Weighted Moments Based Identification of Continuous-Time Systems

摘要

In this paper we present an algorithm for continuous-time model identification from sample data using the weighted power moments of the output signal of a linear, time-invariant system. While most of the latest methods used in identification utilize a discrete-time model, the moments method is an alternative approach to directly identify a continuous-time model from discrete-time data. The method defines a set of relationships between the power series coefficients of a stable transfer function and the power moments of the output signal of this system. Based on these relations, an algorithm for off-line parameter identification is developed. The method is applied to identify the parameters of a real experimental platform.