A simple implementation of adaptive predictive controllers

The ample availability of low cost microcomputer systems and personal computers has provided the opportunity to implement and use control algorithms previously developed for minicomputers. A class of control algorithms which is particularly suitable for being realized in this way is predictive control, especially because the operator is provided with easy to interpret tuning parameters. In the paper two different predictive control algorithms are briefly introduced. The first method is based on direct computation of the optimal future input sequence based on a least-squares error method, the latter method is based on an iterative method to calculate the future input sequence. Further two different internal models are used; the first method is based on a step response model of the process, while the second method is based on an impulse response model. An important function of the internal model is the prediction of the output of the process during a certain time in the future