Posted on November 20, 2008
By Emil Venere, Purdue University News Service
Filed under Purdue News, 101 views

WEST LAFAYETTE, Ind. — Purdue University engineering students used a new method to track traffic by monitoring Bluetooth signals from cell phones and other wireless devices carried by football fans as they drove home from a recent game with Penn State.

The method uses the pervasive Bluetooth signals to constantly update how long it takes vehicles and pedestrians to travel from one point to another. Harnessing the wireless signals represents a potentially low-cost leap in technology to provide information for everything from the speed of the morning commute to the sluggishness of airport security lines, said Darcy Bullock, professor of civil engineering.

Bullock led the class of graduate students who used the new approach for the Penn State traffic study during the Oct. 4 home game in West Lafayette.

The method picks up the identifying “addresses” from Bluetooth devices in consumer electronics. Because each device has its own distinct digital signature, its travel time can be tracked by detectors installed at intersections or along highways and other locations.

The students used a special antenna to identify 1,520 Bluetooth addresses in the crowd of more than 57,000 fans in attendance. The students then used 13 tracking stations to monitor the Bluetooth signals as fans drove home from the game at Purdue’s Ross-Ade Stadium along two routes leading to Interstate 65: a 4.2-mile southern route and a 5.2-mile northern route.

Data from the study documented which route had the fastest travel time.

“We found that the postgame travel time along the southern route was up to 28 minutes, but the travel time along the northern route was only 12 to 14 minutes, even though the northern route is one mile longer,” said graduate student Mary Martchouk. “And there was also much less variability in travel time on the northern route, much less congestion.”

The researchers concluded that the Bluetooth method was far more effective than two more conventional traffic-tracking techniques. Those techniques typically use spotters with camcorders to manually record individual license plate numbers on cars or “probe vehicles” equipped with a GPS tracker.

“Bluetooth matching is much easier to do than license-plate matching, which is very labor intensive,” said graduate student Ronald Davis. “And you can gather much more data using the Bluetooth signals than you can with both license-plate matching and GPS probe vehicles.”

The Bluetooth method is less invasive than license-plate matching, which identifies the person being tracked. People cannot be identified by their Bluetooth signals, Bullock said.

The students presented their findings on Nov. 18 during an annual dinner of local members of the Institute of Transportation Engineers and will be hosting a national webinar, scheduled for 12:30 p.m. Dec. 3, that is open to the public.