Ant colony optimized planning for unmanned surface marine vehicles

Abstract

This paper presents some results achieved from a preliminary study on the use of the Ant Colony Algorithm to plan feasible optimal or suboptimal trajectories for an autonomous ship manoeuvring. The scenario, for this preliminary work, comprises only open sea manoeuvres. The goal involves obtaining the least time consuming ship trajectory between to points, departing from the start point with arbitrary initial speed and attitude values and arriving to the end point with predefined speed and attitude values. The specific dynamic of the ship imposes typical restrictions to its manoeuvrability. In the present case, the non-holonomicity, the rate speed/turn radius, and the imposed forward-only propulsion of the ship make up the main restrictions to the ship movement. For long distances, the problem could be tackled as a classical navigation problem, in which, for the most part of the ship trajectory, techniques such inertial navigation should be enough. The problem arises at short distances when it becomes a manoeuvring problem. In this case to obtain a optimal, --in some cases just a feasible--, trajectory could be a difficult problem.