On Stable Inversion for Linear Systems
Résumé
System inversion is commonly used in many control applications, such as trajectory planning, tracking, optimal control, or feedforward control. In this paper, we introduce a generic stable inversion procedure for linear time invariant systems, using lumped and distributed delays. The solution is BIBO-stable and causal in the time domain. This solution is an approximated inverse of a plant: After a finite time, the error on this approximation will be identically zero. Constructive algebraic procedures and simulation results are carried out.