Offset-free energy-optimal model predictive control for point-to-point motions with high positioning accuracy

This paper discusses Offset-free Energy-optimal Model Predictive Control (offset-free EOMPC) which is a MPC algorithm to realize time-constrained energy-optimal point-to-point motion control with high positioning accuracy for linear time-invariant (LTI) systems. The offset-free EOMPC approach is developed based on our previous research - Energy-optimal Model Predictive Control (EOMPC) - which aims at performing energy-optimal point-to-point motions within a given motion time. A drawback of the EOMPC method is that it cannot achieve high positioning accuracy in the presence of unmodelled disturbances or model-plant mismatch. In order to cope with this problem, a `disturbance model´ strategy is adopted: the system state is augmented with disturbance variables. Based on the `disturbance model´, the disturbances are estimated and the effect of which is cancelled. Numerical validation of the offset-free EOMPC using a model of a linear motor with coulomb friction and cogging disturbances has been performed and the results show that time-constrained energy-optimal point-to-point motion with high positioning accuracy is achieved.