Monitoring traffic through GPS phones
The CCIT (California Center for Innovative Transportation), the department of civil and environmental engineering at Berkeley, the Department of Transport of California (Caltrans) and Nokia have joined, February 8 last, to perform an experiment scale monitoring real-time traffic highway through GPS phones.
The idea of the experiment was to develop a test situation, scale, with hundreds of vehicles equipped with GPS cell phones moving on a restricted portion of highway in the San Francisco Bay.
On February 8th, CCIT, Caltrans, Nokia, and UC Berkeley’s Department of Civil and Environmental Engineering collaborated to conduct an unprecedented experiment in the area of traffic monitoring.
The event was enormously successful, both technically and logistically. Business plans of most major cellular phone manufacturers such as Nokia include embedding GPS in all manufactured cell phones within less than 18 months. Thus, a high penetration rate of GPS-equipped travelers on freeways is expected in the near future. This has major implications for the traffic engineering community, which currently monitors traffic using mostly fixed sensors such as cameras and loop detectors, or location specific sensors such as FasTrak or EZ-pass transponders.
CCIT would like to thank the entire team for their work on this labor intensive project. CCIT would especially like to thank Mobile Century's project manager Ali Mortazavi and principal investigator Alex Bayen for their excellent work.
The difficulty of the experiment was to quantify the amount of data necessary and sufficient to obtain reliable data on traffic flow while minimizing communications and maximizing the confidentiality of communications. During 7 hours, one hundred vehicles driven by students at Berkeley have turned over a portion of motorway 8 miles (12.9 km), sending their addresses and their speed through their mobile phones to remote servers.
The data were encrypted using an asymmetric encryption whose public key is the central server. The information is still sent as a first step to a first server anonymiseur, who withdrew all provenance information in the data. Thus, the first server can decrypt the data, and the central server can not draw a particular person. To minimize the quantity of data to send, the phone software regularly asks the server what is the frequency of transmission in which the system needs, depending on the current speed and the presence of other vehicles in the same sector.
After treatment and smoothing of data through the flow models of vehicles developed by Berkeley, the result can calculate average speeds and travel times between several points. The classic roadside sensors are less reactive, less resistant to breakdowns, very localized and more expensive to maintain that system. This system is reliable enough when 5% of vehicles are equipped with such systems for locating and sending data. As GPS systems will probably become widespread in mobile phones in the coming years, this model has a good potential for growth.
This experience, which was attended by the scientific service of San Francisco, was a successful collaboration between public and private institutions, organization and demonstrating the potential of these technologies.
Sources:
http://www.calccit.org/
http://www.citris-uc.org
Category:
Comments
Post new comment