Urban Travel Reliability Analysis with Consumer GPS Data

摘要: This project explored the use of commercially available consumer global positioning system (GPS) data in travel reliability studies. Specifically, two TomTom traffic products, namely MultiNet and TrafficStats, which produce respectively time-of-day link speed profile time statistics on selected routes their segments, are considered. The quality these commercial products is evaluated using GCM data, a publicly sensor database. There several findings from evaluation First, generate average estimations for highway segments that match those reasonably well. However, estimated times about 10%-15% lower than data. Second, as benchmark, arterial streets found to be severely underestimated Third, daily profiles obtained do not very well most expressways examined this study. In general, tend underestimate segments. Fourth, usefulness provided by TrafficStats raise more concerns. percentile route reports seems spread out too much have variance high realistic cases. These also reported segment statistics, especially terms variance. conclusion research team utility product has considered with cautions, given its price. provide reliable source roads database since it based real observation instead statistical estimation. Reliable routing experiments show results changed significantly after used distributions streets. authors speculate main reason change were underestimated. As result, many arterial-based paths certain risk-taking preferences no longer attractive because they become when used. Based VNET platform, developed RIDEA application, integrates functions access visualize TomTom/GCM databases guidance. provides prototype support future software development analysis.