Robust predictor for uncertain dead time systems

This paper presents a method to stabilize uncertain dead time system based on new robust state predictor. The proposed predictor consists of a state delayed observer. Controller gain and predictor parameters are calculated by solving a nonlinear matrix inequality. More importantly, this method is extended to dead time system with a long time delay or significant uncertainty using sequential sub-predictors (SSP). This predictor composed of collection of sub predictors that each of them predicts the state for a small part of a long time delay. The number of predictors can increase attending the unstability, delay value or uncertainty where the stability condition is satisfied. Examples illustrate the capability of this method.