Discrete modified Smith predictor based on optimal control method for a plant with an integrator

In this paper, a discrete modified Smith predictor based on optimal control design method is proposed. A plant with long dead-time is controlled by means of a predicted-state feedback technique and a modified Smith predictor composed of a plant predictor and an observer. In this method, states of a plant are obtained by an observer because the states are not measured directly in practice. The advantages of the proposed method are that a steady-state error caused by a disturbance can be eliminated and that an unstable plant can be made stable unlike conventional methods such as Smith predictor and Internal Model Control. It also achieves good robustness to modeling errors by adjusting weights of LQI-scheme. In simulation studies, it is shown that the proposed method is effective for these plants. Furthermore, by applying the proposed method to a pneumatic actuator, the effectiveness of the method is confirmed.