台湾海峡船舶AIS轨迹数据挖掘及其航路规划优化

In the new geographic information era, Geo-spatial data such as global/regional ships’ movements are characterized by open access and high dynamics. Ships’ movements data track the change of individual sailing status and its voyage related message, undoubtedly contain rich semantic information. The movement of ships is strongly influenced by external factors in their surrounding environment such as weather, sounding. Unfortunately, traditional trajectory analytics neglected those external factors and could not meet the current need of ships’ movement data science. This study integrate the cutting-edge techniques from machine learning, scientific visualization and GIS to discover cumulative activity patterns of ships' movements. Firstly this study will link ships movement data with the environment to discover the impact of environmental change in cumulative ships mobility ;Then the pattern recognition of ship-ship encounter event and its quantitative risk model are developed, then this thesis proposes a cluster model of ship spots in anchor/berth point and its time-varying characteristics. From the perspective of surface, this study develops a statistic inferring algorithm for delineating principal fairways and evaluating their traffic pressures. The Taiwan Strait is taken as a case study to demonstrate the analysis and visualization methods proposed. This study may be useful to location-based services, maritime safety and ocean spatial planning.

开放性与动态性是新地理信息时代的重要特征。具有良好开放性和高动态性的区域乃至全球AIS船舶轨迹蕴含着丰富的时空语义信息，是地理空间大数据的典型代表。海洋运输过程易受海洋环境的影响和制约，这是海上交通运输领域的独特性。综合考虑船舶群体移动过程和海洋环境的动态影响，对船舶运输安全和集约化运营至关重要。地理信息科学的传统时空轨迹分析方法在处理船舶轨迹时，遇到了巨大的挑战。为此，本申请项目拟发展海洋运输船舶群体性移动模式的数据挖掘方法，从船舶轨迹与海洋环境动态信息的时空融合、船舶运输主航迹带识别、海上航路拓扑网络构建及其交通压力评价、船舶运输冲突隐患事件多发区及其风险评估等方面入手，研究台湾海峡船舶群体移动性时空特征，为台湾海峡航路规划提供优化措施。本课题是地理信息科学、海上交通运输工程学的交叉研究，可为异构轨迹大数据分析提供新的思路和方法支撑。

开放性与动态性是新地理信息时代的重要特征。具有良好开放性和高动态性的区域乃至全球AIS船舶轨迹蕴含着丰富的时空语义信息，是地理空间大数据的典型代表。海洋运输过程易受海洋环境的影响和制约，这是海上交通运输领域的独特性。综合考虑船舶群体移动过程和海洋环境的动态影响，对船舶运输安全和集约化运营至关重要。地理信息科学的传统时空轨迹分析方法在处理船舶轨迹时，遇到了巨大的挑战。项目研究以船舶AIS轨迹数据为研究对象，研究了船舶轨迹与海洋环境动态信息的时空融合、船舶运输主航迹带识别、海上航路拓扑网络构建及其交通压力评价、船舶运输冲突隐患事件多发区及其风险评估等相关方法，实现了海量船舶轨迹数据的存储、查询和空间分析的集成，构建了一种基于光栅法识别海上主航迹带边界的统计推断模型，发展了船舶相遇事件的异步网格遍历检测算法。经过三年的努力，本课题组已发表了5篇论文（其中SCI 论文1篇，EI会议论文1篇）。