Robust Extended Kalman Filter for Positioning Control of Ships in Presence of Parameter Uncertainties

Sea and rivers create several disturbances such as waves, wind and currents that affect ships' movements. In order to obtain an accurate control performance, precise information of ships is required. Since the proposed dynamical models for ships are nonlinear, one should design a nonlinear estimator or use effective techniques to overcome the problems caused by nonlinearity. Furthermore, there are some additive uncertainties in parameters due to lacking knowledge or faulty identification. Accordingly, a robust extended Kalman filter is presented in this paper to estimate required variables in nonlinear state space model while there are considerable uncertainties in the parameters. An advantage of this estimator is that it takes current forces, wind forces, and nonlinear terms into account, which have been often neglected or simplified in previous works. Eventually, simulations for both standard EKF and robust EKF are given and then compared to confirm accuracy of the proposed observer.