Adaptive predictive control based on the ARX model with reduced complexity

In this article, we propose a new adaptive predictive control algorithm based on the ARX-Laguerre model. The vector representation is exploited for the calculation of the predictor which has been combined with the adaptive identification of the Fourier coefficients of the model to exploit it in the development of the adaptive predictive control such that the Fourier coefficients are updated at each sampling time. The convexity of the convergence criterion is shown driving an optimal control. An experimental validation on a 2 nd order electrical circuit focusing on the performance of the proposed algorithm against parametric uncertainties.