| dc.description.abstract | Throughout human history, technological innovation has been an important driver for enhancing human living standards. In the maritime industry, one such technological innovation is the maritime autonomous surface ship (MASS). Its development is seen as an opportunity to increase safety while simultaneously improving environmental performance and enabling more cost-effective shipping.
Research into making MASSs has been happening for some time. Several MASS concepts are being tested around the world. These MASSs must comply with the International Regulations for Preventing Collisions at Sea 1972 (COLREGs), which provide rules that apply to all vessels upon the high seas and in all waters connected therewith navigable by seagoing vessels.
While it has been proven to be no easy matter to construct an algorithm that ensures MASSs comply with the COLREGs, most of the research related to MASSs has focused on overcoming the technological challenges involved.
Research on COLREGs-compliant MASSs is scarce and has often produced algorithms that either clearly contradict the COLREGs or simply ignore rules crucial to seamanship practice, such as the safe speed rule. To support safe implementation of MASSs in the future, this thesis presents two objectives: first, to list identified safety challenges for MASSs and, second, to focus on the safety challenge of how MASSs can autonomously determine safe speed in any situation in accordance with the COLREGs.
A proposed approach to ensuring MASS compliance with the COLREGs was to use automatic identification system (AIS) data to allow artificial intelligence (AI) to learn the most effective, efficient, and COLREGs-compliant ways of manoeuvring. The research conducted in this thesis has investigated if historic AIS data could be used as a reference for safe vessel behaviours. It was determined that vessel speed data from historic AIS data cannot represent safe vessel speeds as per contemporary safe speed understandings. Feeding AI historic AIS data would therefore not be a suitable solution for teaching MASSs safe vessel speeds in different situations.
Consequently, interviews were conducted with experienced navigators to learn how they determine safe vessel speeds. The interview findings demonstrate that the work as done by navigators differs significantly from work as imagined by researchers and legal scholars. Safe vessel speeds are determined differently at sea from how they are perceived in the literature.
These findings have several implications. Programming MASSs to behave according to the work-as-imagined parameters could provoke problems in coordination and cooperation with conventional vessels, thus posing a risk to safety at sea. To operate safely, MASSs must therefore consider the work done in practice by human navigators. It is thus recommended that MASSs follow a goal-based approach when attempting to follow the safe speed rule of the COLREGs. The designers of MASSs must resolve their own method of determining the safe speed for these vessels in different situations. The goal should not be limited to compliance with the rules. They must also ensure that the MASS is in control of the situation and that its actions are transparent and understandable to other vessels in the area. | en_US |
