Embedded estimator in predictive feedback control

In this work, a new approach to design predictive feedback control for SISO systems is presented. The proposed formulation relies on the development of a single step predictor based on an autoregressive moving average with external input (ARMAX) model. Although no explicit observer is actually involved in the implementation, this predictor implicitly includes one since the input-output model subsumes an observer. Exploiting this idea the resulting ARMAX model is extended to include extra outputs to improve the quality of the prediction for systems with large time delay and nonmeasurable disturbances. The resulting predictor is used to develop a predictive feedback controller. This new formulation of predictive feedback control includes feedback and feedforward actions. Simulations of two linear systems illustrate the applicability of the control algorithm.