Multi‐Stage Auxiliary Model Based Recursive Least Squares Estimation for Pseudo‐Linear System With ARMA Noise

ABSTRACT This paper focuses on the parameter estimation problem for pseudo‐linear systems with autoregressive moving average noise. In order to reduce the computational complexity of the identification algorithms, the original pseudo‐linear system is decomposed into three submodels and a three‐stage auxiliary model‐based recursive generalized extended least squares (3S‐AM‐RGELS) algorithm is proposed based on the hierarchical identification principle. The convergence analysis is provided to show that the parameter estimation error can converge to zero under the presented 3S‐AM‐RGELS algorithm. Finally, numerical simulations demonstrate the effectiveness of the proposed algorithms.