Tracking periodic parametric references using model predictive control

This paper presents a model predictive controller for tracking periodic parametric reference curves. The controller is formulated in a single layer so that the time-parameterization of the reference curve, the trajectory planning, and the trajectory tracking tasks are solved in a single optimization problem which is computed at each sampling time. The auxiliary state and input trajectories are introduced into the optimization problem as decision variables, which become the planned trajectory to be effectively followed. By design, the closed-loop system guarantees of recursive feasibility of the optimization problem while being able to handle arbitrary changes of reference during execution time. Such properties are discussed in detail, and new results on the asymptotic average performance are presented. Simulation results show the benefits of the proposed strategy.

Automated summary

This paper presents a model predictive controller that can track periodic parametric reference curves. The controller solves the time-parameterization, trajectory planning, and trajectory tracking tasks in a single optimization problem at each sampling time. The optimization problem includes auxiliary state and input trajectories as decision variables, which are effectively followed as planned trajectories. The closed-loop system is designed to guarantee the recursive feasibility of the optimization problem and handle arbitrary changes of reference during execution time. The paper discusses these properties in detail and presents new results on asymptotic average performance. Simulation results demonstrate the benefits of the proposed strategy.