Parallel distributed identification for block structure nonlinear system

Abstract

The problem of identifying one block structure nonlinear system, consisting of four linear blocks and one static nonlinear block, is considered here. Under the assumptions that static nonlinear is a polynomial form, the block structure nonlinear system can be transformed into one state space form. Then this nonlinear system identification is formulated into identifying the state space form for these four linear blocks. After transforming the formal state space form, we obtain a polynomial nonlinear state space equation. Then the parallel distributed algorithm is used to identify each system matrix, while using the relationship between the formal state space equation and its transformed state space equation. Finally the simulation example is used to prove the efficiency of our algorithm.