations and shown to be better than the conventional nonlinear system identification algorithms.