Latest revision as of 02:52, 2 February 2021

Abstract

We introduce our vision for mining fine-grained urban traffic knowledge from mobile sensing, especially GPS location traces. Beyond characterizing human mobility patterns and measuring traffic congestion, we show how mobile sensing can also reveal details such as intersection performance statistics that are useful for optimizing the timing of a traffic signal. Realizing such applications requires co-designing privacy protection algorithms and novel traffic modeling techniques so that the needs for privacy preserving and traffic modeling can be simultaneously satisfied. We explore privacy algorithms based on the virtual trip lines (VTL) concept to regulate where and when the mobile data should be collected. The traffic modeling techniques feature an integration of traffic principles and learning/optimization techniques. The proposed methods are illustrated using two case studies for extracting traffic knowledge for urban signalized intersection.

Original document

The different versions of the original document can be found in:

• http://www.winlab.rutgers.edu/%7Egruteser/papers/urbTraffic_v12.pdf

• https://www.cs.uic.edu/~urbcomp2012/papers/UrbComp2012_Paper13_Ban.pdf,

https://core.ac.uk/display/103554559,

https://dblp.uni-trier.de/db/conf/kdd/urbcomp2012.html#BanG12,

https://dl.acm.org/ft_gateway.cfm?id=2346514&ftid=1272427&dwn=1,

https://doi.org/10.1145/2346496.2346514,

http://csce.uark.edu/~tingxiny/courses/5013sp14/reading/Ban2012TFU2346496.2346514.pdf,

http://www.cs.uic.edu/~urbcomp2012/papers/UrbComp2012_Paper13_Ban.pdf,

https://academic.microsoft.com/#/detail/2010335233

• http://dl.acm.org/ft_gateway.cfm?id=2346514&ftid=1272427&dwn=1,

http://dx.doi.org/10.1145/2346496.2346514