Design of predictor-based controllers for input-delay systems

In this paper, a predictor-based controller is proposed for both linear and nonlinear systems with constant input delay. The central idea is that the control signal is constructed based on the prediction of the state in the future rather than on the current state. Its performance depends on the estimation accuracy. The controller developed for linear systems is mathematically equivalent to Artstein model reduction or finite spectrum assignment (FSA). For nonlinear systems in companion form or lower triangular form, stability can be guaranteed provided that the prediction errors are bounded.