Automatic Berthing Control System Using Network and Knowledge-base

Abstract

Automatic berthing system for ships using neural networkand knowledge-base is developed and confirmed its availability by simulation. Three layer neural net work is used and error back propagation method is employed for learning of weighting factors and offsets. According to the simulation, it is verified to get the successfull results not only for the given pattern as teaching data but also for interporated or exterporated cases. It is further confirmed to work satisfactorily even under the wind disturbance.