A New Least Squares Iterative Estimation Algorithm for CARAR Systems

Abstract

Mathematical models are the base for the system analysis and the controller design. This paper focuses on the identification problems of controlled autoregressive models with autoregressive noise (CARAR system for short). By applying the iterative method and the hierarchical principle, a least squares identification algorithm is investigated. The key of this algorithm is replacing the unknown noise terms in the information vector with their estimated residuals. The effectiveness of this approach is demonstrated by the simulation experiment.