PROJECT TITLE :

Continuous Trajectory Similarity Search for Online Outlier Detection

ABSTRACT:

Within the scope of this paper, we investigate a novel form of trajectory similarity search by situating it within the framework of continuous query processing. In order to identify any online detours that may have been taken by an object that is in motion from point S to point D while adhering to a reference route Tr, we keep track of the trajectory similarity between the reference route and the current partial route at each timestamp. We propose a partial trajectory similarity measure as a means of bridging the gap left by existing trajectory distance measures, which are unable to adequately capture the difference in trajectory that exists between a partial route and a complete route. In particular, we determine the minimum distance between each of the possible routes that can be extended from the partial route to reach the destination d and then enumerate all of those routes. Calculating deviations in Euclidean space and on road networks are both part of our investigation. In Euclidean space, we are able to directly infer the optimal future path that has the shortest distance between two points on the trajectory. Regarding road networks, we suggest an effective expansion algorithm that is accompanied by a collection of pruning rules. In addition, we propose effective strategies for incremental processing in order to facilitate continuous query processing for objects that are constantly moving. The effectiveness of our query processing algorithms is validated by the results of our experiments, which are carried out on a variety of real datasets.