Robust nonlinear predictive controller for multivariable nonlinear systems with different relative degree

Abstract

This paper presents a robust nonlinear predictive controller (RNPC) for a class of multivariable nonlinear systems with different relative degree. The classical nonlinear predictive control (NPC) based on Taylor approximation method is reformulated and a new cost function is proposed to design the controller. The parameters of this controller are obtained using both generalized minimum variance (GMV) and generalized predictive control (GPC). It is shown that an integral action arises naturally in the controller. Then, zero tracking error can be achieved as long as the closed loop system is stable. The design procedure is illustrated by designing a pilot for a 0.25 kW permanent magnet synchronous motor (PMSM). Validity of the proposed GMV control strategy was experimentally tested on a dSPACE DS1104 Board. Experimental results have demonstrated the achievement of good speed tracking performances and shown that the regulation of the d-axis component of the armature current is guaranteed.