Long-Range Predictive Control Design Based on CARMA Model

Many different long-range predictive controllers have been proposed and investigated for almost two decades. The method presented in this paper is based on a stochastic optimal design which allows the closed-loop characteristics to be explicitly identified. The CARMA model is used instead of the more commonly used incremental parametric model. The prediction horizon or cost function index can be sized according to overall system damping requirement, and even dynamically unfavourable plants can be stabilized by adequate selection of this index. The conclusion has been borne out by extensive simulation tests, and representative results are presented in the paper.