L. Zhao and X. Fu, “Sequential multi-ship encounter scenarios for testing automatic collision avoidance systems [Abstract],” in Proc. 6th Int. Conf. Maritime Autonomous Surface Ship (ICMASS), 2024.
Abstract:
Simulation-based digital testing is considered a fundamental step and approach for safety
testing of autonomous vessel collision avoidance systems. It provides a cost-effective means to
simulate a wide range of scenarios, facilitating quantitative performance assessments and
enabling traceable detection of failures. In the construction of safety-critical scenarios for
testing, besides basic single-ship encounter scenarios, attention has increasingly turned to the
complex encounter scenarios involving multiple ships and their systematic construction. Existing
solutions primarily focus on situations where the tested vessel encounters multiple other vessels
simultaneously. However, in scenarios occurring within high-density traffic areas, such as
intersection areas in narrow waterways, the maneuvering space of the tested vessel may be
constrained by surrounding vessels or result in other close-quarters situations. The background
vessels in these scenarios, although not directly encountered in the initial setup with the tested
vessel, are still vital components of safety-critical scenarios. Their systematic construction has
yet to be fully discussed. This article categorizes such scenarios as Sequential Multi-Ship
Encounter Scenarios and proposes a scalable automated scenario generation method, aiming to
enhance the diversity of digital collision avoidance testing scenarios
License type:
Attribution 4.0 International (CC BY 4.0)
Funding Info:
This research / project is supported by the Singapore Maritime Institute - Programme of Maritime AI Research
Grant Reference no. : SMI-2022-MTP-06