“I’ve learned to do a lot of deep breathing, and if I’m really stressed I try to put on classical music,” says Walter Neary, who commutes about 40 miles each way from his home south of Tacoma to the UW.
“Things tend to snarl between Federal Way and Southcenter, slow at Boeing Field, and at the first Seattle exit it crawls,” he says. “A lot of people make that drive white-knuckled—you see them, fighting for every inch. It can get bad.”
How bad? Last November, Seattle was ranked third among U.S. cities for the amount of time drivers spend stuck in traffic. The intersection of Interstates 5 and 90 was designated as one of the nation’s most congested.
But little does Neary or the other hundreds of thousands of drivers know that a UW engineering project is making their life a little easier.
A UW computer program using “fuzzy logic” can deliver exacting answers about the ever-changing status of area freeways to help move traffic more efficiently, according to Deirdre Meldrum, UW associate professor of electrical engineering and principal investigator on the project. The system uses a computer algorithm that can balance conflicting objectives and anticipate problems before they become critical, resulting in smooth, continuous control to prevent or delay congestion.
In tests last spring, the ramp-metering algorithm produced an 8.2 percent reduction in freeway congestion on a stretch of I-90 from Issaquah to Bellevue—significant enough to be noticeable on a day-to-day basis. It also prevented a bottleneck near the Eastgate on-ramp, a task at which the old system failed. Fuzzy logic now dictates traffic flow on all 126 of the metered freeway on-ramps in the Seattle area.
Fuzzy logic is just one of the latest in a long string of innovations at the UW designed to make the going easier for freeway travelers.
And it’s caught the attention of traffic engineers across the country, according to Meldrum and Cynthia Taylor, a UW research engineer who designed the algorithm that makes fuzzy logic tick.
“We’ve received a lot of positive feedback,” Meldrum says. “There’s been quite a bit of interest and a lot of people have made inquiries, wanting to know more.”
Fuzzy logic is just one of the latest in a long string of innovations at the UW designed to make the going easier for freeway travelers, work that has helped put Seattle on the map as a center for cutting-edge traffic research, according to Mark Hallenbeck, director of the Washington State Transportation Center.
“The state Department of Transportation’s (DOT) information systems are the best in the country,” says Hallenbeck, whose center seeks to match state agencies’ needs with researchers’ expertise. “And the University has either directly or indirectly built the vast majority of that. The DOT folks have written the code for some of the Web sites that are physically running, but it was based on research by UW faculty—mostly engineering faculty—and it’s really state-of-the-art work.”
That work includes Web sites that display diagrams of Seattle’s major freeways with color codes for current levels of traffic congestion, sites that show bus locations and predict arrival times for delayed vehicles, a cable television channel dedicated to real-time information on area traffic conditions, and analytical tools that give transportation engineers and policy makers hard data to consider when balancing transportation issues.
Dan Dailey is a good person to talk to about using the Information Superhighway to aid travelers on the old-fashioned asphalt freeway.
In fact, if you use “Web” and “traffic” in the same sentence at the UW, you’d likely be talking about Dailey, an electrical engineering research professor and director of the school’s Intelligent Transportation Systems program who has been a key player in bringing Internet savvy to the traffic equation. His initial interests didn’t lie in traffic, however—he worked with bubbles, specifically modeling the propagation of bubbles in nuclear reactors. The topics aren’t as disparate as one might think.
Using data from the sniffer and digital maps, Dailey and his students figured out how to position the buses in real time.
“From an analytical or mathematical point of view, the propagation of traffic and the propagation of bubbles look kind of similar,” Dailey says in his office in the new Electrical Engineering/Computer Science Building on campus, still dressed in cycling regalia from his morning commute. “You have different sizes and different speeds, they’re passing one another, and you have a hard time getting a good look at the medium. You can only sense it occasionally and externally.”
Dailey put his bubble expertise to use when he received funding from the DOT to understand traffic behavior as a random process. In 1990, he and his students created a computer program that shows how fast cars are going on the freeway. Other projects quickly followed. When King County Metro Transit installed a bus location system in the early ’90s, Dailey was intrigued. He obtained permission to access the data—not an easy task because it was on a closed system designed by a defense contractor.
“We sent a student over with a ‘sniffer,’ ” a computer program designed to covertly listen in and gather information from a system, “to eavesdrop on the network at their building downtown,” Dailey recalls. “I don’t think they thought we’d be able to do much with it.”
Using data from the sniffer and digital maps, Dailey and his students figured out how to position the buses in real time.
“It took about a year, and we went back to them and showed them what we had,” he says. “They said, ‘Oh my God, we’re not sure we want everyone to know that.’ ”
The application, called Busview, allows computer users to track their Metro bus in real time via a simplified map that runs on their computer screen. People found that interesting, according to Dailey, but what they really wanted to know was when their bus would arrive.
“That’s a lot harder,” he says. “Predicting the future is a much more complicated proposition.”
The resulting project, Transit Watch, placed computer monitors at selected transit centers to display arrival times for the buses. The screens are similar to airport monitors, a format already familiar to most people. The next step involved putting the information where people really want it, according to Dailey: “At their desk.”
“Wireless Web devices are one of the hottest things in the world right now. They're really neat, but they're only neat if you can actually get something useful out of them.”
Mark Hallenbeck, director of the Washington State Transportation Center
MyBus is a Java program that shows arrival times for buses at eight Seattle locations. That way, commuters can get travel information directly from their computer screens. While the programs won’t make buses run on schedule, they do decrease anxiety and make commuting a more attractive option, Dailey maintains.
Another success has been Traffic TV, a program featuring real-time traffic information broadcast on the UW’s experimental channel (76 on TCI in Seattle) twice a day during morning and evening commutes. Sort of a Weather Channel for traffic, the broadcast provides real-time traffic updates through a combination of maps and video shots of current commuting conditions.
Underlying all those projects is a central system, designed by Dailey, that gathers and processes information gleaned directly from more than 3,000 sensors embedded in area freeways. The Intelligent Transportation Systems Information Backbone then makes that data available in a useable format via the Internet. This makes Seattle unique, according to Hallenbeck, and will be a critical factor in fueling the coming wave of portable traffic help devices.
“Wireless Web devices are one of the hottest things in the world right now,” Hallenbeck says. “They’re really neat, but they’re only neat if you can actually get something useful out of them. Wouldn’t you like to know how to get to work, and how to get home, and where your bus is? To do that, you need two things: a little wireless device that communicates and the information that you want to communicate to it.”
The ITS Backbone provides that information in an easily accessible format, allowing broad participation in the development of new traffic technology, says Dailey. It’s also one factor—along with heavy traffic and a governmental push for alternate transportation research—that makes Seattle ground zero for cutting edge-traffic research.
“Seattle is a great place to examine these sorts of issues,” says Dailey. In fact, Seattle was one of four cities selected as a demonstration site for a multi-million dollar federal pilot project to explore innovative transportation alternatives. The local version of the project is an $18 million undertaking known as Smart Trek. “Nationally, we’re seen as a good testing ground because of those things that converge here.”
The numbers show that, believe it or not, in some areas congestion around the University District has shown slight improvements.
Dailey’s current work includes an effort to use cameras to gather information for traffic modeling. If successful, the method would provide a way of gathering traffic data in communities that lack Seattle’s extensive network of highway sensors. It’s a fantastically complex proposition because the cameras can be manipulated by operators, changing the view.
“But that’s what makes it interesting,” Dailey says. “That’s what good science is about.”
Hallenbeck also gathers and processes traffic information. He describes his work as similar to Dailey’s but from an opposite approach and for a different audience.
“He’s the real-time person, interested in getting information to people about what’s happening now,” Hallenbeck says. “I’m interested in the past. I’m the forensics guy. The information we provide is something policy makers can use.”
Hallenbeck’s latest project takes traffic data and converts it into images and numbers that describe the various dimensions of congestion: time (how long are you stuck?), severity (how badly are you stuck?), geography (how far are you stuck?) and temporal (how long does the problem last?).
The graphics provide a way to gauge whether transportation strategies are doing what they were intended to do.
For example, the numbers show that, believe it or not, in some areas congestion around the University District has shown slight improvements, Hallenbeck says.
A comparison of data from 1995 to data taken in 1997 shows a distinct drop in how often State Route 520 becomes congested during the morning commute to the University District—a commute heavily influenced by UW employees. Significant congestion occurs one or two days a month less often in 1997 than it did in 1995. Crowded conditions still routinely occur from about 6:30 a.m. to 9:30 a.m., but this time frame has not expanded since 1995.
In sharp contrast, the afternoon commute in the same direction (coming into Seattle), a flow heavily influenced by people who work on the Eastside and live in Seattle, shows marked increases in both how often significant congestion occurs and what times during the day that congestion occurs. Heavy afternoon congestion now routinely occurs as early as 2 p.m. and can last until nearly 8 p.m. There is a 20 percent greater chance of finding heavy congestion at 3:30 p.m. in 1997 than there was in 1995.
Programs such as U-Pass, which provides UW employees with a heavily subsidized bus pass, are a major reason why the morning commute into Seattle on 520 is bucking the otherwise growing trend of congestion in the region, Hallenbeck says.
The analyses also bring insight into current hotly contested transportation issues, such as High Occupancy Vehicle (HOV) lanes. Critics of the lanes, which are supposed to be used only by vehicles with two or more persons, say the lanes should be opened to all to make maximum use of highway space. Supporters say the lanes encourage carpooling and mass transit and are a necessary part of the area’s overall transportation strategy.
“We can show that the HOV lanes in this area have been a huge success during peak hours,” Hallenbeck says. “We can also show that if you take those people in HOV lanes out of the buses and put them in cars, it will be an utter disaster.”
According to Hallenbeck’s figures, during peak commute hours freeway HOV lanes move more people in fewer vehicles than the general-purpose lanes next to them.
During the afternoon commute north on I-5, for example, the HOV lane carries just under 16,000 people, while the general-purpose lane next to it carries about 8,000. If the HOV lanes were opened, many people would likely opt to drive their own vehicles. The data on I-5 indicate that if just 7 percent of the people now carpooling or riding the bus decide to drive their own cars, congestion would increase, even though all vehicles could use the extra capacity available in the HOV lane.
Hallenbeck’s next report is due out in a couple months.
Congestion solutions don’t come easy for a freeway system as large and variable as Seattle’s. Projections call for a continued flow of people into the Pacific Northwest, with the Puget Sound population increasing as much as 50 percent by 2020. More people will mean more cars and potentially more gridlock. Can technology rescue us from highway hell? That depends, researchers say. The real question is whether commuters and residents will opt to participate.
The bottom line, according to Dailey, is to get information out so people can make informed choices.
“We all live in traffic,” he says. “The better we understand it, the better we can make good decisions to cope with it.”
Mark Hallenbeck, director of the Washington State Transportation Center, suspects that Seattle, and other major U.S. cities, will follow patterns seen in many European and Asian cities with rampant congestion:
“No country in the world has a lack of congestion as a result of mass transit,” Hallenbeck says. However, it will be a means to move quickly in an increasingly crowded community.
Dan Dailey, UW electrical engineering research professor, says that he sees “a lot of promise in human-powered transportation, although I have an obvious bias in that direction.” More and safer pathways for bicycles are a must, he admits, before a lot of people will consider that option.
And, Hallenbeck adds, roadways will eventually be expanded. But that process will be slow, expensive and painful. And it won’t solve the problem.
“You can’t build your way out of congestion,” Dailey says. “It’s been proven over and over again—there’s no ambiguity about that. No city in the world has ever done it.”
They agree that the key to dealing with congestion will be changing how we do things. In light of such events as the passage of I-695, which slashed taxes that provided transportation funding, change may not come easily. Or quickly.
Says Dailey: “I think we’re going to drown in our traffic a little longer before we learn how to swim.”