Predictor-based self-tuning control (Q794598): Difference between revisions

This paper deals with a single-input single-output stochastic system driven by a digital self-tuning controller. It is required that the controller is automatically tuned also with respect to possible changes of the setpoint (command signal), which may be either a-priori known (program control) or uncertain (servo), or it may be required that the process output follows the output of a reference model (model following). The self-tuning controller may include the feedforward from the measurable external disturbance, if available. Linear finite-memory output predictors updated in real-time appear to be a suitable internal representation of the system in the digital self- tuning controller. In most cases of industrial process control the Incremental-predictors should be preferred to Positional-predictors. Two main arguments for this recommendation are presented. A new time- domain approach to the quadratic-optimum control synthesis for systems described by such predictors is presented. The method leads to algorithms (or explicit formulae in low-order cases) which are numerically robust and therefore suitable for real-time computation using microprocessors with reduced wordlength.