Ship collision is a classical problem for maritime practitioners and researchers. Human error is a major cause of collision accidents, which motivates researchers to develop automation systems replacing navigators on board. However, before autonomous ships fully replace conventional ships, supporting the situational awareness of human operators is still a rigid demand. Moreover, how to prevent automation performing out of the human’s expectations (e.g., violation of regulations) is another challenge. Therefore, the design of human-machine interactions (HMIs) becomes crucial.

This dissertation developed the Human-Machine Interaction oriented Collision Avoidance System (HMI-CAS) that allows human operators and automation to share their intelligence. Specifically, the HMI-CAS not only offers one (optimal) solution to human operators but also visualizes the solution space with both dangerous solutions and feasible solutions. Thus, the decision process of automation becomes transparent for human operators. The human operators can not only read and understand the solutions offered by the machine but also validate and modify the solutions via the interface of the HMI-CAS. Without human interventions, the HMI-CAS also can work automatically. Moreover, to support the humans take evasive action in time, the measure of collision risk utilizing a concept called “room-for-maneuver” is proposed, which offers alerts before collisions become inevitable.

In brief, instead of replacing humans on board, the proposed HMI-CAS aims at bridging the intelligence of humans and machines, which enriches the choice of collision avoidance systems for supporting human operators and for developing autonomous ships.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
Supervisors/Advisors
Thesis sponsors
  • China Scholarship Council
Award date18 Nov 2019
Place of PublicationDelft
Edition1
Publisher
  • Delft University Publishers - TU Delft Library
Print ISBNs978-94-6323-937-0
DOIs
Publication statusPublished - 18 Nov 2019

ID: 62722332