Trending: After-party at Jeff Bezos’ D.C. mansion attracts Bill Gates, Ivanka Trump and other notable guests
Zhibin Li, Kris Henrickson and Yinhai Wang from the University of Washington's Department of Civil and Environmental Engineering are developing a way to use the Wi-Fi and Bluetooth signals from bus passengers in order to better understand bus use. Photo: Lisa Stiffler.
Zhibin Li, Yinhai Wang and Kristian Henrickson (left to right) from the University of Washington’s Department of Civil and Environmental Engineering are developing a way to use the Wi-Fi and Bluetooth signals from bus passengers’ mobile devices in order to better understand bus use. Photo: Lisa Stiffler.

Transportation officials are forever struggling to figure out how best to meet bus riders’ needs within their limited budgets. One of their biggest challenges has been determining where and when people are getting on and off buses, whether they’re transferring to other routes and how long their ride is taking.

The trouble is there hasn’t been an affordable, reliable way to get that data — until now.

University of Washington researchers are doing experiments in which they collect the Wi-Fi and Bluetooth signals sent from passengers’ cell phones and other mobile devices to create a clearer picture of transit use. In an experiment conducted this past spring, they were able to use these signals to determine when and where passengers were getting on and off UW buses. They presented their study last week at the annual meeting of the Transportation Research Board in Washington, D.C.

“This will fill a very important gap in transit operations,” said Yinhai Wang, a UW professor of civil and environmental engineering and director of the Pacific Northwest Transportation Consortium, or PacTrans, who worked on the project.

“The previous way of [doing] data collection is very challenging,” he said. “Basically you hand out questionnaires to people.” The surveys are expensive to administer and many riders fail to complete and return them.

With this new approach, Wi-Fi and Bluetooth signals are collected by a smartphone installed at the front of a bus. The phone runs an app to gather the signals, which contain media access control (MAC) addresses that are unique to each device. The smartphone also has GPS locating abilities, so the researchers can determine where as well as when the signal starts and stops.

Bus from the University of Washington fleet. Photo: UW.
Bus from the University of Washington fleet. Photo: UW.

The whole setup cost less than $60 per bus. That included the smartphone, which was an older, less expensive model, as well as the wiring and power adapter so it could be plugged into the bus’s electrical system.

In addition to determining where and when people get on and off the bus, the new approach could one day be used to figure out what percent of riders are regular commuters or use transit infrequently, and if they’re transferring to other buses. That could be a boon to systems like King County Metro Transit, a public transportation agency that provides more than 120 million bus rides annually in the greater Seattle area.

But there is a potential catch. Not everyone is comfortable with the idea of agencies collecting and tracking their bus riding habits through their cell phones.

“People are concerned about this kind of thing,” said Kristian Henrickson, a UW civil and environmental engineering doctoral student and research assistant who helped with the study, which was also coauthored by doctoral student Matthew Dunlap and research associate Zhibin Li, also from the UW Department of Civil and Environmental Engineering.

The team is eager to ease the public’s worries and has a plan for doing so. They have developed an algorithm to turn the MAC addresses into different, unique sets of numbers that they can assign to individual riders, but don’t identify a specific mobile device in a way anyone else would recognize. In contrast to other traditional methods for tracking traffic and highway use, including cameras that snap images of license plates, the converted MAC addresses should present fewer privacy concerns, they said.

“If it’s done in the right way,” said Henrickson, “and we actively communicate the benefit and make very transparent how we do it, it’s not a problem.”

The initial tests were done on buses run by the UW that travel between the main Seattle campus in the University District neighborhood and Harborview Medical Center, covering a total of eight stops.

This map shows the detection of Bluetooth signals collected from one of the UW buses. The gaps in the map were caused when the bus traveled through a tunnel and the GPS signal was lost. Image: UW.
This map shows the detection of Bluetooth signals collected from one of the UW buses. The gaps in the signal toward the bottom of the map were caused when the bus traveled through a tunnel and the GPS signal was lost. Image: UW.

Josh Kavanagh, UW Director of Transportation Services, was eager to partner with the researchers in order to support their work and gain some new insights into rider behavior.

“Our ability to plan is only as good as the inputs coming into it, and this gives us a wonderful tool,” Kavanagh said. He has shared the research with King County Metro, who he described as having “a lot of interest in it.”

Metro currently can collect some data from the ORCA cards that many passengers use to pay for their rides. But the cards will only show when a rider got on to the bus, and not where they got off. And if the bus is off schedule, it won’t give accurate information for where they boarded, the UW’s Wang explained.

The UW engineers found that one of the difficulties in collecting useful data was the noise from signals coming from non-passengers, including people in cars near the buses, pedestrians and people in buildings along the route. The scientists were able to filter out the non-rider signals based on criteria such as whether the signal was too short or long lived, and if it started or ended at locations that weren’t bus stops.

There are limitations to the data. Based on other research, the engineers expect that roughly 10-20 percent of passengers are carrying mobile devices with their Wi-Fi and Bluetooth turned on.

And some people — particularly older and lower-income riders — might not have mobile devices, or devices with Bluetooth or Wi-Fi. Other passengers could have multiple devices, causing some population segments to be underrepresented while others are overrepresented. The researchers said they could correct for some of that bias by supplementing with questionnaires or by doing headcounts to determine what percent of riders are being detected.

For Kavanagh’s part, he thinks the new strategy holds a lot of potential.

“There is not a lot of hardware investment or labor time setting the equipment up, and that makes it a very cost effective tool,” he said. “We are just beginning to understand how we can use the technology for understanding consumer behavior. This will provide us value for many years to come.”

Like what you're reading? Subscribe to GeekWire's free newsletters to catch every headline

Comments

Job Listings on GeekWork

Find more jobs on GeekWork. Employers, post a job here.