Application of Moving Horizon Filter for Dynamic Positioning Ship

Abstract

The dynamic positioning (DP) ship maintains its position and heading by active thrusters. Some measurements of the ship motion variables can be obtained from certain reference sensor systems. However, the other measurements may be expensive or impossible to be obtained, such as ship velocity and acceleration. Fortunately, they can probably be estimated using the available measurements and the ship dynamic mathematical model. The dynamic mathematical model of a ship is based on a hydrodynamic and aerodynamic description. In order to handle estimation errors due to ship modeling uncertainty, the moving horizon filter is designed by combining the ship dynamic model, the concept of moving horizon and optimum control. A simulation of a dynamic positioning supply ship is carried out. The modeling uncertainty is considered respectively with time-invariant and time-varying mass matrix and damping matrix. By comparing moving horizon filter with Kalman filter, the effectiveness and robustness of the moving horizon filter is demonstrated for DP ship.