Taming Traffic with Your Phone: The Mobile Millennium Project

by Gordy Slack


Cell phone users download free software that automatically and anonymously contributes their position and velocity data to a central location.

Traffic in the San Francisco Bay Area is bad. But if the early results from the Mobile Millennium project are any indication, our ability to navigate through that traffic is about to get a lot better.

The idea behind Mobile Millennium is simple. Cell phone users download free software that automatically and anonymously contributes their position and velocity data to a central location. In exchange for this raw data, the phone owner gets a map-based, real-time view of traffic flow all over the Bay Area that can help them navigate around traffic jams and find the most viable
alternative routes.

UC Berkeley Civil Engineering Assistant Professor Alexandre Bayen, in collaboration with Nokia, Navteq, the California and Federal Departments of Transportation, and with support from CITRIS, launched the Mobile Millennium pilot project in November 2008. The project uses GPS-equipped cell phones to provide real-time traffic information all over the Bay Area and is hosted by the California Center for Innovative Transportation (CCIT), a deployment-focused research center at UC
Berkeley’s Institute of Transportation Studies.

Until now, most of our traffic information has come from static sensors built into the pavement and radar devices or cameras mounted on highways and bridges. While these create a network of traffic information (see traffic.com) on the main arteries traveled by commuters, they are mute about less traveled routes not wired with the expensive equipment.  

“We hope to eventually provide traffic data for areas currently not covered by the traffic monitoring infrastructure, such as arterial roads frontage roads, rural roads, expressways, and so on,” says Bayen. “GPS can bring traffic information everywhere where there are mobile phones.”

Within a few weeks of the project’s launch, more than 3,000 users downloaded the software. “At first we saw use growing organically both from the university in Berkeley and from the Nokia Center in Palo Alto. Since then the rest of the Bay Area has begun to fill in and light up on our maps,” notes Bayen.

Users of the software, which is available at http://traffic.berkeley.edu/, will get maps with all of the current static data already available on http://www.traffic.com, but they will also get the new layer of analyzed data coming from contributors like themselves.


Leaders of the Mobile Millennium partnership cut the ribbon on the software download station that launches a pilot traffic-information
system based on anonymous GPS data from participating members of the
[Thomas West, director of UC Berkeley’s California Center for
Innovative Transportation; Quinn Jacobson, lead scientist at Nokia
Research Center; Harry Voccola, senior vice president of government and
industry relations at NAVTEQ; Randell Iwasaki, chief deputy director of
Caltrans; Alexandre Bayen, UC Berkeley assistant professor of civil and
environmental engineering; and Henry Tirri, Head of Nokia Research Center.]

CCIT expects more dots to be lighting up the data maps soon. “Nokia is producing more than one million phones a day, about 15 phones a second. We are dealing with enormous scales that are able to completely penetrate the transportation network very rapidly,” says Thomas West, the Director of CCIT.

When the software is in use, your phone will display a map with a dot in the middle. That dot then moves wherever you move and automatically scrolls the map. The roads and highways are color-coded for their degree of traffic: green means fast, yellow means slower, red means congested, and black means totally congested.  You can zoom in, zoom out, search, and follow your commute.

Because the system’s algorithms produce speed information, the program can quickly estimate the time a particular trip will take given current traffic conditions.

The team has developed a voice-activated function for hands-free operation.  Because the GPS knows where you are, the program can generate trip-relevant audio messages. If you are headed north from Palo Alto, say, the program may tell you that the Dumbarton Bridge is backed up so you should take the San Mateo Bridge instead. Or that you should get off the freeway at an early exit and take a frontage road. Or even that, given what traffic is typically like between Palo Alto and Berkeley at that time of day, it may be better to wait an hour before you leave at all.

Not only will the new technology cover now blank areas on the map and give users immediate and position-specific information, but it could also save California money. The current system relies on equipment that is both expensive to purchase and costly to maintain. The capital investment for the cell-phone-based system is distributed among its users. Although the service will be run through an information processing office, to be distributed between CCIT, the new CITRIS building on the UC Berkeley campus, and several Nokia and Navteq facilities, it will be much less expensive to maintain.

As security issues are of topmost concern in projects like these, the team is being very careful when it comes to users’ privacy, says West. By the time the project receives data from any user, it is thoroughly anonymized.

“The database that I have is a big bucket of measurements, in pairs, of positions, times, and velocities. It would be virtually impossible to figure out where they were coming from,” says West.

Also, before the data is sent to the system, it is encrypted with the same techniques that bank systems use. So, even if someone is eavesdropping on the line they will not be able to interpret the data, says Bayen.

And finally, so that they do not get private data that is not relevant, the team has devised a “virtual tripline” that turns sampling off when cell phone are outside of areas of interest to the mapping system. So, when you pull off the street and into the donut store parking lot, or into your own driveway, the system stops paying attention to you altogether.

“Scientists only need information on very specific places in the transportation network (highways, highway exchanges, or intersections) so this is where we want to sample people,” says Bayen. The virtual trip line is geographic marker defined by GPS coordinates, a line in the middle of the road that triggers the phone to send an update whenever it is crossed.

Bayen uses the system whenever he is in a car, even when it is not his car. For example, he went to the New York City in November to present the Mobile Millennium project to the Federal Department of Transportation. On his way to the San Francisco airport, the cab he was in was suddenly deadlocked in traffic. He checked his cell phone and was much relieved see that the cab was going to be out of the traffic jam in just a few minutes.