利用DBSCAN优化船舶领域算法的实时碰撞预警模型  

A Ship Domain-Improved Real-time Collision Warning Model Based on DBSCAN

作　　者：何渡[1] 肖丝雨 任俊伟 刘子恒 沈昕 李凯文 陈智军 HE Du;XIAO Si-yu;REN Jun-wei;LIU Zi-heng;SHEN Xin;LI Kai-wen;CHEN Zhi-jun(Hubei Academy of Scientific and Technical Information,Wuhan 430063,China;School of Computer Science and Information Engineering,Hubei University,Wuhan 430062,China)

机构地区：[1]湖北省科技信息研究院,湖北武汉430063 [2]湖北大学计算机与信息工程学院,湖北武汉430062

出　　处：《软件导刊》2021年第2期107-113,共7页Software Guide

基　　金：湖北省自然科学基金面上项目(2019CFB757);国家水运安全工程技术研究中心开放基金项目(A2019011);内河航运技术湖北省重点实验室基金项目(NHHY2017001);湖北省教育厅科学研究计划重点项目(D20161001);中央高校基本科研业务费专项项目(2019III050GX,2019III007GX)。

摘　　要：船舶碰撞预警对于智慧城市中智能航道的实现具有重要意义。航船行驶极易受天气、水域、航速等复杂多变因素的影响,导致船舶碰撞预警的时效性和准确性不高。针对该问题,提出一种基于DBSCAN改进船舶领域算法的实时碰撞预警模型。该模型基于大规模船舶AIS轨迹,结合船舶行驶速度、船舶形状等因素,优化改进了传统船舶领域计算方法,并利用DBSCAN空间聚类思想,同时考虑船舶位置和船首向,在大量实时AIS数据下进行碰撞范围预测与潜在危险船舶预警。该模型使用长江流域真实AIS数据进行实验,并采用碰撞点时间估计法进行对比实验,结果表明,该模型能够有效提高碰撞预警的准确性和时效性。Ship collision warning is of great significance to the realization of intelligent waterways in smart cities.Ship movement is easily affected by weather,water area,speed and other complex and variable factors,resulting in the low timeliness and accuracy of ship collision warning.Aiming at this problem,this paper proposes a ship domain-improved real-time collision warning model based on DBSCAN algorithm.This model improves the calculation method of traditional ships domain by considering the dynamic factors(e.g.speed,shape,direction,etc.).By utilizing the DBSCAN—a spatial clustering method,the ships’location and heading direction at the one moment are taken account into a collision range prediction model for the potential dangerous warning.In order to verify the efficiency of the model,a large scale of real AIS data around the Yangtze River Basin is used in the experiments.The experimental results show that the accuracy and timeliness of the proposed model can be obviously enhanced compared with the traditional collision warning model.

关 键 词：时空大数据 AIS 船舶领域 实时航情 DBSCAN 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]