Friday, July 1, 2022
HomeNewsPeer-to-Peer Technology Applied To Traffic Forecasting

Peer-to-Peer Technology Applied To Traffic Forecasting

- advertisement -
- advertisement -

The International Journal of Vehicle Information and Communication systems recently published the research of four UC Irvine graduate students concerning a project capable of clearing Los Angeles Traffic. Tentatively dubbed “Autonet,” a portmanteau of “automatic/automobile” and “Internet,” the program is the brainchild of Trevor Harmon, James Marca, Pete Martini and Raymond Klefstad.
According to a 2006 study by the Bureau of Transportation, there are over 250 million registered cars in the United States. In addition to this, 30 percent of all intersections in the U.S. have no electronic monitoring, which means 70 percent of all traffic lights are not monitored.
Currently, magnetic sensors along roads detect the speed and size of the cars that pass over them and relay that information to the California Department of Transportation (CalTrans). CalTrans then uses that information to assess traffic conditions. While this system has been in place for decades, it is highly inefficient in organizing and disseminating prompt and accurate information to those traveling at high speeds on freeways. Autonet could mitigate that.
Autonet began in 2000 as two of Marca’s basic conundrums in life: the need for a cell phone and the lack of money to obtain one.
“WiFi’s everywhere, so why can’t I just make a call over WiFi?” Marca recalls asking himself as he pondered his plight. “People should just have WiFi in their cars. And if everyone has WiFi in their cars, why can’t cars just talk to other cars?”
Rather than let the thought slip away, Marca, a post-doctoral researcher at UCI’s Institute of Transportation Studies, pitched it to Professor Will Recker, the institute’s head. Recker then brought together the interdisciplinary dream team of Marca, Harmon, now a graduate of UCI’s computer science program, Martini, a staff engineer at UCI, and Klefstad, an assistant adjunct professor and expert in real-time embedded software.
At its core, Autonet functions much as any modern web-capable laptop would by using the same radio frequencies as WiFi signals. The idea is that as cars constantly move past and around each other, they are constantly connected through WiFi. This would be done in conjunction with GPS systems like Garmin or Magellan, most of which can geometrically determine position, the speed limit on each road and the user’s car’s speed. Using this, Autonet can determine if an incident has happened on the road on the principle that if a car loses 80 percent of its speed in relation to the speed limit, it has encountered an obstacle on the road. Surrounding cars then acquire this information and broadcast it everywhere they go, allowing other Autonet-using vehicles to plot different routes.
The proposition was not, however, without its dissidents.
“Previous work had concluded that this sort of technology wouldn’t be possible because of the Doppler effect, but we proved that it could,” Harmon said.
The Doppler effect, in the case of Autonet, would mean that the frequency of WiFi radio waves transmitted changes as the receiver moves away from the source, meaning less information would be transmitted. Yet, when this was tested in real cars, it proved not to be a problem. Even at highway speeds, cars were able to exchange over 3,500 incident reports each.
“This system will be realistic because it’s scalable,” said Wenlong Jin, professor of civil and environmental engineering, who worked on the project’s cost-benefit analysis and mathematical simulations.
While Autonet’s primary goals are to stabilize traffic flow and reduce congestion, its periphery benefits include increased safety, lower fuel consumption and less air pollution. Jin also pointed out the possible cost of integrating Autonet into the current infrastructure and concerns about privacy invasion, but noted that the speed of technological development should mitigate that.
In some ways, the need for Autonet is comparable to the high demand for free downloadable music over peer-to-peer networks like Kazaa. To the average UCI student, the problems Autonet and Kazaa solve are strikingly similar to one another. Whereas those who freely download music loathe the Recording Industry Association of America, their demands that people buy CDs and their relentless threats of lawsuits, commuters loathe heavy traffic and having to rely on often outdated or incorrect traffic reports. Both are slow, expensive and often frustrating processes.
“Say you’re driving down the 405 into L.A. and the traffic report tells you there’s a crash downtown. Well, who cares what’s happening downtown? By the time you get there, the traffic report’s outdated. Autonet would tell you what’s happening along the way to downtown L.A.,” Harmon said.