ABSTRACT
Locating and navigation devices now are ubiquitous. Huge amounts of geo-referenced spatial location data and moving object trajectory data are being generated at everincreasing rates. Patterns discovered from these data are valuable in understanding human mobility and facilitating traffic mitigation. The state-of-the-art techniques in managing such data are based on spatial databases and moving object databases to index and query geometrical coordinates directly. However, as most of human movements are constrained by built infrastructures and road networks, an alternative approach to matching the location and trajectory data with infrastructure data for subsequent processing (Richter et al. 2012) can potentially be more efficient from a computing perspective. For example, once a global positioning system (GPS) trajectory is transformed into a sequence of road segments, it will be possible to apply well-studied frequent sequence mining (FSM) algorithms (Agrawal and Srikant 1995) to identify popular routes for different groups of people at different scales for different purposes, for example, ridesharing, hailing taxies for riders, making more profits for drivers, and understanding functional zones in metropolitan areas to facilitate city planning and traffic mitigation.
