An Autonomous Collision Avoidance Algorithm Based on Perception and Maritime Rules

Ship collision avoidance algorithms are the primary difficulty in autonomous navigation technology. We propose an autonomous collision avoidance method based on perception and maritime rules. By evaluating the collision danger, we provide early warning for ships to ensure navigation safety. In order to effectively solve the problem of uncoordinated collision avoidance behaviors caused by inconsistent division of encounter situations, our research optimizes the discrimination method based on the combination of the angle of gangway and heading intersection angle, establishes the role-symmetric encounter situation division algorithm, and determines the principle of collision avoidance compliant actions in multi-ship situations. The experiments of multi-ship collision avoidance in mixed complex situations comprehensively verify the effective performance of the autonomous collision avoidance algorithm based on perception and maritime rules, and provide strong support for its feasibility and superiority in practical applications.