Time-varying neural networks based indirect adaptive ILC for discrete-time varying nonlinear systems

By using the iterative learning projection algorithm with dead-zone for training time-varying weights, a time-varying neural networks (TVNNs) based indirect adaptive iterative learning control (I-AILC) scheme is presented for a class of uncertain discrete-time varying nonlinear systems with unknown control gain sign. The control singularity has been overcome through a modification of the control gain estimation which can be bounded away from zero. The proposed TVNNs-based I-AILC doesn´t require the strict initial resetting condition and the reference trajectory can vary along the iteration axis. Theoretical analysis proves the boundedness of all signals of the closed-loop system and convergence of the tracking error to a bounded region. The numerical results presented verify effectiveness of the proposed method.