GPS Data Analysis: Travel Time Estimation in Sparse Data Environments

Human urban trajectory data collected from GPS-enabled mobile devices or vehicles are extensively used in several applications such as urban planning, traffic management, location based services. We describe several GPS trajectory data analysis problems that come up in such settings including the problem of accurate, automatic map creation from GPS trajectories, and the problem of identifying travel patterns. We focus on the problem of estimating, in real time, the vehicle travel time along a given path. We describe a novel technique for combining historical data information and real-time data observations that are collected from a small number of floating cars.