Cross-scale Spatial Enrichment of Trajectories for Speeding Up Similarity Computing

Abstract

Different types of cross-scale analytics have been applied to spatial enrichment and aggregation of trajectories using geographical context sources as both subjects can present different spatial patterns at different scales. This paper clarifies the taxonomy of different types of “cross scale”. A conceptual framework is then proposed on summarizing the key components for spatial enrichment of trajectories supporting different cross-scale types. Following a workflow guided by the proposed framework, POIs are used for enrichment of GPS waypoints, in a proof-of-concept case study. The preservation of pairwise trajectory similarity between the raw trajectory and the enriched trajectories is investigated. Empirical results show a good preservation while the time on computing the distance/similarity matrix is significantly reduced. This shows the potential for applications relying on an efficient trajectory clustering strategy.