Volterra Series identification Based on State Transition Algorithm with Orthogonal Transformation

Abstract

A Volterra kernel identification method based on state transition algorithm with orthogonal transformation (called OTSTA) was proposed to solve the hard problem in identifying Volterra kernels of nonlinear systems. Firstly, the population with chaotic sequences was initialized by using chaotic strategy. Then the orthogonal transformation was used to finish the mutation operator of the selected individual. OTSTA was used on the identification of Volterra series, and compared with particle swarm optimization (called PSO) and state transition algorithm (STA). The simulation results showed that OTSTA has better identification precision and convergence than PSO and STA under non-noise interference. And when there is noise, the identification precision, convergence and anti-interference of OTSTA are also superior to PSO and STA.