Si̇stem Ki̇mli̇klendi̇rme İçi̇n Geli̇şti̇ri̇lmi̇ş Bi̇r Wiener Model

Abstract

Wiener block structure is formed by cascade of linear and nonlinear models. A novel and improved Wiener model structure for system identification area is proposed in this study. In proposed Wiener model, Finite Impulse Response (FIR) model is used as linear part and Soft Switching based Hybrid (SSH) model is used as nonlinear part. The SSH structure consists of a Second Order Volterra (SOV) nonlinear model, a Memoryless Polynomial (MP) nonlinear model, and a soft-switching part through a Neuro-Fuzzy (NF) network. In simulation studies, different types systems are identified by presented novel model. In addition to the mentioned identified systems, the performance of the improved model is also compared with Volterra model and Wiener models presented in the literature. Simulation results find out the success of the proposed model.