A maneuvering velocity obstacle algorithm for ship collision avoidance

Safe and effective collision avoidance is the key for ships to achieve the autonomous navigation. This paper proposes a novel maneuvering velocity obstacle (MVO) algorithm, which considers the ship velocity area more accurately with a ship maneuvering mathematical model group (MMG) model and current navigation states. Specifically, the MVO algorithm considers ship MMG model and uses planar geometry methods to determine the reachable collision avoidance area of the velocity vector of own ship (OS). It calculates the collision avoidance velocity and course targets within a specified time step, uses a proportional integral derivative (PID) controller to execute the target instructions, and updates the states of OS to achieve dynamic collision avoidance. Three cases are carried out with different encounter situations. The results indicate that the OS using the MVO algorithm can safely avoid both static and dynamic obstacles. The collision avoidance performance is superior to that of the traditional velocity obstacle (VO) algorithm and complies with navigation rules and ship maneuvering capabilities.