The Performance of a Crowdsourced Transportation Information System
Anthony Tomasic, John L Zimmerman, Charles Garrod, Yun Huang, Terence Nip, Aaron M. Steinfeld · Transportation Research Board 94th Annual MeetingTransportation Research Board · 2015
Crowdsourced mobile sensing systems provide a counterpoint to the idea of fully automated sensing systems by transferring some or all of the sensing duties to the end users. Humans can easily sense in some ways that are impossible for machines to sense, leading to hybrid crowdsourced-automated systems. However, this transfer of sensing to humans comes with design trade-offs in terms of the sparsity of the sensed information and human entry errors. To better understand these design trade-offs, the authors developed a real-time arrival information system for a local transit agency that crowdsources the location of transit vehicles by having riders share location traces from their smart phones. The authors then deployed the system and measured the public’s use of it for 10 months, gathering data on 296,283 interactive sessions. Analysis shows that relying on users can produce very sparse information for the whole system but much better information relative to the times and places that users access the service. Reflecting on this deployment, the authors conclude that crowdsourced-sensing systems perform well when (i) the value of an observation persists over time, (ii) many people make observations, and (iii) observations would be difficult or expensive to sense with an automated sensing system.