Arriving on time: estimating travel time distributions on large-scale road networks

摘要: Most optimal routing problems focus on minimizing travel time or distance traveled. Oftentimes, a more useful objective is to maximize the probability of on-time arrival, which requires statistical distributions times, rather than just mean values. We propose method estimate large-scale road networks, using probe vehicle data collected from GPS. present framework that works with large input data, and scales linearly size network. Leveraging planar topology graph, computes efficiently correlations between neighboring streets. First, raw traces are compressed into pairs times number stops for each traversed segment `stop-and-go' algorithm developed this work. The then used as training path model, couples Markov model along Gaussian random field. Finally, scalable inference algorithms obtaining composite MM-GMRF model. illustrate accuracy scalability our 505,000 link network spanning San Francisco Bay Area.