Simulation and Analysis of Nonlinear System Identification Using the LMS Volterra Filter

Abstract

Nonlinear system Identification based on Volterra filter are widely used for the nonlinearity identification in various application. A standard algorithm for LMS-Volterra filter for system identification simulation, tested with several convergence criteria is presented in this paper. We analyze the steady-state mean square error (MSE) convergence of the LMS algorithm when random functions are used as reference inputs. In this paper, we make a more precise analysis using the deterministic nature of the reference inputs and their time-variant correlation matrix. Simulations performed under MATLAB show remarkable differences between convergence criteria with various value of the step size. Along with that the least mean squared (LMS) adaptive filtering algorithm may experience uncontrolled parameter drift when its input signal is not persistently exciting, leading to serious consequences when implemented with finite word-length. The second order LMS Volterra filter with variable step size for system identification are analyzed and comparing the theoretical value with experimental value. Copyright © 2009 IFSA.