Identification of Prandtl-Ishlinskii hysteresis models using modified particle swarm optimization

Abstract

A modified particle swarm optimization algorithm (MPSO) is proposed in this paper. The algorithm is implemented to identify the parameters of the hysteresis nonlinearity, which is described by a modified Prandtl-Ishlinskii model. This new algorithm redefines the global best position and personal best position in the traditional PSO algorithm by an effective informed strategy, in order to balance the exploitation and exploration of the algorithm. Furthermore, a mutation operator is employed to increase the diversity of the particles and prevent premature convergence. Experiments have been conducted to verify the effectiveness of the proposed method. The comparisons with other variants of the PSO demonstrate that the identification of hysteresis based on the MPSO is effective and feasible. © Springer-Verlag Berlin Heidelberg 2012.