Ulysse Vautier

Abstract

The purpose of this study is to investigate novel methods on model identification, navigation, guidance and control of autonomous sailboats in order to sail robustly and efficiently in a wide variety of situations. The study has focused on the development of a robust platform, a model identification process for sailboats, robust control and path-planning methods. A hardware and software system was developed in order to build a research plat- form on which to experiment. Five sailboats were effectively adapted to be able to sail autonomously. Two one-meter long model sailboats, a two-meter long sailboat, a trimaran and a catamaran were tested using this system. Of which, a one-meter long sailboat was used to test developed model identification and navigation methods. A complex sailboat model identification process has been proposed to be, afterwards, used in Model Predictive Control (MPC) scheme. Moreover, a path-planning method for the shortest-length manoeuvre was proposed using the Dubin’s path. This method was then applied in an area scanning task formulated as a Traveling Salesman Prob- lem. The ability for an autonomous sailboat to be used in a wide range of applications was demonstrated through the design and development of novel control methods, a robust hardware configuration and a modular software system. This study has proven that the autonomous sailboat can be a genuine solution to challenging marine applications by providing a durable, sustainable and highly-controllable platform.

Keywords

Robotics, Autonomous Vehicle, Navigation, Sensors, Path-Planning, Optimisation, Sailing Robot, Control, MPC, Dubin's Vehicle, Dubin's Path, Guidance, Model Identification

Document Type

Thesis

Publication Date

2021

Download

Share

COinS