MIT researchers have developed a model that can predict traffic jams.
The have come up with a formula of 'phantom' jams, based on the equations that describe detonation waves caused by explosions.
'Phantom' jams are those that occur for no apparent reason, in the absence of an accident or roadworks.
In high-density traffic, a small disturbance, such as a driver braking too hard, becomes amplified as the cars behind react more strongly.
Eventually, many cars behind the initial disturbance, the small incremental increases become a fully-fledged, self-sustaining traffic jam.
The mathematicians call such phantom jams 'jamitons'.
Aslan Kasimov, a lecturer at MIT's department of mathematics, and the team found that jamitons behave like detonation waves in that they have a 'sonic point', a sort of bottleneck, which separates the traffic flow into upstream and downstream areas.
In the downstream section, drivers have no way of telling that there is no external cause for the jam.
"You're stuck in traffic until all of the sudden it just clears," the Telegraph quoted Morris Flynn, the lead author of the paper, as saying.
The MIT team said: "The equations, similar to those used to describe fluid mechanics, model traffic jams as a self-sustaining wave. Variables such as traffic speed and traffic density are used to calculate the conditions under which a jamiton will form and how fast it will spread."
The mathematical model will not help break up jams once they have formed - drivers will simply have to wait them out. But it will allow civic planners to predict where jams are most likely to occur.
This knowledge could help them to determine safe speed limits, as well as identifying potential accident hotspots where the traffic density is greatest.
Further down the line, it could allow engineers to design roads that prevent traffic from building up dangerously, say scientists.