One of the biggest debates among the autonomous vehicle community is when we’ll start to realize the long-term benefits of self-driving cars (like safer roads) on a widespread basis. Many people believe we won’t begin to reap these benefits until the majority of vehicles on our roads are driverless, leaving us to deal with unpredictable human drivers that continue to cause accidents and traffic jams that can stretch for miles.
Contrary to these popular beliefs, a new study from the University of Illinois at Urbana-Champaign has concluded the addition of even a few autonomous vehicles can help ease congestion on roadways. Demonstrations of “phantom traffic jams,” where vehicles drive in a circle, simulate the effects of a single vehicle slowing down on a traffic-filled road. During these simulations, one vehicle applies its brakes for no reason, causing a slowdown that resonates through the rest of the cars in the circle.
Researchers demonstrated how placing even just one autonomous vehicle in one of these circular traffic simulations can ease the effects from a phantom traffic jam. The experiment’s results showed that having an autonomous vehicle intelligently control its speed when a phantom traffic jam starts, it’s possible to mitigate the amount of braking that occurs further down the line. The presence of just one driverless vehicle reduces the standard deviation in speed of all cars in the simulation by 50 percent. The amount of sharp hits to the brakes even drops from around nine to 2.5 (and sometimes zero) per vehicle for every kilometer traveled.
A car uses more fuel when the vehicle slows down and then gets back up to speed. Fuel consumption is something that autonomous vehicles also reduce when present in these phantom traffic jam scenarios. University researchers even determined that normal vehicles involved in the simulation saved up to 40 percent more fuel. The fact these improvements can occur when a single autonomous vehicle is inserted into a flow of 20 other cars speaks volumes of this concept’s potential.
