A New Method for Finding Similar Patterns in Moving Bodies

An important consideration in similarity-based retrieval of moving object trajectories is the definition of a distance function. The shape based trajectory retrieval is gaining importance due many real time applications. The existing distance functions consider partial matching of the shape of trajectories while retrieving trajectories from database. In this paper we have proposed Angular distance measure as a new shape based distance functions to retrieve trajectories. All the existing distance functions are based on global similarity and there is need for local similarity since it identities sub-trajectory with maximum similarity locally. We have proposed T-BLAST, a local similarity algorithm, which takes into account sub-trajectory matching. Further we have pruned T-BLAST technique by reducing the length of trajectories using Minimum Description Length (MDL) principle. Experimental results show that our proposed Angular distance effectively captures the shape of the trajectories while matching trajectories. T-BLAST accurately and efficiently performs sub-trajectory matching with maximum local similarity. Also, the experimental results show that Pruned T-BLAST algorithm is more efficient compared to T-BLAST.