Ph.D. student Kristian Henrickson, Professor Yinhai Wang and alumnus Yegor Malinovskiy (Ph.D. ’13), from left, in the STAR Lab, where they developed newly patented mobile sensing technology that will soon be applied to large-scale transit networks.
The future of gathering data from bus riders is, quite literally, headed in a new direction. Thanks to newly patented technology developed by a team of UW CEE researchers, it is now easier and less expensive than ever to learn about bus rider travel patterns, which can be used to improve bus service.
“We want to reduce people’s transition time,” Professor Yinhai Wang said. “In planning efforts, people’s delay times can be reduced.”
Providing better bus service is an ongoing challenge, as it is often difficult and expensive to collect key travel pattern data from passengers. Current data collection methods in practice consist of manual counts and surveys, which are expensive to administer and have low coverage or completion rates. Some transit agencies also gather data from passengers who pay for rides by swiping smart cards, but this only records trip origin information, not destination details.
The new data collection technology, which costs about $150 per unit, gathers data about where bus riders board and disembark, how much time passes before they board another bus, and which bus stops are the busiest. The technology was developed by a team of UW CEE researchers including Professor Yinhai Wang, alumnus Yegor Malinovskiy (Ph.D. ’13), Ph.D. student Kristian Henrickson, research assistant Matthew Dunlap and research associate Zhibin Li.
“This technology has the potential to give us rich information that can be used to improve service and prioritize investments,” Henrickson said. “Unlike manual data collection or rider surveys, with this approach we can collect data cheaply and continuously on a large number of transit vehicles.”
The technology consists of a mobile device with a GPS unit that is installed inside a bus. When passengers board the bus, the sensor scans Bluetooth and Wi-Fi signals from their smartphones. The technology works by detecting the unique Media Access Control (MAC) address of mobile devices. In order to protect the privacy of passengers, the original MAC addresses are not disclosed. To filter out nearby signals from people waiting at bus stops, or even pedestrians in close proximity to buses, processing algorithms were developed.
The system was first tested in May 2015 on UW hospital bus routes traveling between campus and downtown Seattle. Since then, the software has been improved with new sensor hardware that is lower cost, offers more stable sensing and captures more data.
The primary challenge of the system is that not all passengers, especially the elderly or low-income, may have smartphones or other mobile devices with detectable signals. Even so, the sample size is still large enough to get a reliable picture of travel patterns, said Wang.
“We can’t capture all users because not everyone has a smartphone,” Wang said. “But we can get 10-20 percent for measurement, which is enough.”
The technology will soon be applied to large-scale transit networks. A number of organizations, both in the United States and abroad, are interested in using the technology to improve bus service, including organizations in the United States, Brazil, China and the Netherlands.
Not limited to just buses, the technology also has future applications for gathering data about pedestrians and cars, said Wang. The technology has the potential to replace license plate recognition systems, which cost upwards of $10,000 apiece.
The research is funded by PacTrans and the Washington State Department of Transportation.