Abstract: Clustering trajectories as a whole can miss common sub-trajectories.Discovering common sub-trajectories is very useful in many applications，such as weather forecast and traffic control.A new method T-CLUS is proposed to clustering trajectories，which partitions a trajectory into line segments，and then the augmented cluster-ordering of the line segments is generated，finally cluster structure of sub-trajectories is identified by means of reachability plot.Experimental results demonstrate that T-CLUS is scalable and accurate to discover common sub-trajectories from a trajectory database.

摘要： 将运动轨迹作为整体聚类会丢失相似子轨迹段，而相似子轨迹段在实际应用中用处很大，如天气预报、交通控制等。提出一种新方法T-CLUS进行轨迹聚类，先将长轨迹分割成许多较短的直子段，再产生子段的增广聚类顺序，最后根据可达性图识别子轨迹聚类结构，得到子轨迹运动模式。实验结果表明T-CLUS方法在从轨迹数据库中发现相似子轨迹上是可伸缩的和正确的。