Explainable AI and Autonomy for the Maritime Domain

Academic Institution: Heriot Watt University

Academic Supervisor: Professor Helen Hastie

Industry Partner: SeeByte

PhD Student: Konstantinos Gavriilidis

Start Date: 16th September 2019

Abstract

Techniques will be investigated to enable the human operator to interact in natural language to collaborate and team effectively with autonomous systems with varying autonomy and communication levels in the Marine domain. Current use of unmanned systems within the marine or maritime domain typically relegates the operator to the role of a spectator, while the unmanned system carries out pre-planned autonomous operations. On-board autonomy and AI capabilities are now developing to the point where elements of the mission execution and decision making process can be delegated to the unmanned system. Increased expectations around mission complexity and duration require that the interaction between the human operator and unmanned system is richer, with the autonomous system able to provide explainability and collaborative feedback, necessary for transparency and accountability. This research gap will be the focus of the investigatory work carried out as part of this PhD project.

Heriot-Watt University will collaborate with SeeByte for this research, a software company based in Edinburgh with world recognised domain expertise in Maritime Robotics. This collaboration will provide end-user access, as well as data and software products relevant to the research challenge.

The research impact will be through the generation of IP relevant to the management and interaction of human operators with (multiple) unmanned systems. This research domain is of critical importance as unmanned systems begin to be deployed in communication-limited environments for long, complex missions. This research has the potential to significantly advance unmanned systems technology and produce output that can be fed into SeeByte’s product suite.