Internal model control using neural networks

This paper deals with the internal model control of non linear dynamic systems based on artificial neural networks. The proposed control scheme is based on the neural network model and the inverse model of the process. These models are determined off line using input output data. The back propagation algorithm is used to train the neural networks. The neural network internal model control is applied to a level control of a laboratory process. The performances of the proposed controller are compared to a standard PI controller. The combination of the PI controller and an anticipation action given by the inverse model of the process has been also tested on the experimental process.