Automated berthing (parking) of autonomous ships

Abstract

Shipping operations can increase their efficiency by automating standard operations. This thesis explores the concept of automated navigation in a static harbor environment and automating the berthing procedure for a commercial ship. The ship is actuated with two stern azimuth thrusters and a bow tunnel thruster, giving it full maneuverability.

A literature study is done on berthing procedures, collision avoidance systems, and path following control. Several methods of collision avoidance are evaluated. An A* algorithm for path planning has been implemented and extended upon. A path following kinematic controller has been implemented in order to steer the ship along the planned path. The VPM has been implemented collision avoidance with stationary but unforeseen obstacles, not accounted for by the path planner. Finally, a nonlinear PID DP controller has been implemented to steer the ship the final distance to the berth.

Simulations have performed for a detailed hydrodynamic ship model by use of MATLAB Simulink. The results show that the A* method with extensions is a suitable path planning tool. The path following controller and the VPM for collision avoidance perform satisfactorily but are not robust. The DP controller does not perform satisfactorily as of now. Improvements for all the methods are suggested for future work.