Experimental validation of a velocity obstacle based collision avoidance algorithm for unmanned surface vehicles

Abstract

This paper presents experimental validation results of autonomous collision avoidance algorithms using an unmanned surface vehicle (USV). For autonomous collision avoidance while following given waypoints, the existing line-of-sight (LOS) guidance and velocity obstacle (VO) algorithms are modifed and applied to this USV. The proposed collision avoidance algorithm considers the rule 13 to 17 in the international collision regulations (COLREGs) and provides a rule-compliant evasive path. The performance and practical validity of the developed autonomous navigation and collision avoidance capabilities are demonstrated and discussed using the results of real-sea experiments conducted with the USV M-Searcher, developed by the Agency for Defense Development, Republic of Korea.