Precomputation of generalized predictive self-tuning controllers

A method for implementing generalized predictive self-tuning controllers is presented which avoids heavy computational requirements. The method makes use of the fact that a general predictive controller using a quadratic function results in a linear control law that can be described by a few parameters. These parameters are computed over the range of interest of the process parameters and a function is used to obtain an approximation to the real controller parameters. The controllers' parameters are given for processes which can be modeled by a static gain, a time constant, and an effective dead time, that is, for the majority of processes in industry. >