Robust identification of first-order plus dead-time model from step response

In this paper, a simple yet robust identification method for a linear monot onic process, derived from a step test, is proposed. New linear regression equations are derived, from which the parameters of a first-order plus dead -time model can be obtained directly. No iterations in calculation are need ed. The proposed method outperforms the existing estimation methods that us e step-test responses. The estimation error is smaller in both the time dom ain and the frequency domain. Furthermore, the method is robust in the pres ence of large amounts of measurement noise. The effectiveness of the identi fication method has been demonstrated through a number of simulation exampl es and a real-time test. (C) 1999 Elsevier Science Ltd. All sights reserved .