Parametric identification of Box-Jenkins structured closed-loop Hammerstein systems using gravitational search algorithm

Abstract

This paper presents a Gravitational Search Algorithm (GSA) based accurate parametric identification approach to identify a closed-loop Box-Jenkins structured Hammerstein model. The main objective of the employed algorithm is to estimate the parameters associated with the model by optimizing the fitness function which is the output mean square error (MSE) in this work. Efficient identification of a generalized practical closed-loop Hammerstein model has been achieved from the outcomes of the simulation studies. Convergence curves of the output MSE and the parameters show the consistency of the performance of the proposed GSA based approach. Effective identification in the presence of colour noise shows the robustness of the GSA based system identification problem.