Verification of collision avoidance algorithms in open sea and full visibility using fuzzy logic

Autonomous Surface Vehicles (ASV) have been a subject undergoing intense study since they provide good potential regarding reducing the cost, increasing safety, reliability, efficiency and sustainability. The complexity of the ASVs needs to be tested to gain trust and to prevent fatal consequences. This paper proposes a method using fuzzy logic for evaluating the compliance of Collision-Avoidance Systems (CAS) with regard to the Convention on the International Regulations for Preventing Collisions at Sea (COLREG). For the proof-of -concept purpose, the set of rules for open sea and full visibility is considered; but the framework from this paper can be extended to include narrow channels, traffic separation schemes and restricted visibility by added more fuzzy rules. Fuzzy logic allows handling vague terms and represents linguistic variables and values written in COLREG in a quantitative mathematical form without introducing sharp assumptions or specifications to original COLREG. The paper develops four evaluation systems (1) evaluation of all vessels in pairwise encounters, (2) ownship (OS) evaluation in pairwise encounters for the interest of Class societies, (3) evaluation of all vessels in multiple-vessel encounters, and (4) OS evaluation in multiple-vessel encounters for the interest of Class societies. The designed systems were verified on a set of simulated scenarios. The obtained results were validated against visual assessment of the trajectory, heading and speed plots. The results showed that the evaluation systems provided variables that would be challenging or impossible to obtain by visual assessment.

Automated summary

This paper proposes a fuzzy logic-based method for evaluating the compliance of Collision-Avoidance Systems (CAS) in Autonomous Surface Vehicles (ASV). The evaluation systems were verified on simulated scenarios and found to provide variables that would be challenging or impossible to obtain by visual assessment.