Generalized PID Learning Control

Abstract The class of single-input, single-output, minimum phase, nonlinear, time-invariant systems with uncertain output-dependent nonlinearities, uncertain parameters and known relative degree one or two is considered. The output regulation problem by output error feedback in the presence of periodic reference and/or disturbance signals with known common period is addressed and solved by following recent developments in the iterative learning control theory. A simple iterative learning control algorithm is derived: it can be interpreted as a generalization of the classical PID control which, for such a class of systems, solves the problem when reference and disturbance signals are constant. An application to autonomous vehicle control is presented to illustrate the potentiality of the proposed technique.