A novel design of iterative learning control with feedback structure

The purpose of this work is to attain a unified design framework of PID-like ILC with pure feedback structure for a class of nonlinear repetitive NARMA systems. To serving the controller design and analysis, the nonlinear NARMA system is transferred into a linear data model. Then, we suppose that there exists a desired nonlinear controller such that the system output tracks the desired output exactly. By using the mean-values theorem, two time-dynamical linearization methods are proposed for the ideal nonlinear controller and a CFDL-ILC and a PFDL-ILC are presented, respectively. Comparatively, the PFDL-ILC can use more tracking errors of previous time instants of the same iteration and thus may achieve a better control performance. However, the trade off is that the computation may become more complex. The proposed approaches are data-driven and no process model is required for the controller design and analysis. The availability of the proposed approaches is further confirmed by simulation results.