Adaptive iterative learning control for nonlinear uncertain systems with both state and input constraints

This work proposes a new adaptive iterative learning control (AILC) scheme for nonlinear systems with both state and input constraints, where the time-varying parametric uncertainties, external disturbances, and random initial errors are also considered together. The proposed AILC consists of a learning control law and two fully projected parameter learning laws. By incorporating a barrier composite energy function into the learning control law and using a projection mechanism for the parameter learning laws, the proposed AILC can flexibly and actively manipulate the states and inputs of the system into their pre-specified and constrained ranges, respectively. It is theoretically shown that the asymptotic and pointwise convergence properties are guaranteed without violating any state and input constraints. The validity of the proposed AILC scheme is further verified with a practical train operation system.