基于神经网络的异常船舶航迹特征因子模型

Characteristic Factor Model of Abnormal Ship Track Based on Neural Network

船舶异常行为是海事安全科学理论研究的重要组成部分,它对于船舶监管以及海上安全通行具有重要意义,目前是航海领域的热点研究方向之一。根据船舶航迹分析船舶航行状态以及规范性,研究了船舶航迹行为特征,发现隐藏于航迹数据集中的船舶异常行为模型。针对目前危险船舶向外播放伪造数据造成的船舶监控不准确的异常船舶行为,提出了基于神经网络的异常船舶航迹特征因子模型。通过中分纬度算法分析轨迹数据的隐藏关系,选取实际舶船舶自动识别系统(automatic identification system,AIS)的数据对多层前馈神经网络(back propagation,BP)进行训练,使其完成对关系规则的学习,通过测试集数据对模型准确性进行判断。结果表明,该模型对于伪造船舶数据判断的准确率高于0.985。

The abnormal behavior of ships is an important part of the scientific theory of maritime safety,which is of great significance to ship supervision and safe passage at sea.It is currently one of the hot research directions in the field of navigation.In order to found the abnormal behavior model of the ship hidden in the trajectory data set,the ship's navigation status and standardization was analyzed that according to the ship's trajectory,and the characteristics of the ship's trajectory behavior was studied.Aiming at the abnormal ship behavior that the inaccurate ship monitoring is caused by the dangerous ships broadcasting forged data currently,a neural network was used to investigate the feature factor model of abnormal ship's trajectory.The mid-latitude algorithm was used to analyze hidden relationship of the trajectory data.In addition,the actual ship's automatic identification system(AIS)data was selected to train the back propagation(BP)neural network that in order to learning of the relationship rules,and the test set data was used to judge the accuracy of the model.The experimental results show that the accuracy of the model for judging forged ship data is higher than 0.985.