Improving Transit Performance with Automated Data Collection Systems
Nigel Wilson relies on the London Underground Oyster card, but not for his daily commute. Wilson uses data from transit fare cards to research and improve transit services in London. Wilson, professor of Civil and Environmental Engineering at MIT, was at Volpe, The National Transportation Systems Center, to participate in a Straight from the Source talk about automated data collection systems like the Oyster card, the Massachusetts Bay Transportation Authority CharlieCard, and Hong Kong Octopus Card to improve public transit performance.
Wilson examines data collected by a variety of automated systems to understand passenger movements. The process requires creative use of data from systems that were not created for the purpose of counting passenger trips or the time a person spends waiting for a transit vehicle.
"A lot of what we've done in London is focus on how we manage reliability from the customer's perspective, rather than the operator's perspective," said Wilson. His team has been studying London transit for eight years. "These automated systems enable us to estimate reliability for the first time," he explained, because they provide a very large and complete sample set.
"On buses, with headways of ten minutes or less, we can look at waiting time, in-vehicle time, and attributes for individuals, and aggregate all that to represent overall performance," he explained. On the Underground, knowing that the headway—the time between buses or trains on the same line—is three to four minutes, he said, "We can use tap in and tap out times to estimate the journey time distribution between a particular origin-destination pair…this might be the time the customer would have to allow for, because of unreliability of the service."
Wilson's Origin-Destination Prototype Tool
Wilson has also been creating an origin-destination matrix for transit users to understand how the system is actually being used. "The objective is to estimate, at the root and network levels, the way the system is being used. We're particularly interested in linked journeys," said Wilson. "You chain together successive taps from a particular card." For buses it's trickier because "you don't have the actual location where a tap occurred." In London there are 6 million Oyster card taps per day. "We're getting a very good base picture from which to expand to the whole network," said Wilson.
The team's most challenging project is estimating the number of train passengers using train weighing systems. "If a train is full, as opposed to half-full, that may affect our [operations] decisions," explained Wilson. "For customers, it's important to know when the next train will be there and whether it will be crowded."
Wilson's research is producing tangible results: London will give up bus passenger surveys in favor of implementing Wilson's prototype origin-destination tool. And reliability can now be measured with data from automatic data collection systems. While train loads cannot currently be estimated for the Underground, Wilson stated firmly that it will be quite realistic in the foreseeable future.
To learn more, view the video highlights from Dr. Wilson's talk.