Journal
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
Volume 21, Issue 3, Pages 1167-1179Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TITS.2019.2902927
Keywords
Autonomous surface vehicles; collision avoidance; path planning; multi-objective optimisation; COLREGs
Categories
Funding
- Innovate UK Project Robust Machine Executable Collision Regulations at Sea [TSB 102308]
Ask authors/readers for more resources
This paper presents a multiobjective optimization approach for path planning of autonomous surface vehicles (ASVs). A unique feature of the technique is the unification of the Convention on the International Regulations for Preventing Collisions at Sea (COLREGs) with good seamanship's practice along with hierarchical (rather than simultaneous) inclusion of objectives. The requirements of collision avoidance are formulated as mathematical inequalities and constraints in the optimization framework and thus collision-free manoeuvres and COLREGs-compliant behaviours are provided in a seafarer-like way. Specific expert knowledge is also taken into account when designing the multiobjective optimization algorithm. For example, good seamanship reveals that if allowed, an evasive manoeuvre with course changes is always preferred over one with speed changes in practical maritime navigation. As a result, a hierarchical sorting rule is designed to prioritize the objective of course/speed change preference over other objectives such as path length and path smoothness, and then incorporated into a specific evolutionary algorithm called hierarchical multiobjective particle swarm optimization (H-MOPSO) algorithm. The H-MOPSO algorithm solves the real-time path planning problem through finding solutions of the formulated optimization problem. The effectiveness of the proposed H-MOPSO algorithm is demonstrated through both desktop and high-fidelity networked bridge simulations.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available