The adaptive predictive controller design for the problem of input delay based on the Lyapunov approach is discussed in the present paper. The main problem of designing a predictive controller is to determine the values of control gains and to compensate for the effect of input delay for the predictable state. We first propose a novel predictive controller called the adaptive predictive controller and use it in place of the traditional adaptive controller used in the nonlinear control system without any input delay. Second, we adopt a simplex algorithm that can search the optimal timing for a prediction to compensate for the effect of input delay. In the meantime, we use the techniques of neural network predictive model and direct adaptive scheme to design the adaptive predictive controller for some class of nonlinear systems with input delay. To enhance the input-delay systems, stability to the size of input delay, a supervised control scheme and an adaptive predictive controller of unknown nonlinear systems with input delay are proposed. The supervised adaptive control law aims to pull the states back to the pre-specified state region and reduce the overshoot towards an extra stabilizing input. Finally, some computer simulation results are provided to illustrate the proposed methods.
PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART I-JOURNAL OF SYSTEMS AND CONTROL ENGINEERING