A STUDY ON AN AUTOMATIC SHIP BERTHING USING A NEURAL NETWORK WITH DECOUPLE STRUCTURE

Abstract

Automatic ship control is a tendency, which towards to fully automatilization of ship control process, to make unmanned ships in future. With ship collision avoidance, ship berthing is complex problem affecting to maritime safety, human life, sea pollution, and ..etc. Automatic ship berthing has been studied early; however, existing systems always exist certain drawbacks. In this research, a decouple neural network is proposed for automatic ship berthing. With the saparerate structure of neural network, the performance of automatic berthing process is improved compared to existing berthing systems with the non-decouple structure in the controller, on regulating the rudder and the propeller. The result of numerical simulation shows the perfomance of proposed system based ship berthing in comparison with previous controllers.