Title :
Parameterized Convergence Bounds for Volterra Series Expansion of NARX Models
Author :
Zhenlong Xiao ; XingJian Jing ; Li Cheng
Author_Institution :
Dept. of Mech. Eng., Hong Kong Polytech. Univ., Hong Kong, China
Abstract :
Evaluation of the convergence bound in the frequency domain for Volterra series expansion of nonlinear systems described by NARX models is studied. This provides new convergence criteria under which the nonlinear system of interest has a convergent Volterra series expansion, and the new criteria are expressed explicitly in terms of the input magnitude, model parameters, and frequency variable. The new convergence criteria are firstly developed for harmonic inputs and then extended to multi-tone and general input cases. Based on the theoretical analysis, a general procedure for calculating the convergence bound is provided. The results provide a fundamental basis for nonlinear signal processing using the Volterra series theory.
Keywords :
Volterra series; autoregressive processes; frequency-domain analysis; nonlinear filters; NARX model; Volterra series theory; convergent Volterra series expansion; frequency domain; multitone case; nonlinear autoregressive with exogenous input mode; nonlinear signal processing; nonlinear system; parameterized convergence bound; vergence criteria; Analytical models; Computational modeling; Convergence; Educational institutions; Frequency-domain analysis; Harmonic analysis; Nonlinear systems; Convergence criteria; NARX models; Volterra series; frequency domain;
Journal_Title :
Signal Processing, IEEE Transactions on
DOI :
10.1109/TSP.2013.2277838