Over the last few decades, urban commutes have emerged as a genuine nightmare. In the U.S., motorists in the Washington, D.C. area waste an average of 67 hours a year stuck in traffic, while drivers in Los Angeles and the San Francisco Bay area squander 61 hours a year that way, according to a Bloomberg and Texas A&M Transportation Institute report. Meanwhile, in cities such as Mumbai, Beijing, and Sao Paulo, commutes can stretch to several hours per day and traffic backups can extend more than 100 miles (160 kilometers).
The result? Wasted time and fuel, higher levels of pollution, frayed nerves, safety concerns, and inefficiencies that rock the foundation of modern economies. A study by Texas A&M University found that in the U.S. alone, $27 billion worth of time and fuel was wasted in 2011 as a consequence of all traffic-related delays, resulting in economic loss exceeding $121 billion. In fact, despite telecommuting, home offices, and flexible work hours, the problem is growing worse. "We are facing extremely serious obstacles and a growing list of problems, says Dominique Bonte, a vice president and practice director at market research firm ABI Research.
Traffic engineers traditionally have focused on installing timed traffic lights and road sensors to detect vehicle flows, and developing algorithms to manage automobiles and infrastructure more effectively. While these systems help alleviate some congestion, they remain relatively unsophisticated. As a result, transportation engineers and computer scientists are now designing a smarter network of roads, traffic lights, signs, and vehicles. As Sei Kato, a managing consultant at IBM Global Business Services and an advisor for the City of Kyoto in Japan, puts it: "There is a need to maximize existing infrastructure."
Transportation engineers and computer scientists are designing a smarter network of roads, traffic lights, signs, and vehicles.
Make no mistake, smart traffic networks promise to redefine driving and urban environments by introducing smart systems that can react in real time to changing conditions. Yet, the technology is not without risks. Security researchers say these networks are vulnerable to hacks and attacks that could wreak havoc and endanger public safety. "Successful hacks could seriously disturb traffic and cause widespread problems," notes Andre Weimerskirch, a research scientist at the University of Michigan Transportation Research Institute.
The march toward traffic automation is nothing new. In 1914, the City of Cleveland, OH, installed the world's first electric traffic signal, which ushered in a level of automation that ultimately transformed transportation. Over the ensuing years, green-yellow-red signals became ubiquitous, providing a reliable way to manage traffic flow and pedestrians. Engineers typically timed these stoplights to match traffic volumes. In recent years, they also have turned to sensors embedded in pavement and, in some cases, video cameras and other technologies, to determine when a light should switch to green and how long it should stay green.
Along the way, cities have introduced increasingly sophisticated traffic networks. For example, the City of Los Angeles switched on its Automated Traffic Surveillance and Control System (ATSAC) system in preparation for the 1984 Summer Olympics. ATSAC has undergone significant upgrades since then and it now provides remote control capabilities for all of the city's 4,500-plus traffic lights. The systemwhich relies on sensors and closed-circuit camerasreduces commute times by about 12%, according to the city. In addition, it is connected to rapid transit and bus networks so, for example, if a bus is running behind schedule, an engineer can alter the traffic light schedule so it can catch up.
Singapore has also forged ahead with traffic monitoring and control systems. It has already tested a system that relies on induction loops in roadways to monitor traffic. Later this year, working with IBM, the city will unveil a system that plugs in data from induction loops, video cameras, and taxi GPS systems to provide more accurate data about traffic patterns. The technology is also being used in Lyon and Montpellier, France, where it creates thousands of data points. The traffic control algorithm relies on a number of key factors, including road category, density of traffic on the road, speed limits along with traffic data and incident data, says Laura Wynter, director of the IBM Research Collaboratory in Singapore.
In the U.S., about 79% of vehicle detection systems currently use video or induction loopsalso known as in-ground loopsto track the movement of vehicles on roads. Some communities also rely on microwave, radar, and ultrasonic sensors. Typically, a mix of technologies provides deeper insights into movements and behavior. According to a 2003 survey by the Institute of Transportation Engineers, 62% of traffic intersections across the U.S. operate in an interconnected fashion. However, these stoplights typically function in specific traffic corridors or as part of a network, and they lack the full spectrum of data to revolutionize traffic management.
"The problem with many of these systems is that they provide some benefits, and they help manage traffic better, but they do not tap into the efficiencies that are possible with a fully networked infrastructure and vehicles," explains Bonte, who is responsible for ABI Research's global telematics, location, and navigation market coverage, as well as for the firm's coverage of smart cities. "It's one thing to monitor traffic and make adjustments; it's another to change the underlying dynamics of the overall system. There is a need for more highly networked and advanced technology and software that can work across an entire city and interface with other systems."
This notion appeals to Carolina Osorio, an assistant professor of civil and environmental engineering at the Massachusetts Institute of Technology. She and colleagues are developing more advanced technology to address traffic management on a broader scale. One of the problems with current traffic systems is that when engineers adjust one set of stoplights, it has a ripple effect on other intersectionsalong with driver behavior, she notes. For example, if wait times increase along a particular route, drivers will likely seek "alternative routes that feature fewer red lights." Too often, engineers wind up shifting problems from one set of motorists to another.
Osorio is currently testing a monitoring system and algorithms in Lausanne, Switzerland and New York City. She approaches the problem through the use of traffic simulators that analyze the behavior of drivers in response to changing conditions. In the Lausanne simulation, the approach has so far netted a 22% reduction in travel time compared to commercial traffic-light timing software. It encompasses 12,000 individual drivers, approaches the task as a wholetaking into account the motion and movement of the entire traffic system, rather than focusing on a collection of individual motorists displaying distinct and complex behaviors. The initial simulation used 150 runs taking place during a typical 5 P.M. to 6 P.M. weekday period. City engineers use the simulation data to make changes to actual lights and other infrastructure.
"There is a need for a more highly networked and advanced technology and software that can work across an entire city and interface with other systems."
While some cities already use high-resolution simulators, known as microscopic simulators, these systems often become too computationally intensive. Osorio's "homogenous" method approaches the accuracy of high-resolution models with the computational efficiency of a low-resolution model. Instead of merely reducing commute times, the algorithm provides data to improve fuel consumption and suggest where to place vehicle-sharing hubs and parking. It incorporates inputs such as a given time period, green signal phase durations, behavioral factors, and more.
Practical results occur at the intersection of connected infrastructure, vehicles, and algorithms, Osorio says. The traffic management systems she and others are developing attempt to embed a high-resolution description of both traveler behavior and network systems, including "prevailing traffic control strategies in the network," she notes. Key factors include: travel behavior, traffic flow theory, simulation, nonlinear optimization, and traffic control methodologiesincluding the response of motorists. "These strategies can lead to improvements in [traffic management] at a large scale, including the scale of a full city."
Of course, the problem is that drivers in Moscow are different than those in Auckland, New Zealand, or San Francisco, so there is no "one size fits all" algorithm. What is more, as conditions in a city changesay a new high-rise apartment building openstraffic flow patterns may change with them. As a result, researchers increasingly examine two primary components: capturing data that identifies behavior and trends, while also accounting for deviations from trends.
Traffic: The Next Generation
Over the next few years, as the Internet of Things and connected machines take hold, traffic management is likely to speed forward. Bonte says no infrastructure will likely be left untouched; using sensors, RFID, and other connected technologies, it is possible to transform everything from traffic lights and road signs to street and vehicle design. "Virtually everything can become part of a connected infrastructure," he explains. This, in turn, will introduce other changes, including whether signs and other infrastructure (often referred to as "road furniture") are actually necessary. Instead, a traffic network may rely on virtual traffic lights and cues delivered to drivers and pedestrians via smartphones, he says.
In addition, vehicle-to-vehicle (V2V) communications would allow autonomous or semiautonomous vehicles to connect to each other in real time, and slide the dial from a centrally managed system to a more distributed computing and communications model. With traffic data connected to stoplights and vehicles communicating with each other, "You quickly move past the incremental gains afforded by smarter traffic signals and unlock huge gains," Bonte says. Within this scenario, cars can "platoon" a half a meter or less apart, thus stretching the capacity of existing roads and infrastructure. These networks will adjust traffic lights, subways, light rail systems, and bus routes across an entire city in real time in order to optimize conditions. Moreover, "You create a much safer environment. The number of collisions goes down to near zero," Bonte explains.
The framework for this future is already taking shape. In Pittsburgh, PA, traffic engineers working with Carnegie Mellon University installed 49 smart signals in 2014. These stoplights use radar rather than street sensors to monitor the flow of traffic, and they communicate with each other in real time. Officials estimate the system will trim wait times in traffic by 42%, and total travel times by 24%. In Utah, the state's Department of Transportation has installed a computerized traffic network that uses radar and video to manage traffic lights. The result? Only 28% of vehicles now hit a red light at any given intersection.
Meanwhile, technology companies such as IBM and Cisco Systems are developing software to run smart cities. At Kyoto University and IBM's Tokyo Research Laboratory, for instance, developers have introduced software that allows city officials to run simulations and adjust traffic systems in real time to respond to different variables, such as weather, sporting and entertainment events, and pedestrian flows. One simulation conducted by researchers at Carnegie Mellon University resulted in a 60% improvement in traffic flow in Porto, Portugal. Not surprisingly, smart vehicles connected to these networks could fuel further gains.
Practical results occur at the intersection of connected infrastructure, vehicles, and algorithms.
For now, Kato says one of the biggest technical challenges for traffic engineers and computer scientists revolves around analyzing and understanding "macroscopic behavior." Current simulations can do a reasonably good job of understanding behavior involved with lane changes, merging traffic, and congestion, "but many problems remain when it comes to understanding route modeling and demand modeling. More sophisticated models and algorithms are now under development based on cognitive and big data approaches," he explains.
There also is a need to address potential security and hacking risks associated with connected vehicles and traffic networks. Researchers have already demonstrated the ability to commandeer locks, brakes, and the steering wheel for existing automobiles. Moreover, in 2006, during a labor dispute, two traffic engineers in Los Angeles broke into the city's transportation computer and clogged traffic at key intersections in the city for several days. Although experts say they are not aware of widespread system hacks to date, it is clear that connected and automated systems create new and bigger risks.
In fact, Cesar Cerrudo, an Argentinian security researcher, has identified vulnerabilities in traffic control systems in 40 U.S. cities and in nine other countries. He reported these systems lack key encryption and authentication, and it is possible to sniff out and intercept clear text and data running over 802.15.4 wireless networks. Breaking into these systems required only a few basic tools that cost less than $100, he noted in a blog post in April 2014. "These [incidents] could cause real issues, even deadly ones, by causing accidents or blocking ambulances, firefighters, or police cars going to an emergency call," he explained.
Nevertheless, smarter transportation networks are taking shape and the development of smart vehicles is speeding forward. Bonte says these initiatives will pick up steam in the years ahead and, in many cases, will fold into broader smart city initiatives that seek to reduce waste, pollution, and energy demands.
Says Kato: "Automobiles brought us rapid transportation and improvements in transportation efficiency. On the other hand, they introduced problems such as chronic traffic jams, noise, air pollution, and safety concerns. Smarter traffic networks help cities reduce the impact of these problems."
Ghena, B., Beyer, W., Hillaker, A., Pevarnek, J., and Halderman, J.A.
Green Lights Forever: Analyzing the Security of Traffic Infrastructure, Electrical Engineering and Computer Science Department, University of Michigan, 8th USENIX Workshop on Offensive Technologies, August 2014
Hacking U.S. (and U.K., Australia, France, etc.) Traffic Control Systems, April 2014 blog.ioactive.com/2014/04/hacking-us-and-uk-australia-france-etc.html.
Zimmermann, T., Wirtz, H., Punal, O., & Wehrle, K.
Analyzing Metropolitan-Area Networking within Public Transportation Systems for Smart City Applications. In 6th International Conference on New Technologies, Mobility and Security 2014, (p. 15), IEEE, March 2014
©2015 ACM 0001-0782/15/01
Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and full citation on the first page. Copyright for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers, or to redistribute to lists, requires prior specific permission and/or fee. Request permission to publish from [email protected] or fax (212) 869-0481.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2015 ACM, Inc.