Receding horizon based trajectory planning and two-degree-of-freedom tracking control for fast sampling constrained systems

This paper presents a trajectory planning method inspired by receding horizon based control - for example model predictive control. To plan a trajectory that is perfectly trackable by an underlying controller, a cost functional with input and state constraints is minimized. The proposed planning method is decoupled from a very fast sampling control system in the 100 kHz range because of the relatively high computational demand. This structure allows to profit from the advantages of model predictive control, such as the handling of constraints, even for very fast sampling systems. Here the state and input trajectories which are naturally available in receding horizon based trajectory planning are used to improve the tracking behavior with the help of a dynamic feed-forward path. This concept of dynamic feed-forward control enables perfect tracking of the planned trajectory even in dynamic operation. Experimental results demonstrate the effectiveness of the proposed planning and control approach for a highly dynamical two-mass system. The proposed scheme is compared to two conventional control techniques.