Urban Map Inference by Pervasive Vehicular Sensing Systems with Complementary Mobility
summary
coMap infers road geometry and road class from four vehicle fleets (18,000 taxis, 10,000 private cars, 6,000 trucks, 14,000 buses) by weighting GPS points by vehicle type and using the complementary mobility features entropy to detect intersections. A graph-based map sketching stage links weighted points to form segments (15.9% improvement over KDE and trajectory-clustering baselines), then a learning-based map painting stage uses map-matched trips and trip features to classify road categories (74.58% average accuracy).
AI-written summary · about
related talks
15mTowards automated mapping of pedestrian path networks at scale
15mComputational Mobility: Learning, Prediction, and Algorithms
19mHow we leverage AI to build great maps at massive scale
06mExtending Coverage of Stationary Sensing Systems with Mobile Sensing Systems for Human Mobility ...
28m"Where, When, Why, and How: Directi
19mChallenges & Opportunities for Building Crowdsourced Mapping Services for Autonomous Driving
25mMap transit systems on transportation networks using GTFS Data and Infomap
13mUtilizing Vehicle Trajectory Data to Characterize Performance Measures of a Highway Corridors