Predictive Controller with UKBF for Marine Dynamic Positioning System

For surface vessel dynamic positioning (DP) system controls, a design method for predictive controllers is proposed based on the unscented Kalman-Bucy filtering (UKBF) algorithm and on a non-switching analytical model predictive control (NSAMPC) method. To avoid linearization and discretization errors in the model, the UKBF algorithm was applied to determine motion-state estimates of the ship. On the basis of the relative degree concept and non-linear ship motion models, an NSAMPC controller was designed in order to maintain a floating vessel at a specific position. The proposed method was verified by performing simulations on a marine supply vessel. The simulation results exhibit a smoother output of the designed controller, which is helpful for reducing the wear on the propeller. Moreover, the proposed controller results in a settling time of less than 40 s for the north position of the vessel with an overshoot of less than 5%. Likewise, the controller results in an east-position settling time of less than 60 s with an overshoot of less than 5%. In addition, the maximum offset of the vessel heading angle is less than 1.5°. The proposed controller allows the ship to reach the specific position quickly.