量质融合的移动轨迹相似性查询技术研究

项目名称： 量质融合的移动轨迹相似性查询技术研究

项目编号： No.61502324

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 自动化技术、计算机技术

项目作者： 郑凯

作者单位： 苏州大学

项目金额： 21万元

中文摘要： 定位技术的快速发展以及智能手机、移动互联网的大规模普及，使得记录移动对象历史位置的轨迹数据产生了爆炸式增长，轨迹数据的存储与查询业已成为现代数据管理领域的重要分支。随着大数据时代对数据质量的管理提出了前所未有的高要求，轨迹数据也亟需从“基于数量”的管理模式向“基于质量和数量”的管理模式转变，其难点在于如何在海量、实时、多源、异质的轨迹数据上有效地提升数据质量以及进行高效地相似性查询。本项目拟从四个方面进行探索，包括：1. 系统性地分析、量化轨迹数据中存在质量问题以及对轨迹相似性度量函数所带来的影响；2. 设计轨迹校准框架和校准参考系统；3. 设计轨迹校准模型、校准算法以及增量式的维护方法；4. 设计基于校准轨迹的新型索引结构和查询优化技术。相关研究成果预计将对轨迹数据管理提供新的解决思路、显著提升相似性查询结果的可用性，因而具有重要的理论与实践意义。

中文关键词： 时空数据；移动对象；数据质量；移动轨迹

英文摘要： Rapid development of positioning technology as well as proliferation of smartphones and Mobile Internet has resulted in explosive growth of trajectory data, i.e., historical movement records of moving objects. Storage and query processing of trajectory data have consequently become an important branch in the field of data management. With the unprecedentedly increasing demand for data quality management in Big Data era, trajectory data management systems are required to pay attentions on both quality and efficiency issues. Its main challenges include how to effectively improve the quality and efficiently process similarity queries over large-scale, real-time, heterogeneous trajectory data. This project aims to investigate this problem from four aspects: 1. systematically analyse the impact of trajectory data quality on similarity measures; 2. design trajectory calibration framework and reference system; 3. propose trajectory calibration models, algorithms and incremental maintenance schemes; 4. develop novel indexing structure and query optimisations based on calibrated trajectories. Research outcomes of this project are expected to provide new methodology and toolkit for trajectory data management, especially in terms that they can enhance the effectiveness of trajectory similarity queries remarkably. As a consequence this project will have great significance in both theory and practice.

英文关键词： spatio-temporal data;moving object;data quality;moving object trajectory