Interactive Trip Planning Using Activity Trajectories

Abstract

We present an interactive trip planning system called @FINDER which uses an exemplar trajectory query to find the most related top-k spatial-textual trajectories. @FINDER is implemented to support various degrees of user information needs for trip planning. For users with zero knowledge about places to travel, @FINDER provides a heatmap of popular points of interest (POIs) as well as popular activities from a trajectory database. The system helps users quickly explore the places, and helps formulate an exemplar trajectory query, which specifies preferred places to go and activities of interest. Then @FINDER provides efficient query processing of the top-k related spatial-textual trajectories using a new approach to spatial-textual trajectory indexing recently developed at RMIT University. For each of the top-k results found in the form of a set of POIs and activities, @FINDER further computes the optimal route (in term of the travel time) covering all of the POIs, and returns an album to the user. Lastly, users can further interact with @FINDER by adding or deleting POIs/activities in the original exemplar query, and the system will update the results in a timely manner.