2011/05/19 by Gautam S. Thakur, Pan Hui, Thakur, Gautam S. +5
Social Sciences · Computer Science · Engineering · #Human Mobility and Location-Based Analysis #Video Surveillance and Tracking Methods #Vehicular Ad Hoc Networks (VANETs)
paper · pdf · doi:10.48550/arxiv.1105.4151
Realistic modeling of vehicular mobility has been particularly challenging\ndue to a lack of large libraries of measurements in the research community. In\nthis paper we introduce a novel method for large-scale monitoring, analysis,\nand identification of spatio-temporal models for vehicular mobility using the\nfreely available online webcams in cities across the globe. We collect\nvehicular mobility traces from 2,700 traffic webcams in 10 different cities for\nseveral months and generate a mobility dataset of 7.5 Terabytes consisting of\n125 million of images. To the best of our knowl- edge, this is the largest data\nset ever used in such study. To process and analyze this data, we propose an\nefficient and scalable algorithm to estimate traffic density based on\nbackground image subtraction. Initial results show that at least 82% of\nindividual cameras with less than 5% deviation from four cities follow\nLoglogistic distribution and also 94% cameras from Toronto follow gamma\ndistribution. The aggregate results from each city also demonstrate that Log-\nLogistic and gamma distribution pass the KS-test with 95% confidence.\nFurthermore, many of the camera traces exhibit long range dependence, with\nself-similarity evident in the aggregates of traffic (per city). We believe our\nnovel data collection method and dataset provide a much needed contribution to\nthe research community for realistic modeling of vehicular networks and\nmobility.\n