Robust learning control algorithms for uncertain time-varying systems

The trajectory tracking problem of uncertain time-varying systems is addressed, where the same tasks are performed repeatedly within a finite duration of time, or periodic references are followed over an infinite interval. Through the Lyapunov-like synthesis, two robust learning control algorithms are developed based on the control tasks, and their stability and convergence results are established. Both algorithms can compensate for the shortcoming when either one is applied separately. The robust control component guarantees all the variables in the closed-loop to be bounded, while the learning control component ensures that the tracking error converges to zero. Numerical results are presented to demonstrate effectiveness of the proposed learning algorithms.