Estimated Parameters Of Auto Regressive eXogenous Model Using Two Stage Least Square and Recursive Least Square Algorithm

In this investigation, output error models based on two-stage and auxiliary models are investigated with those based on least squares and recursive least squares algorithms. The main principle is to use the decomposition technique and the auxiliary model identification notion to separate a system into two subsystems, each of which comprises a parameters variable. Then, employ each algorithm separately. Root mean square error is applied to calculate model validation procedures. Compared with the two stage model-based recursive least squares algorithm, the TS-RLS algorithm has more parameter estimated accuracy.