Study On Adaptive Nonlinear Filters Based On Volterra Series Model And Their Application

In recent years, more and more attention was paid to nonlinear phenomena and there are many nonlinear problems in practical situations. Different from the linear systems, there is no normal model of nonlinear systems. The Volterra series model is a useful nonlinear model and can model a large class of nonlinear systems.Firstly, the Volterra series theory and its development are introduced. The conditions in which nonlinear systems can been modeled by Volterra series model are also introduced.Secondly, after studying modeling nonlinear systems using Volterra series model, the adaptive algorithms are presented. It is also investigated that step factor controls the convergence of the algorithms. A simulation is given which uses the Volterra adaptive model for modeling an unknown nonlinear system and indicates that the method is efficient.In order to sovle the niose's effect to system identification, total least mean square adaptive algorithm is presented. What is more, this paper derives a multistage Volterra filter to solve the coupling problem. And the multistage Volterra filter is shown to be equivalent to the partially decoupled Volterra filter.Lastly, a multichannel adaptive lattice filter is applied to the equalization of nonlinear channel. In digital communication, inter-symbol interference and noise make the channel poor, so it is necessary to use the equalizer. The simulation indicates that the multichannel adaptive lattice filter is efficient.