Robust physical parameter estimation by using MRAC based on the inverse normalized delta operator

The conventional identification algorithm based on the shift operator lacks a physical meaning. Thus, we proposed previously the new method that estimates directly physical parameters by unifying MRAC and the delta form. Since this proposed method needs to calculate the difference on the error between the plant and model output, the algorithm cannot be applied to the high-order systems due to measurement noise. Thus, in this paper we propose the method based on the integral operation by using the inverse normalized delta operator. As an illustration, second-order systems are considered. Simulation results show that this proposed method can accurately estimate the plant parameters in comparison with the conventional RLS under measurement noise. Since the amount of the calculation on the proposed method is very low, the algorithm can be calculated in real-time by using fixed-point microprocessors with short word-length.