Abstract

Traditional adaptive inverse control systems are often used FIR filters, such filters will make the system costing increasing as long training sequence is needed and the slow convergence occurs at the same time. Thus, it is unable to adapt the requirements of the real-time control systems. In this paper, an IIR filter is designed using an improved adaptive inertia weight particle swarm optimization (PSO) algorithm to reduce the computing costs and improve the convergence speed of the filter weight. Also, this PSO-IIR filter is used in discrete adaptive inverse system for online system identification and online inverse controller adjusting. Simulation results show that, for discrete uncertain systems, this method can make the system achieve control objectives quickly and effectively.

Paper Details