Nonlinear-disturbance-observer-based predictive control for trajectory tracking of planar motors

Abstract

To improve the trajectory tracking performance of planar motors against disturbances, model predictive position control (MPPC) methods using the non-linear disturbance observer (NDO) are proposed in this study. Based on the single-axis dynamic model with disturbances, a single-axis NDO is designed using an extended state observer approach. The designed NDO is expressed as a third-order non-linear state-space equation in which the position error, velocity error, and lumped disturbance in the single axis are taken as the state variables. Two MPPC methods are developed based on the NDO. In the first MPPC, the disturbance is embedded into the prediction model using the NDO, and a controller is designed to minimise a quadratic cost function, which is established by applying the prediction model with disturbance. The output of the controller is the control action. In the second MPPC, a controller is used to minimise the quadratic cost function, which is built by employing the prediction model without disturbance. The sum of the output of the controller and the compensated disturbance estimated by the NDO is the control action. The comparative experiment is performed on a planar motor system self-developed in the laboratory. The effectiveness of the proposed methods is verified via the experimental results.