Fuzzy c-regression models based on Euclidean particle swarm optimization in noisy environment

This paper addresses the effectiveness of fuzzy c-regression models algorithm and Euclidean particle swarm optimization to nonlinear system identification in a noisy environment. The fuzzy c-regression models (FCRM) clustering algorithm is sensitive to initialization that leads to converge to a local minimum of the objective function. In addition, The particle swarm optimization can be easily trapped in local optima and premature convergence. In order to overcome these problems, the Euclidean particle swarm optimization is proposed to optimize the initial states of FCRM algorithm. Thereafter, weighted recursive least squares is employed to fine tune parameters of the obtained fuzzy model. Finally, the proposed approach is tested by studying a nonlinear modeling problems to verify the identification performance.