Generalised Parameterisation for MPC

Bilal Khan and Anthony Rossiter


MPC, generalised parameterisation, feasibility


The paper generalises approaches to predictive control based on Laguerre and Kautz functions. It is shown that Laguerre and Kautz are special cases of generalised orthonormal basis functions and thus one can give a more general parameterisation using higher order orthonormal basis functions. Specifically, a simple but efficient algorithm that uses generalised functions to parameterise the degrees of freedom in an optimal predictive control is presented. The efficacy of the proposed parameterisation within existing predictive control algorithms that use a similar strategy, is demonstrated by examples.

Important Links:

Go Back