Extended and Unscented Kalman filters for parameter estimation of an autonomous underwater vehicle

In this paper, a high performance procedure for estimating of hydrodynamic coefficients in Autonomous Underwater Vehicles (AUV's) is proposed. In modeling of an AUV, experimental data should be verified and validated using appropriate techniques. Due to implementation complexity in calculating methods, computation of hydrodynamic parameters is challenging. This paper presents analytical approaches for estimating an AUV's hydrodynamic coefficients. Nonlinear Kalman Filter (KF) algorithms are implemented to estimate unknown augmented states (coefficients). A comparative study is conducted which shows the superior performance of Unscented Kalman Filter (UKF) in comparison with Extended Kalman Filter (EKF).