Why Does Traffic Jam?
Traffic jams caused by tollbooths, accidents and rubbernecking rank high on the list of ways to ruin a summer vacation, but for sheer exasperation, it's hard to beat congestion that has no visible cause. You're cruising at 55 and suddenly you're bumper to bumper. You inch along for a mile. Then, daylight: just as suddenly you're back up to 55 (or more), never having seen even so much as a salacious billboard to explain the slowdown. Once such delays were merely annoying. Now their cost is mounting. Hours lost to traffic jams exceed 2.7 billion a year; idling cars and stop-and-go driving add to air pollution and guzzle gas. Before highway designers can prevent traffic jams, though, scientists have to understand where they come from. "We have equations to explain traffic," says Michael Cassidy of the University of California, Berkeley, "but many are wrong. Only a new approach -- looking at traffic as wave motions over time and space -- might reveal the cause of bottlenecks."
Scientists like Cassidy are racing to keep up with technology-loving engineers who have decided that the solution to traffic jams is the Intelligent Vehicle/Highway System. With IVHS's on-board computers and in-the-roadbed sensors, drivers would receive information about a tie-up via a dashboard display; they could choose an alternate route. Last week a congressional subcommittee held a hearing at which Transportation Department officials argued the merits of IVHS, which is receiving $310 million in 1994 for pilot programs and research. While enthusiasts envision rush hours as smooth as new pavement, some experts are skeptical. No wonder: it was only after overloaded highways were widened -- at huge expense -- that engineers realized that made traffic worse (drivers flocked to the nice new road). That's why theorist Gordon Newell of Berkeley calls smart highways "a pile of nonsense. The schemes for re-routing people will just cause congestion somewhere else."
So traffic scientists are figuring out why smart highways may be just as doomed to congestion as stupid ones. The answer may be to watch the waves, looking for the similarity between traffic and fluids. "Traffic behaves like waves in a river," says Cassidy. Waves of congestion and decongestion pass through traffic much as sound waves travel through air. Sound waves increase the density of air; traffic waves increase the density of cars. From a helicopter, one can actually see waves of congestion propagate steadily upstream.
Congestion can be triggered by the slightest disruption to smooth flow and how drivers react to it. In heavy traffic, for instance, drivers are worried about smashing the bumper in front of them. If traffic slows just a little -- at a curve, say -- the driver behind may overcompensate and brake more than needed to keep the same spacing between him and the next car. The driver next in line does the same, and a shock wave of deceleration propagates "until someone has to stop," explains Fred Hall of McMaster University in Ontario. The shock wave eases just as mysteriously as it began: drivers see space in front of them and accelerate to close the gap, on down the line. When drivers pass a bottleneck, "you get a reverse shock wave," says Hall. Those near the blockage stay stopped a few extra seconds so they have daylight in front of them. The result of that caution: the wave of decongestion propagates only slowly to the back of the line.
Building on this work, Berkeley's Cassidy suspects that similar driver behavior can cause jams a mile or more past a merge -- exactly where intuition says congestion should ease up. "Some [researchers] think that bottlenecks form simply because the on-ramp produces excess capacity," says Cassidy. "But I suspect it may have a lot to do with driver behavior." As everyone knows, drivers lose their manhood if they allow another car to slip in front of them; to avoid this, they accelerate at the first glimpse of that dreaded "merging traffic" sign, according to Cassidy's preliminary work. Once they have passed the merge without losing their place in line, they ease off. This deceleration starts perhaps a mile past the merge point, quite the opposite of traffic experts' assumption that the effect of merges is felt in back of them. If Cassidy is right, traffic flow can be understood as a unique combination of wave mechanics . . . and testosterone poisoning.
Scientists suspect that treating traffic as sound waves traveling through air might reveal how mysterious bottlenecks arise and then dissipate.
Cars approaching a bottleneck have not yet hit the back end of the propagating shock wave.
The lead cars decelerate, rounding a curve. Lagging cars slow even more, sending a shock wave to the back of the line of cars. They slow to a crawl.
Vehicles pass through the bottleneck and speed up; a wave of decongestion soon travels downstream.