一种基于高斯混合模型的船舶航迹聚类方法  

Ship Trajectory Clustering Method Based on Gaussian Mixture Model

作　　者：甄荣[1,2,3] 石自强 ZHEN Rong;SHI Ziqiang(Navigation College,Jimei University,Xiamen 361021,Fujian,China;Hubei Key Laboratory of Inland Shipping Technology,Wuhan 430063,China;Intelligent Transportation Systems Research Center,Wuhan University of Technology,Wuhan 430063,China)

机构地区：[1]集美大学航海学院,福建厦门361021 [2]内河航运技术湖北省重点实验室,武汉430063 [3]武汉理工大学智能交通系统研究中心,武汉430063

出　　处：《船舶工程》2021年第11期139-143,共5页Ship Engineering

基　　金：国家自然科学基金(52001134);福建省自然科学基金(2020J01661);内河航运技术湖北省重点实验室基金(NHHY2020001);福建省教育厅中青年科技项目(JAT190293);集美大学博士科研启动基金(ZQ2019014)。

摘　　要：文章提出一种基于高斯混合模型(GMM)的船舶航迹聚类表达方法。设计区域船舶自动识别系统(AIS)数据选取方法,从AIS数据中选取特定区域的信息。分析数据挖掘中高斯混合模型聚类的方法和原理,通过运用GMM方法和期望最大(EM)算法,对厦门港船舶AIS信息中包含的航迹点进行聚类处理。研究结果表明,聚类得到的船舶航迹簇分布与实际海上交通环境下表现的船舶行为是一致的,同时能获得船舶轨迹点数据分布均值和协方差等数理分布的参数。该方法有助于从机器学习的角度理解海上交通环境和船舶行为,能为海上交通环境中船舶行为的智能识别、理解和监控提供参考。A method for expressing ship trajectories based on Gaussian Mixture Model(GMM)is proposed.The data selection method of the regional ship automatic identification system(AIS)is designed,and select the information of a specific area from the AIS data.The method and principle of Gaussian mixture model(GMM)clustering are analyzed in data mining.By using GMM method and expectation maximum(EM)algorithm,the trajectories contained in the ship’s AIS information in Xiamen Port are clustered.The research results show that the distribution of the ship trajectories obtained by clustering is consistent with the ship behavior under the actual maritime traffic environment.At the same time,the mathematical distribution parameters such as the mean value and covariance of the ship trajectory point data distribution can be obtained.This method is helpful to understand the marine traffic environment and ship behavior from the perspective of machine learning,and can provide theoretical basis and support for the intelligent identification,understanding and monitoring of ship behavior in the marine traffic environment.

关 键 词：海上交通 高斯混合模型 期望最大算法 船舶航迹聚类 船舶行为 

分 类 号：U675.7[交通运输工程—船舶及航道工程] TP391[交通运输工程—船舶与海洋工程]