Adaptive predictive controllers based on orthonormal series representation

This work presents and compares a set of active and passive adaptive predictive controllers based on orthonormal series representations. The controllers are designed according to the receding horizon principle, where the parameters of the predictive model are considered as random variables. This general framework allows the design of active and passive adaptive controllers, according to the assumptions made over the prediction of the posterior densities of the random variables available in the system. Several simulated examples illustrate the characteristics of each algorithm. The performance improvement by active adaptive control is demonstrated.