Non linear predictive control algorithm based on Laguerre-Volterra models

This paper proposes a non linear predictive control algorithm based on Laguerre-Volterra model. This latter results from the expansion of discrete Volterra kernels on independent Laguerre basis. This expansion enables to alleviate the Volterra model complexity resulting from the high parameter number. As the yielded Laguerre-Volterra model is linear with respect to parameters, we develop a predictive control strategy using the classical technique that minimizes a quadratic criterion with respect to constraints due to sensors and actuators which are summarized by bounds on control, control increment and output signals.