Curing congestion: we can now tackle traffic as a complex network problem.
Globally, people average about 60 minutes of travel each day, much of that consumed by stops or slowdowns. Some delays are regular and recurring, others unexpected. They are all more than just a personal inconvenience: human capital is underutilized, freight distribution delayed, meetings missed, fuel wasted, and nerves frayed.
Many important roadways regularly operate near their breaking points, and as populations and economies continue to expand, travel demands will rise and waits lengthen. Stressed transportation systems become less resilient and create cascades of real costs. As freight is slowed, food prices rise, for example; and as travel times exceed acceptable thresholds, certain destinations lose their attraction.
Thanks to new technologies we are now able to observe these systems and congestions effects in real time, across the large, complex scales at which they operate. We are also able to intervene at the same scale, not just locally, an approach that raises the prospect of making congestion a thing of the past.
Transportation system managers can now turn to a combination of sensors, algorithms, and system simulators to predict traffic demands from minute to minute, day to day, and year to year. In a highly instrumented system like New York City's, signal times, tolls, and left-turn permissions can be fine-tuned in real time.
Truly reducing, rather than just managing, congestion requires active intervention to change travelers' behavior and make the most of scarce roadway real estate. Varying tolls throughout the day in response to traffic patterns, offering slot reservations for space in some lanes, and compensating those who turn to alternatives to their own vehicle are one set of options. GPS-enabled smart-phones and other technologies make such strategies much more realistic, since drivers can receive and act on information rapidly. The arrival of autonomous, self-driving vehicles that can safely travel close together should make it even easier to enhance traffic flow (see "Self-Driving Cars," p. 41). With road space limited, over the long term travelers will need to shift to smaller vehicles, public transit, and nonmotorized modes of transport. But by acknowledging the true costs and complexity of road congestion, we can moderate it quite effectively now.
SOLUTIONS * TRAFFIC
Cars on the world's roads, actual and projected 2010 1 billion 2030 2 billion Note: Table made from bar graph. Traffic's 2010 cost in the United States, in lost productivity and wasted fuel $101 billion Actual and projected worldwide road deaths, in thousands 1990 542 2000 723 2010 957 2020 1,204 Note: Table made from bar graph.
Cars that need no human drivers have the potential to increase productivity, unsnarl traffic, and sharply reduce accidents. Such cars could safely travel much closer together on the road, and people could work--or nap--while being whisked along. Two-way data feeds could help them learn the best travel routes while warning them of new obstacles and sending them instructions about how to queue at intersections. Onboard sensors, such as the ones shown here, would work in concert to create a clear picture of the environment that could inform autonomous acceleration, braking, and steering. Self-driving cars are probably a decade from being ready for the mass market, but they work: Google's prototypes have driven more than 300,000 miles without accidents, albeit with humans sitting in the driver's seat just in case.
(1) Cameras Stereo cameras can help the car see lane markings. Visual data can be fused with data from other sensors to help the vehicle avoid obstacles.
(2) Lidar A light detection and ranging system can assemble a very thorough 360[degrees] picture of the surrounding environment.
(3) Millimeter-wave radar High-frequency radar, installed in bumpers or at intersection control units, can detect pedestrians even under poor lighting and background conditions,
(4) Radar Sometimes used today by cruise-control systems to maintain a fixed distance from cars ahead, radar can help maintain the car's position relative to surroundings.
Kara Kockelraan is a professor of transportation engineering at the University of Texas, Austin.
|Printer friendly Cite/link Email Feedback|
|Publication:||MIT Technology Review|
|Date:||Nov 1, 2012|
|Previous Article:||Alzheimer's Ahead: an epidemic of age-related dementia looms.|
|Next Article:||Repeating history: we have fully eradicated only one disease. Let's do it again.|