高速公路出入口运动车辆轨迹分层聚类算法

发布时间：2018-02-20 16:44

  本文关键词： 交通运输系统工程 高速公路出入口 轨迹分析 改进Hausdorff距离 聚类算法　出处：《吉林大学学报(工学版)》2017年06期 　论文类型：期刊论文

【摘要】：为了提高对高速公路出入口车辆运动行为的理解和分析水平,根据出入口车辆运动轨迹的时空特征,提出了一种运动轨迹层次聚类算法。结合出入口轨迹方向一致、长短不一的特点,提出采用改进Hausdorff距离来衡量轨迹间的相似性。建立了改进模糊C均值轨迹分层聚类算法,首先根据轨迹的空间几何位置进行路径聚类,然后根据车辆的速度信息对已有路径聚类进一步聚类获得具有时空区分度的最终结果。真实高速公路出入口的试验结果表明:本文提出的轨迹聚类算法对于场景固定运动行为模式不仅具有较强的适用性,而且能够保障聚类结果的准确性和可靠性。
[Abstract]:In order to improve the understanding and analysis of the movement behavior of vehicles at the entrance and exit of freeway, a hierarchical clustering algorithm of motion trajectory is proposed according to the temporal and spatial characteristics of the moving track of the vehicle at the entrance and exit, and the direction of the entry and exit trajectory is the same. An improved Hausdorff distance is proposed to measure the similarity of trajectories, and an improved fuzzy C-means trajectory clustering algorithm is established. Firstly, the path clustering is carried out according to the spatial geometric position of the trajectory. Then according to the speed information of the vehicle, the existing path clustering is further clustered to obtain the final result with space-time discrimination. The experimental results of the real freeway entrance and exit show that the trajectory clustering algorithm proposed in this paper is useful to the scene. The fixed motion behavior mode is not only applicable, Moreover, it can ensure the accuracy and reliability of clustering results.
【作者单位】： 同济大学道路与交通工程教育部重点试验室;
【基金】：“863”国家高技术研究发展计划项目(2013AA12A206)
【分类号】：TP311.13;U491
，

本文编号：1519475