Robust estimation of hydrodynamic coefficients of an AUV using Kalman and H∞ filters

Abstract

In this paper, a robust procedure for estimating of nominal hydrodynamic coefficients of an autonomous underwater vehicle (AUV) in presence of parameter uncertainties is proposed. Some parameters of the model of an AUV, such as its hydrodynamic coefficients, may perturb from their nominal values because of changes in its operating conditions and the surrounding environment. This paper presents two different filters for robust identification of the AUV parameters. The robust extended Kalman filter (REKF) and the extended H-infinity (EH∞) filter are implemented to estimate the nominal hydrodynamic coefficients of the AUV. By using a model of a sample AUV and simulating the motion of the vehicle in the diving mode involving parameter perturbations, the performance of the filters is compared. The results show that the REKF outperforms the EH∞ filter in this application. Also, the performance of these estimation algorithms is compared with the standard extended Kalman filter (EKF), using the data obtained from experimental test of a real AUV. It is observed that the model using the identified values of hydrodynamic coefficients of both filters can predict the actual trajectory of the vehicle and the robust filters are more accurate as compared to the standard EKF.