Performance characterization and comparison of conventional and machine-learning-based techniques for control of a USV

This study is part of ongoing work on situational awareness and autonomy of a 16' WAM-V USV. The objective of the study is to assess the merits of application of two different path-following methods for the motion control of the USV, including characterizing and evaluating their performance, robustness and potential in actual field implementation. The first method uses a conventional control technique based on linear theory, while the second method uses Reinforcement Learning. The performances of the two methods for selected case studies are compared and discussed.