Inverse model neural network-based control of dynamic systems

A direct inverse neural network-based control scheme is presented and its performance tested by computer simulations using a nonlinear system. The adaptation of the neural network weights involved in the control scheme is carried out by means of a continuous-time variable structure control-based learning algorithm that enables an online approximation of the inverse transfer operator of the unknown system to be obtained. Further research is underway to prove conditions for assuring the stability of the proposed control scheme.