A shape-based map matching approach for geographic transferability of discriminative subtrajectories

This paper addresses the challenge of map matching and geographic transferability in trajectory analysis. Existing methods often face limitations tied to specific coordinates or road networks. In response, we propose GASM, a shape-based map matching method that relies solely on trajectory shapes, irrespective of geographic origin. GASM introduces a symbolic road network representation, facilitating efficient searches based solely on trajectory shapes. Our experimentation, spanning over 5,000 km of roads, demonstrates GASM's ability to accurately position trajectories with an impressive accuracy exceeding 90%. Notably, GASM stands as the first in the literature to perform shape-based symbolic map matching without prior knowledge of the geographic region.