基于出租车轨迹数据的最优路径规划方法

发布时间：2018-08-20 14:37

【摘要】：针对传统的路径规划算法并不一定能计算得到现实中最优路径的问题,提出一种融合了出租车驾驶经验并以时间为度量的路径规划算法。该算法的实现是将路径规划这个以计算为中心的技术变为以数据为中心的数据驱动挖掘技术。首先,从大量的出租车轨迹数据中提取真实的载人轨迹数据,并将载人轨迹数据匹配到路网数据中;然后,根据地图匹配结果计算路段的访问频次,选取前Top-k个路段作为热点路段;其次,计算热点路段间行车轨迹的相似度,对轨迹进行聚类分析,在路网的基础上构建该k个路段的热点路段图;最后,使用一种改进的A~*算法实现路径规划。实验结果表明,与传统的最短路径规划算法和基于驾驶经验路网分层的路径规划算法相比,所提出的基于热点路段图的路径规划方法有效地缩短规划路径的长度及路径行驶时间,提高路径规划的用时效率。
[Abstract]:In order to solve the problem that the traditional path planning algorithm can not always calculate the optimal path in reality, a path planning algorithm is proposed, which combines taxi driving experience and takes time as the metric. The implementation of the algorithm is to change the path planning, a computation-centric technology, into a data-centric data-driven mining technology. Firstly, the real human trajectory data are extracted from a large number of taxi track data, and the manned trajectory data are matched to the road network data. Then, according to the map matching results, the access frequency of the road section is calculated. Select the former Top-k section as the hot section; secondly, calculate the similarity between the hot road sections, cluster analysis of the trajectory, on the basis of the road network, build the hot road map of the k sections; finally, An improved Awe * algorithm is used to implement path planning. The experimental results show that compared with the traditional shortest path planning algorithm and the path planning algorithm based on driving experience road network stratification, The proposed route planning method based on hot spot road map can effectively shorten the length and travel time of the planning path and improve the time efficiency of the path planning.
【作者单位】：武汉理工大学计算机科学与技术学院;武汉理工大学航运学院;
【基金】：国家自然科学基金资助项目(51579202) 中国博士后基金资助项目(2015T80848)~~
【分类号】：U495

【参考文献】