轨迹相似性度量方法研究新进展

New Progress of Research on Trajectory Similarity Measurement Methods

轨迹数据广泛应用于智能交通、自然保护、疫情防控等领域。轨迹相似性度量是轨迹查询分析中最复杂和耗时的操作之一,是轨迹数据管理领域的研究基础。文章首先将轨迹相似性度量方法按对时间信息是否敏感划分为时间敏感型和非时间敏感型,同时介绍了基于语义和深度学习的新型轨迹相似性度量方法;然后对每类度量方法进行了综合对比分析,并给出了各自的优缺点;最后对本领域未来的研究趋势进行了展望。

Trajectory data is widely used in fields such as intelligent transportation,natural conservation,and epidemic prevention and control.Trajectory similarity measurement is one of the most complex and time-consuming operations in trajectory query analysis,and is the research foundation in the field of trajectory data management.This paper first divides trajectory similarity measurement methods into time sensitive and non time sensitive based on whether they are sensitive to time information,and introduces a new trajectory similarity measurement method based on semantics and deep learning;then,a comprehensive comparative analysis is conducted on each type of measurement method,and their respective advantages and disadvantages are given;finally,prospects for future research trends in this field are presented.