基于回声状态网络的船舶交通事故预测

Vessel traffic accident prediction based on echo state network

船舶交通事故的预测结果对船舶交通智能管理具有指导性意义,针对当前船舶交通事故的预测误差大,建模过程耗费时间长等难题,设计基于回声状态网络的船舶交通事故预测模型。首先对当前船舶交通事故预测研究现状进行分析,指出各种船舶交通事故预测建模方法的局限性,然后收集大量的船舶交通事故历史数据,并进行一定预处理,构建船舶交通事故预测样本数据,然后通过回声状态网络的学习建立船舶交通事故预测模型,并采用具体船舶交通事故预测仿真实例分析其性能,回声状态网络的船舶交通事故预测精度超过95%,预测结果十分稳定,缩短了船舶交通事故预测建模过程耗费的时间,是一种高精度、速度快的船舶交通事故预测方法。

The prediction results of ship traffic accidents are instructive to the intelligent management of ship traffic.Aiming at the problems of the prediction errors of current ship traffic accidents and the long time spent in the modeling process,a prediction model of ship traffic accidents based on echo state network is designed.Firstly,the current research status of ship traffic accident prediction is analyzed,and the limitations of various ship traffic accident prediction modeling methods are pointed out.Secondly,a large number of historical data of ship traffic accident are collected and pre-processed to construct sample data of ship traffic accident prediction.Then,the ship traffic accident prediction model is established by learning the echo state network,and the specific ship is used.The performance of the simulation example of traffic accident prediction is analyzed.The prediction accuracy of the echo state network is more than 95%.The prediction result is very stable,and it shortens the time consumed in the modeling process of the ship traffic accident prediction.It is a high precision and fast prediction method of ship traffic accident.