轨迹预测技术的现状及发展综述  

Review of the status and development of trajectory prediction technology

作　　者：张志远 倪国新[1] 徐艳国[1] Zhang Zhiyuan;Ni Guoxin;Xu Yanguo(Nanjing Research Institute of Electronics Technology,Nanjing 210000,China)

机构地区：[1]南京电子技术研究所,南京210000

出　　处：《电子测量技术》2020年第13期111-116,共6页Electronic Measurement Technology

摘　　要：位置获取技术的进步产生了大量的轨迹数据,这些数据蕴含了丰富的信息。在过去一段时间里,人们提出了很多处理、挖掘轨迹数据的技术,推动其快速发展,应用更加广泛。为帮助人们更迅速了解轨迹数据挖掘这一领域。对轨迹数据挖掘的主要研究进行了系统的综述,首先介绍了轨迹数据的相关概念及其预处理方法,然后阐述了几种模式挖掘的方法,比如序列模式、伴随模式挖掘等。其次,介绍了现有的轨迹预测模型并且讨论分析了各个方法的优缺点,最后说明了轨迹预测的潜在挑战和未来的发展方向。The progress of location acquisition technology has produced a large amount of trajectory data, which contains a wealth of information. In the past, many techniques for processing and mining track data have been proposed, which have promoted its rapid development and wider application. The purpose of this paper is to help people understand the field of trajectory data mining more quickly. In this paper, the main research on trajectory data mining is systematically summarized. Firstly, the relevant concepts and preprocessing methods of trajectory data are introduced, and then several methods of pattern mining are described, such as sequential pattern mining and moving together pattern mining. Secondly, the existing trajectory prediction models are introduced and the advantages and disadvantages of each method are discussed and analyzed. Finally, the potential challenges and future development direction of trajectory prediction are illustrated.

关 键 词：轨迹预测 数据挖掘 预测算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]