Least square and Instrumental Variable system identification of ac servo position control system with fractional Gaussian noise

Abstract

In this paper, the classical Least Square Estimator (LSE) and its improved version the Instrumental Variable (IV) estimator have been used for the identification of an ac servo motor position control system. The data for system identification has been collected from a practical test set-up for fixed command on the final angular position of the servo motor with varying level of velocity and acceleration. The measured data is corrupted then with externally induced random noise having a Gaussian distribution, commonly known as white Gaussian noise (wGn). Performance of the LSE and IV estimators are also compared for fractional Gaussian noise (fGn) which have heavy tails in its statistical distribution and are capable of modeling real world signals having spiky nature.