A Fast and Efficient Parameter Estimation Algorithm for Generalized Output Error Models

Abstract

One kind of the colored noise interference systems is generalized output error model (OEARMA). This paper presents a two-stage recursive least squares algorithm for OEARMA. Aiming at the OEARMA, this paper puts forward a two-stage recursive least squares algorithm. The basic idea of the algorithm is to combinie the auxiliary model identification idea and the decomposition technique to decompose a system into two subsystems. Each subsystem contains a parameter vector. With auxiliary model-based recursive extended least squares theory, an unknown intermediate variable output instead of the auxiliary model identification model vector, instead of unmeasurable noise terms in the information vector with the estimated residuals, which can use recursive identification idea to estimated all the parameters of the system, the algorithm has a high computational efficiency. The example of simulation states the effectiveness of the proposed algorithm.