城市客流移动数据的模式分析、可视化和决策优化支持

Most cities in China are facing severe problems of traffic congestion, environment pollution and shortage of energy. One of the key solutions to these problems is to effectively and systematically improve the efficiency of urban transportation, reduce vehicle traffic while meeting human mobility needs. This project focuses on developing new methodologies for analyzing big urban mobility data of floating cars, discovering and understanding inherent rider patterns of urban transportation (including buses and taxi cars), and developing a computer-assisted and interactive route planning optimization system to support public transportation planning with the extracted data and discovered patterns. Specifically, passengers' origin-destination (OD) informatin will be extracted from GPS trajectories of taxi cars, buses and anonymous mobile phone location records in an urban area, for different time periods. Surveillance videos on buses will also be analyzed to assess the occupancy rate for each bus for a time and location (from floating car data). The project will develop a new multi-scale flow clustering method that is efficient and effective to discover inherent flow patterns from massive OD data, which can help understand urban passenger patterns and optimize bus route planning. With the discovered patterns of urban passenger flows, in-depth analysis will be carried out to understand the relationship between taxi car ridership and existing bus routes and capacities. Particularly, we are interested in answering questions such as: How much can existing bus routes accommodate the needs of taxi passengers? What are the possible reasons that taxi riders chose to use a taxi instead of a bus (e.g., no available bus, too many transfers, time saving, or simply having no knowledge of available buses)? Can currently bus routes be improved to accommodate more needs? Finally we will develop an optimization method and system to facilitate the assessment, improvement and new planning of bus routes, with consideration of multi-scale urban passenger flow patterns, multiple objective functions and constraints. The optimization system leverages the power of state-of-art spatial optimization algorithms, informative patterns discovered from urban mobility data, and experts' domain knowledge and envisions, through an interactive optimization framework. The ultimate goal is to develop new theory and methodology to help scientists and practitioners analyze urban passenger flow data, obtaining insights, and better design urban transportation systems.

我国大中城市普遍面临着交通拥堵、环境污染、能源紧缺等问题。科学地提高城市交通效率，减少车流、疏通人流是解决这些问题的关键之一。本项目围绕城市大数据分析和城市交通规划决策支持，从理论探讨、方法研究、可视化技术、规律探索、规划优化等不同层面开展基本理论和关键技术研究。本项目基于出租车和公交车GPS 轨迹数据，结合手机信号和公交车视频，提取出租车和公交车载客情况及乘客的起终点（OD）信息，开发新的多尺度流向聚类分析和制图综合可视化方法，综合分析城市出租车客流和公交车客流的时空模式，评估公共交通线路配置，构建基于多尺度流向规律的多目标交互式公共交通路线优化方法，为城市交通改善和规划优化提供切实有效的理论、方法和技术支持。

目前中国，大中城市普遍面临着交通拥堵、环境污染、能源紧缺、疾病传播等问题，严重影响了城市的宜居程度和持续发展。这是城市化进程中面临的挑战，也是必须要依靠科学技术解决的问题。本项目围绕城市浮动车大数据分析和城市交通规划决策支持开展基本理论和关键技术研究。 本研究提出并将设计多尺度流向聚类方法、流向地图制图综合方法，自动提取多层次的流向模式，实现自动的多尺度流向制图综合，能够准确且快速地提取并可视化海量OD数据中的多尺度的复杂模式。基于上述理论创新的同事，本项目开发基于海量OD数据和多层次流向模式的公共交通规划优化方法。本项目的研究成果将对城市规划、交通规划、应急管理等多个领域具有现实意义。