Requirements for Optimal Local Route Planning of Autonomous Ships
Sung-Wook Ohn, Ho Namgung
Abstract
Ships transport large volumes of cargo, and are therefore major contributors to the global economy. Ship collisions can cause significant economic losses. Path-planning algorithms can prevent such collisions by suggesting the optimal path for navigation. Conventional path-planning algorithms are disadvantageous, because they do not consider the navigation practices followed by experienced navigators. Therefore, in this study, we developed the requirements for optimal local path planning of autonomous ships by considering the open sea, restricted waters, and two-ship and multi-ship interactions, in addition to the navigation practices adopted by navigators and rules in COLREGs part B. First, the navigation practices under various scenarios were collected. Subsequently, these practices were linked to COLREGs part B to extract the key rules and keywords related to collision avoidance. Finally, the requirements for generating the optimal local path were drafted based on the rules and keywords. The utility of the requirements was demonstrated by applying them to representative path-planning algorithms for the timely and accurate evaluation of their effectiveness. The proposed requirements can be utilized to improve the existing path-planning algorithms and develop superior algorithms in the future.