BP neural network-based ship track real-time prediction

The paper proposed the thought of predicting differences of co-ordinates to realize the ship track prediction,and the ship track prediction model based on three-layer BP network was developed.Course and speed as well as difference in longitude and latitude were regarded as the input and output parameters, respectively.Results indicate that it costs less with high accuracy and less parameters,which meets requirements of accuracy,real-time performance,universality for VTS.It shows that the method is effective and feasible.