Neuro-adaptive control for general nonlinear discrete-time systems

The authors propose a type of adaptive control with neural networks, called neuroadaptive control. This neural control is based on a framework of model reference adaptive control (MRAC). MRAC is briefly outlined, and an inverse representation of a discrete-time general nonlinear system is derived. A design is presented of the neuroadaptive control system, which estimates a function describing the inverse representation and generates an input so that the output of the controlled object tracks the command input. A simple numerical simulation illustrates the performance of the neuroadaptive control.<
>