PhD in Computer Science: Model-Based Assurance of Autonomous Surface Vehicles

I started my PhD at Linköping University in August 2024, in the Department of Computer and Information Science (IDA), division of Software and Systems (SAS), within the Programming Environments Laboratory (PELAB). My principal supervisor is Prof. Dániel Varró and my co-supervisor is Dr. Ulf Kargén. The position is funded by CUGS, the National Graduate School in Computer Science, and I am affiliated with the Wallenberg AI, Autonomous Systems and Software Program (WASP).

The thesis topic in one paragraph

Autonomous surface vehicles (ASVs) have to complete their missions without endangering other traffic, which at sea means following the International Regulations for Preventing Collisions at Sea (COLREGS). The COLREGS were written for human operators. They rely on judgement ("safe distance", "readily apparent" manoeuvres) and become ambiguous once three or more vessels meet. Showing that an ASV complies with them therefore calls for systematic, simulation-based testing. Historical traffic data and random scenario generation miss the rare combinations that matter most. My thesis, Enhancing Safety Assurance for Autonomous Surface Vehicles through Model-Based Test Generation, works on closing that gap.

Research objectives

1. A specification language for system-level ASV test scenarios. Model vessels, obstacles, geometric relations and COLREGS situations using qualitative abstractions.
2. Automatic derivation of assurance suites. Generate relevant, functionally complete, diverse and robust multi-vessel scenarios.
3. Execution and evaluation. Run the scenarios in naval simulators, monitor COLREGS compliance and produce assurance evidence.

The detailed story of the research, including what is done and how the pieces fit, is in Model-driven scenario-based assurance for COLREGS compliance.

What a PhD year looks like here

• Research (80%): based on my individual study plan, the work is done with three-week sprints, retrospectives, weekly meetings with my supervisors, and extra sessions around paper deadlines.
• Teaching (20%): lab and project supervision in TDDC88 Software Engineering, examination grading, and thesis supervision.
• Courses: CUGS core courses and WASP courses (autonomous systems, software engineering and cloud, topological data analysis, ethics), currently 79.5 hp.
• Community: WASP winter conferences and cluster workshops, MODPROD, and WARA-PS field and demonstration weeks in Västervik.

Funding and related project

My PhD position is funded by the CUGS graduate school. The research is closely related to ScentUSV (Automated test scenario synthesis for verifying collision avoidance of autonomous vessels), a 2024–2026 project of my supervisors funded by the Department of the Navy award N62909-24-1-2006 issued by the Office of Naval Research. The award supports my supervisors' work, so our joint papers acknowledge it, but it does not fund my PhD. The PhD continues after the project ends, with planned completion in 2029.