If you drive in the Greater Toronto Area, indeed, in any traffic snake pit, you may be interested in this:
During rush hour, highways typically operate at only 25-per cent capacity;
“Incidents,” such as a stalled vehicle or an accident, cause 60 per cent of traffic gridlock;
A 10-minute “incident” creates approximately 100 minutes of congestion;
There are solutions that could reduce your commute time by as much as 50 per cent, without adding a single new mile of road or transit service.
Baher Abdulhai can list half a dozen ways, off the cuff, that his research could help you get home faster and safer. At 36, this “roads scholar,” as he has been dubbed by the media, is already one of the world’s leading experts in intelligent transportation systems. And he and his researchers are just at the threshold of what they might accomplish.
This past November he launched his brainchild, U of T’s new $4.3-million Intelligent Transportation Systems (ITS) Centre and Testbed, which will turn Toronto’s transportation system into a laboratory for creating and testing high-tech solutions to gridlock. The centre, Canada’s transportation flagship, already rivals the world’s best: the advanced traffic research centre at University of California at Irvine, where Abdulhai earned his PhD in 1996 and derived his inspiration.
Will Recker, who heads UC Irvine’s centre, confessed to TechWeekly that he was “blown away but not surprised” by what Abdulhai has created since U of T recruited him away from the sunny beaches of southern California in 1998.
“When [Abdulhai] walked in the door, he had a clear vision,” says Eric Miller, director of U of T’s Joint Program in Transportation, which oversees the ITS arm. “He had the idea for the lab, and he said this is how he would build it, this is what it would look like and this is what it would accomplish…. The research coming out of this lab will influence the way we do our urban and transportation planning in the city. It will also attract bright young minds to the field, which can only be good for transportation.”
Abdulhai started studying highways in his birthplace – what he calls “the battle zone” – Cairo, Egypt, a city of 14 million with a snarled transportation system. “There are few traffic controls,” says Abdulhai, who completed a master’s degree in civil engineering at Cairo University. “It’s a big, huge mess. You have to force your way through.”
He initially focused on highway infrastructure – designing better highways and roads – which brought him to the University of Alberta to do a PhD in pavement management. There, he heard about the emerging field of ITS, which uses large-scale computer systems to simulate real traffic patterns on real roadways, then tests alternative flow patterns and control mechanisms. For instance, an ITS computer program can simulate the addition of on-ramp traffic lights regulating access to highways and predict the effect on traffic flow. The goal is to apply information technology to maximize capacity on existing roadways and transit systems.
“I had always liked developing computer applications for infrastructure analysis and management, so this [ITS] was immediately fascinating,” says Abdulhai. “It gave me an opportunity to deal with a massive system. In Egypt, our approach [to solving gridlock] was ad hoc, and I knew we could do a lot better if we applied intelligent systems.”
When he got a call from Steve Ritchie, a professor of civil engineering at UC Irvine, to join an ITS research project – “he said he could give me a week to think about it” – Abdulhai went home and told his wife, Nayera, and that night, they packed their bags for California.
For his dissertation, Abdulhai approached traffic flow (or, more often, chaos) as if it were a computer science problem, using pattern recognition to decide if traffic was “normal” or “abnormal.” “Abnormal” happens when traffic on a freeway suddenly slows or stops and remains knotted in gridlock, sometimes for no discernible reason. These interruptions are called “incidents.” It could be a stalled vehicle, an accident, even a motorist braking suddenly, which causes a chain reaction of braking, which in turn can bring traffic to a standstill a mile behind the original “incident.”
As in chaos theory, traffic “incidents” are the butterfly’s beating wings that can wreak havoc miles away. An incident that impedes traffic for just one minute can create 10 minutes of congestion, and incidents account for 60 to 65 per cent of all delays in congested urban areas. The effects are catastrophic but, says Abdulhai, “a computer program that could detect the problem immediately and notify traffic operators to remove the problem would minimize the effect on traffic.”
To detect incidents, he collected data relayed from loop detectors embedded at 500-metre intervals along freeways in San Francisco and Orange County. These loop detectors, found in most highways, indicate how many cars pass over them in any given time period. Abdulhai incorporated these numbers into an algorithm (a set of instructions that tells a computer how to solve a problem). The program worked like a human brain, executing thousands of instructions simultaneously, rather than sequentially, like a normal computer. It operated like a set of eyes on the freeway, observing when traffic slowed down and sped up. If the computer “observed” traffic moving over one detector at a regular rate, but moving slower at a detector behind it, it inferred that an “incident” or accident had occurred in the stretch of roadway in between.
“His research represented a major advance and arguably one of the most significant developments in, and contributions to, the field of freeway incident detection to date,” says Ritchie. “[Abdulhai] has demonstrated an extraordinary ability to forge new strategic partnerships and create win-win relationships with supporters and participants in the ITS Centre at Toronto.”
The ITS Centre and Testbed Abdulhai founded at U of T is like an all-seeing eye on Toronto traffic. It is linked, by fibre-optic cable, to the Ontario Ministry of Transportation and City of Toronto traffic centres, and receives real-time data fed from the detector loops embedded in highways and roadways all over the Greater Toronto Area (GTA). It also receives real-time video feed from the ministry’s 200 traffic cameras situated throughout the GTA. The centre, located in the Sandford Fleming building, looks like a NASA control room, with 30 computer stations facing a wall of 20 video monitors. If a computer signals an “incident” from the detector feed, researchers will be able to call up the video feed to study the congestion.
The centre is still in its first phase: building a massive computer simulation of the GTA transportation system, a virtual Toronto, with highways, city roads, rail and transit routes that are “alive” with virtual vehicles, GO trains and subways moving through the system. The “virtual port land” is complete – you can now watch a simulation of “real” cars clogging up the Spadina entrance to the Gardiner Expressway during a typical evening rush hour.
“The centre will deal with Toronto as a huge experimental ground,” Abdulhai says. “And this city is a great traffic lab because it has extensive freeways, a huge surface network, a comprehensive transit system, and yet it’s nicely congested. It is not using its capacity properly.”
The centre reached its research capacity almost immediately, with 30 graduate students at work on various studies. Prasenjit Roy, who studied civil engineering at the Indian Institute of Technology in Kharagpur, came to U of T to do his master’s degree because of the centre and Abdulhai. He is working directly with Abdulhai to create an “automated incident detection” program, which will not only detect traffic incidents, but also notify proper authorities to deal with them. “He gave me a huge, challenging assignment and I panicked,” says Roy. “But he’s guiding me through it, step by step, to get results little by little.. The results are very promising.”
Roy says his algorithm can detect 92 to 94 per cent of all incidents, and it is not site specific – meaning that the same program will learn traffic patterns at different sites, then adapt itself to detect incidents there. The program will be ready in July or August, and it’s conceivable that the Ministry of Transportation could evaluate and implement it within the next year. “This is cutting-edge research that’s going to be applied,” says Roy, “and I’m very much a part of what we’re developing. It’s very satisfying.”
Sitting behind Roy in the centre is Veljanovska Kostandina, from the University of Skopje in Macedonia. She’s working toward a master’s degree in transportation engineering. Her assignment is developing an algorithm that will “learn” the best way to control freeway access through on-ramp metering. A normal highway lane has a capacity of 2,000 vehicles an hour, but during rush hour, when traffic slows to stop-and- go, that capacity drops to 250 to 400 vehicles an hour. Ramp metering on the Gardiner Expressway, for example, would control the number of vehicles entering it and prevent it from becoming clogged, says Abdulhai. “During rush hour, freeways can drop to about 25 per cent of capacity, exactly when you need it most. You could build four Gardiner Expressways to get that capacity back, or you could introduce ramp meters. If you do proper ramp metering, you can cut travel time by as much as 30 per cent – and that’s a conservative estimate. Waiting on an on-ramp for a few minutes is a small price to pay for a smoother trip home.”
Kostandina is just starting her work and admits she’s also feeling a little overwhelmed. She’s new to the field of traffic engineering, but there’s a cool synergy in the centre – she can turn to colleagues for help, and she’s in demand for her computer knowledge. Still, her research is bound to provoke wide reaction. Toronto experimented with metering 10 years ago, and it failed miserably. Abdulhai contends that the metering was applied incorrectly, and this is where the centre proves most useful: in the “virtual Toronto,” researchers can test a solution before it is ever implemented in the “real Toronto.” “If you do [ramp metering] right, you get the benefits,” says Abdulhai. “If you do it wrong, you get disaster.”
But when it comes to transportation, it can take more than solid scientific research to change attitudes. Undergraduate student Jaime Abraham discovered the speed that 85 per cent of drivers are travelling on Ontario’s Highway 401 is well over the posted limit of 100 km/h and close to 130 km/h. Abdulhai then recommended increasing the speed limit to 130, arguing that the higher speed limit would be safer. “What happens now,” says Abdulhai, “is you get most people driving between 110 and 130 and some people driving at 100, so there’s a spread in speed values, which leads to an increase in road rage, aggressive passing on the right and, really, chaos. We said the speed limit should be set at what 85 per cent of the population drives. Then we should enforce etiquette – if you want to drive slower, you have to drive in the right-hand lane. That would keep traffic moving at the same speed and improve lane discipline.”
Abdulhai’s recommendation provoked reaction from several politicians and police forces and landed him in the media hot seat. “I was on one radio show and this woman called up and said, ‘The professor just wants to drive his Mercedes fast,'” laughs Abdulhai. “First, I wish I had a Mercedes. And second, this is not what I want; this is what 85 per cent of the population is already doing. We’re just asking, is 85 per cent of the population crazy? Or is the law inappropriate? We’re not recommending this for school zones. These are freeways we’re talking about. And we’re not even suggesting something new. In England and Germany, where traffic fatalities are lower, the speed limits are similar to what we’re proposing.”
If and how fast the centre’s research is put into practice depends greatly on its partners. It received funding from 15 public and private-sector partners, including Transport Canada, and it will work closely with the City of Toronto and the Ontario Ministry of Transportation. Abdulhai has already met with the new waterfront redevelopment commission, which is considering proposals to bury, dismantle or reconfigure the Gardiner Expressway. With the testbed’s “virtual portland” complete, the centre can test every option and forecast the effect on local traffic.
“We’re part of the loop, part of the network,” says Abdulhai. “We’re the research arm. We’ll come up with ideas, and they’ll ask us to test their ideas. We’ll provide them with technical answers and quantifiable research. How long it takes for research to be implemented is a political issue. My hope is that politicians move aggressively.” He estimates that if research currently underway in the lab were applied today, it could reduce the average commute time by as much as 50 per cent.
Abdulhai’s specific research area – intelligent management of roadway traffic – is only one of four that will be using the centre. Researchers working on transit, commercial transportation and long-term transportation planning will also use the “virtual Toronto” to explore ways of increasing the effectiveness of other transportation systems in Toronto. Eric Miller, the director of the Joint Program in Transportation and an expert in long-range transportation planning, says his team will be able to build large-scale models to simulate the development of Toronto over the next 10 to 20 years. “Ultimately, we’d say with our model, if you built the waterfront, for example, would they come? We can look beyond traffic to what development, businesses, housing and mix of transportation would be there.
“The increasing motorization of travel – even in Toronto, which has good transit service – does not support sustainable development. The increasing surburbanization of business has increased our use of highways. We will test how transportation can be used as a tool to shape city planning and development.”
Indeed, Miller believes the centre will draw researchers from other fields such as economics, urban planning and psychology. The range of projects they might tackle are limitless. Back at the lab, Abdulhai’s researchers are developing a program to ease “administrative gridlock,” enabling city computers to negotiate with Ministry of Transportation computers to determine optimal rerouting in case of freeway incidents and congestion. He envisions the day when motorists can link into computers at traffic control centres to be “guided” on the best route to travel according to current traffic conditions.
“I find the management of transportation a huge challenge,” says Abdulhai. “You’re faced with a problem that is only going to get worse, given that the growth of populations is far outpacing the growth in transportation systems.”
“ITS is where the action is,” says Miller. “It combines research with a practical application. You’re doing something that has a real and tangible impact on millions of people.”
Margaret Webb (BA 1985 UC) is a freelance writer in Toronto.
By bringing artificial intelligence into chemistry, Prof. Aspuru-Guzik aims to vastly shrink the time it takes to develop new drugs – and almost everything else