Using Multiple Objective Optimization for Autonomous Sailing Vessels

Abstract

This research addresses using multiple object optimization, via the established opensource Mission Oriented Operating Suite-Interval Programming (MOOS-IvP) platform, to meet the unique navigational demands and operational constraints of autonomous sailing vessels. Recognizing a gap in the existing IvP Helm framework’s ability to accommodate the intricate dynamics of wind-powered navigation, this thesis initiates with the development of a sailing behavior. The core contribution of this work is the novel introduction of a sine wave-based approach for defining upwind tacking maneuvers. Building from a foundation in mathematical analysis, an algorithm was developed that employs the sine function to model the vessel’s tack plan. Furthermore, the thesis explores the integration of this behavior within the MOOS-IvP architecture, detailing the modifications necessary to support wind-powered navigation. Evaluation of the proposed navigation behavior encompasses simulated environments. The assessments highlight the algorithm’s adaptability to changing wind conditions. Through a combination of theoretical development and simulation, this study not only demonstrates the viability of integrating traditional sailing methods with contemporary autonomous systems but also contributes to advancing the capabilities of the standard MOOS-IvP tool kit and its continued use in various maritime applications.