The aim of this paper is to develop a robust method to identify and control a MIMO nonlinear process in real time. Contrary to some approaches like neural networks technique, the identification of the process is performed by adjusting a linear autoregressive model, which is defined at each time around the operating point of the process. This approach is an efficient way to find quickly and at low computational cost, an approximate model of the plant. It is necessary that a fast adaptation of this model takes into account the parameters changes and the nonlinearities or the process. This is also useful to reduce the effect of disturbances. The control law is then deduced from this model, such that errors between the outputs of the process and the desired outputs are reduced with a specified dynamics. This adaptive controller is applied to a manipulator whose aim is to control a laser beam.
Identification, Least squares, Control Theory, Prediction, Disturbance
All articles published by Studies in Informatics and Control (SIC) are published in accordance with the terms and conditions of the Creative Commons Attribution License (CC BY-NC-SA 4.0).