基于BP神经网络的船舶航迹实时预测

BP neural network-based ship track real-time prediction

提出以预测船位差实现航迹预测的思想,并设计基于三层BP神经网络的航迹预测模型.模型以航向、航速和经度差、纬度差作为输入输出.实验表明,该算法预测精度高、算法耗时短、计算参数少,符合VTS对航迹预测的准确、实时、通用性的要求,表明方法是有效可行的.

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.