繁忙水域船舶交通流特性提取及运行态势研判

Considering the dense vessel traffic, cross waterways, and complex traffic conflicts in crowded waters, facing the development tendency and demand of maritime supervision and management, making a full sufficient reference of advanced concepts and theories from road, pedestrian, and air transportation area, while based on current research achievements in marine transportation area. In this project, on the basis of collection and comprehensive pre-processing of multi-source spatial-temporal vessel trajectory data, implement original studies in area-wide vessel traffic flow dynamic characteristics extracting and analytical modelling and area-wide vessel traffic flow morphological characteristics extracting and measuring, systematically propose vessel traffic flow state judgement theory and method in crowded waters. In terms of theory, innovatively construct area-wide vessel traffic flow multi-parameter model, put forward measuring methods of vessel traffic heterogeneity and complexity, revealing the morphological characteristics of vessel traffic flows, enrich vessel traffic flow theory and method; In terms of technique, improve the comprehensive pre-processing methods of multi-source spatial-temporal vessel trajectory data, propose calibration algorithms of area-wide vessel traffic flow analytical model and measuring algorithms of area-wide vessel traffic flow morphological characteristics, systematically propose vessel traffic flow data analysis and modelling techniques in crowded waters, provide technical support of practical applications. The undertaking of this project is theoretically and practically significant for breaking through bottle-neck of the scientific development of maritime supervision and management and expanding discipline connotation of marine traffic engineering.

针对繁忙水域船舶交通密集、航路交叉穿越、交通冲突复杂等特点，面向海事监管发展趋势和需求，充分借鉴道路、行人和航空交通领域相关先进理念，并立足于水路交通领域已有研究成果，在多源船舶时空轨迹数据收集与综合预处理基础上，从区域船舶交通流动态特性提取与解析建模、区域船舶交通流形态特征分析与度量两方面展开原创性研究，系统形成繁忙水域船舶交通流运行态势研判理论与方法。在理论上，创新地构建区域船舶交通流多参数解析模型，提出船舶交通流不均匀性和复杂性度量方法，揭示船舶交通流形态特征，丰富船舶交通流理论与方法；在技术上，改进多源船舶时空轨迹数据综合预处理算法，提出区域船舶交通流解析模型参数标定算法和形态特征度量算法，系统形成繁忙水域船舶交通流数据分析与建模成套技术，为实际应用提供技术支持。项目研究对突破繁忙水域海事监管科学化发展的瓶颈制约、拓展水上交通工程的学科内涵等方面具有重要的理论与现实意义。

繁忙水域具有船舶交通密集、航路交叉穿越、交通冲突复杂等特点，是各级海事动态监管工作的重点。本项目在繁忙水域多源船舶时空轨迹与运行数据收集基础上，完成了多源船舶时空轨迹与运行数据综合预处理方法研究，提出了基于Bi-LSTM的船舶航行轨迹修复、基于改进DBSCAN算法的船舶轨迹聚类、基于ADP算法的船舶轨迹压缩、基于聚类与深度学习的船舶转速-航速-油耗关系建模等算法；提出了基于三维时空轨迹的区域船舶交通动态特性提取方法，构建了基于特性关系的区域船舶交通流解析模型；基于点集偏差理论，构建了区域船舶交通流空间异质性度量模型，综合DTW、K-Means、轨迹骨架提取算法，构建了区域船舶交通流时空异质性度量模型；在提出船舶间冲突状态和趋势判定方法的基础上，构建了基于船舶冲突面积、冲突角度和冲突距离三指标的船舶交通复杂度度量模型。总体上看，本研究解决了两个关键科学问题：（1）区域船舶交通流动态特性提取；（2）区域船舶交通流不均匀性和复杂性度量。项目研究取得的研究成果对突破繁忙水域海事监管科学化发展的瓶颈制约、创新交通流建模技术、拓展水上交通工程的学科内涵等方面具有重要的理论与现实意义。