A Non-fragile Approach to H ∞ ILC of Linear Discrete-time Systems with Uncertain Learning Gain

This paper studies the problem of iterative learning control (ILC) with uncertain learning gain. The ILC controller to be designed with additive gain variations resulting from controller implementations. Resorting to 2D Roesser system manner and Lyapunov method, a new condition is developed to guarantee the desired stability with the prescribed H ∞ performance. A sufficient conditions for the ILC controller with additive gain variations design in terms of solutions to a set of linear matrix inequalities (LMIs). To overcome the coupling nonlinearities between system parameters and Lyapunov variable, an effective strategy is employed. A notion of structured vertex separator is utilized to approach the problem of additive gain variations. A numerical example is given to illustrate the effect of the proposed method.