Autonomous Surrounding Ship Navigation In Virtual Maritime Environments

Abstract

Full mission marine training scenarios are essential in training marine trainees with ship handling simulators. Most of the current ship handling simulators endure with the common shortage of lacking intelligent navigation of target or surrounding vessels which complies with the COLREG (Convention on International Regulation for Preventing Collisions at Sea) navigational rules. Handling multiple surrounding ships in the environment with an individual instructor leaves a heavy workload to the instructor and assigning multiple in- structors to control surrounding ships makes the environment setting complex and increases the cost of the training scenarios. This research presents a novel approach to automate the surrounding vessels in the marine simulation environment with a central controller to con- trol the behaviours of each surrounding vessel in the marine environment. The controller uses positioning data of real ships which are derived from a set of historical Automatic Identi cation System (AIS) data and map the positioning data with the surrounding ships in virtual environment to obtain the navigation. An Ant Colony Optimization (ACO) algo- rithms is used to avoid the collisions by generating a new path when both trainees ship and surrounding ships are in the vicinity of each other. The collision handling mechanism uses COLREGS rules and avoids both static and dynamic obstacles in the environment. The ACO algorithm generates successful collision avoided paths in all head-on, crossing and over-taking encountering situations. The number of turning points were used to measure the accuracy of the path and no of obstacles in the environment a ects the smoothness of the path. With this results it is revealed that AIS data can be used as an assistant to automate surrounding vessels in a virtual maritime environment together with a proper collision avoidance mechanism.