| 本文已被：浏览 3759次 下载 5401次 |
码上扫一扫！ |
|
|
| 基于偏好的个性化路网匹配算法 |
|
高需1,2, 武延军2, 郭黎敏3, 丁治明3, 陈军成3
|
|
1.中国科学院大学, 北京 100049;2.中国科学院 软件研究所 协同创新中心, 北京 100190;3.北京工业大学 计算机学院, 北京 100124
|
|
| 摘要: |
| 定位技术的普遍应用，使得随时随地获取个人位置成为可能，进一步推动了基于位置的服务等新型应用的发展，产生了海量轨迹数据.精确的路网匹配对提高这些新型应用的服务质量具有重要的研究意义，然而受众多因素的影响，大部分轨迹的采样率较低，比如由签到类应用或低功耗设备生成的低采样轨迹，给路网匹配带来了巨大的挑战.研究基于偏好的个性化路网匹配（driving preference based personalized map-matching，简称DPMM），提出了在动态道路交通网络中的用户驾驶偏好模型.基于该模型，提出了两阶段路网匹配算法：局部匹配搜索用户最可能采用的几条局部Skyline路径；设计了全局匹配的动态规划算法，该算法返回在用户驾驶偏好下最可能的多条全局路径作为最终匹配结果.实验结果充分表明，该方法是有效的和高效的，具有一定的使用价值. |
| 关键词: 时空数据 轨迹 路网匹配 多目标优化 Skyline路径 动态规划 |
| DOI：10.13328/j.cnki.jos.005297 |
| 分类号: |
| 基金项目:国家自然科学基金（61402449，91546111）；中国科学院战略性科技先导专项课题（XDA06010600）；北京市教委重点项目（KZ201610005009） |
|
| Personalized Map-Matching Algorithm Based on Driving Preference |
|
GAO Xu1,2, WU Yan-Jun2, GUO Li-Min3, DING Zhi-Ming3, CHEN Jun-Cheng3
|
|
1.University of Chinese Academy of Sciences, Beijing 100049, China;2.Collaborative Innovation Center, Institute of Software, The Chinese Academy of Sciences, Beijing 100190, China;3.School of Computer, Beijing University of Technology, Beijing 100124, China
|
| Abstract: |
| With the increasing proliferation of position technologies, there comes huge volumes of trajectory data, which are used in many modern applications such as path planning and location based services. Accurate road network matching can improve the service quality of these new applications. However, the low sampling trajectories bring a major challenge for map-matching. This paper studies the problem of matching individual law-sampling trajectory to a dynamic multi-criteria road network based on user's driving preferences. First, a driving preference model in the dynamic road traffic network is proposed. Based on this model, a two-stage map-matching algorithm is developed. While local matching searches multiple local likely Skyline paths, a global matching dynamic programming algorithm is designed and the most probable k global paths are selected as the matching result. Experiments show that the proposed method is effective and efficient. |
| Key words: spatio-temporal data trajectory map-matching multi-objective optimization Skyline path dynamic programming |
