An Iterative Learning Control with Alignment Initial Condition for a Class of Nonlinear Systems

Iterative learning control (ILC) is effective for nonlinear systems to track repetitive trajectories. However, identical initial condition is usually assumed for perfect tracking. This assumption can be relaxed for a class of nonlinear systems that has a unique steady-state response for an input. A contraction mapping ILC with selective learning is proposed to achieve perfect tracking under the alignment initial condition, such that the end state of the preceding iteration becomes the initial state of the current iteration. The input updating law and the sufficient condition of monotonic convergence of the input sequence are given. The tracking performance is illustrated by a simulated example.