Given the plethora of GPS and location-based services, queries over trajectories have recently received much attention. In this paper we examine trajectory joins over streaming spatiotemporal data. Given a stream of spatiotemporal trajectories created by monitored moving objects, the outcome of a Continuous Spatiotemporal Trajectory Join (CSTJ) query is the set of objects in the stream, which have shown similar behavior over a query-specified time interval, relative to the current timestamp. We propose a novel indexing scheme for streaming spatiotemporal data and develop algorithms for CSTJ evaluation, which utilize the proposed indexing scheme and effectively reduce the computation cost and I/O operations. Finally, we present a thorough experimental evaluation of the proposed indexing structure and algorithms.
Download Full PDF Version (Non-Commercial Use)