On-line neural network control applications

This work aims to present an exploration of potential applications of on-line trained neural networks. The work consists of computer simulation results from implementing an internal model neural network-based control scheme, and a model reference neural network-based adaptive control scheme. The design of the on-line learning algorithm need to adapt the neural networks weights is based on the theory of continuous-time variable structure control systems. The main aspect in the formulated internal mode control scheme using on-line trained neural networks is the possibility of obtaining simultaneous on-line estimations of the forward transfer operator as well as the inverse transfer operator of an unknown dynamic plant that allows the output variable of the plant to be regulated according to the specifications provided by a reference input signal.