Learning Feed Forward Control for High Performance of Tracking Error

The ILC subject is to improve transient responses for repetitive systems .In traditional control, a given repetitive system will produce the same cycle-after-cycle path error for a given trajectory, since the error and the command will be the same. The ILC algorithms try to improve this phenomenon by changing the input control of the next cycle, based on the error of the previous cycle. This technique has the advantage that it can be used without prior knowledge of the system to be controlled. In this article a comparative study will be done , including the previous cycle learning configuration, and current cycle learning configuration. The λ-norm is adopted as the topological measure in the demonstration of the asymptotic stability of this technique of control , when the iteration number tends to infinity. The theoretical are illustrated by simulation. The results of simulations prove clearly the efficiency of the control by iterative learning. Keywords - Iterative learning control, controlled system, previous iteration, λ-norm, learning filter