Addressing COLREG Rule 2b: COLREG and Risk Aware Manoeuvre-Based Predictive Planning
This paper presents a novel collision avoidance method for Autonomous Surface Vessels (ASVs) that considers nuanced risk and COLREG profiles to determine an optimal safe manoeuvre. The COLREGs are a set of collision regulations defined by the International Maritime Organisation by which all compliant vessels should manoeuvre safely. However, COLREG Rule 2(b), in summary, states that a vessel may, if needed, make “a departure from these rules necessary to avoid immediate danger.” Our proposed COLREG and Risk Aware Manoeuvre-based Predictive Planner (CRAMPP) collates risk and COLREG compliance, along with other metrics, into a single cost function. The cost function features a novel dynamic weighting system that facilitates COLREG Rule 2(b). A motion model is used to simulate the ASV dynamics at varying speed and course setpoints as inputs to the cost function, which is optimised to find a manoeuvre with minimum risk and COLREG compliance cost. Field and simulated experiments were conducted to evaluate and demonstrate the efficacy of CRAMPP in one-on-one and complex multi-ship scenarios.
Authors
- Matthew Dunbabin (ORCID: https://orcid.org/0000-0003-0806-7720)
- Jason Ford (ORCID: https://orcid.org/0000-0001-5847-6253)
- Jordan Gleeson (ORCID: https://orcid.org/0000-0002-9296-5699)
Institutions
- Queensland University of Technology (AU)
Publication Details
- Journal
- IEEE Robotics and Automation Letters
- Published
- 2025-05-29
- DOI
- https://doi.org/10.1109/lra.2025.3575003
- Primary Topic
- AI-based Problem Solving and Planning
- Type
- article
- Field-Weighted Citation Impact
- 0.00