This dissertation presents an approach to the H ? control of linear input-delay systems. First, a novel asymptotic state predictor is introduced for dead-time systems to compensate for the delay in the input. Next, the idea of sequential predictors () is suggested for unstable systems with a long time delay. is based on series of coupled asymptotic predictors, each of which is responsible for prediction of one small portion of the delay, such that the predictors collectively forecast the state for a long time. Moreover, H ? performance of prediction error can be improved by increasing the number of predictors. Consequently, an H ? controller is designed for the standard H ? problem of dead time systems, using sequential predictors. More importantly, method is extended to robust H ? control in presence of uncertainties. Examples illustrate the superiority of to the robust H ? control of input time delay systems. Key words: Input time delay systems, Robust H ? control, Sequential Predictors
