⇧ [VIDÉO] You may also like this affiliate content (after ads)
A very small proportion of autonomous vehicles present on the roads may be enough to influence all traffic and improve its smoothness. This conclusion was reached by researchers from different universities who studied the impact of their presence on traffic.
Traffic flow is also one of the keys to improving the energy efficiency of vehicles. The scientists, who are still testing their hypothesis, have published an interim report on arXiv in which they argue that self-driving cars can, through their presence on the roads, increase the energy efficiency of all drivers, whether they drive or not. autonomous vehicle.
This impact is due to what they call “waves” or “waves”: “Traffic flow exhibits various instabilities at high density and this is called the congestion phase. Such instability can turn into constant waves of stops and starts and move upstream in the flow of traffic, ”they explain in their article. “Over the past decades, many efforts have been made to explain the origin of these waves and their potential relationship with the non-linearity of the system. Waves can be caused by network characteristics (bottlenecks, ramps, etc.) as well as driver behavior (changing lanes, hard braking, etc.). These waves cause inefficient movement and increased fuel consumption.”
So the team first ran a simulation of a very simplified layout: a two-lane ring road. They were able to see that having autonomous cars can reduce fuel consumption by 40% for all traffic, even if only 5% of drivers use these vehicles.
They then conducted tests outside of the simulators on a road that met the same conditions. Their findings are that autonomous vehicles not only circulate more smoothly, but also somehow “force” other drivers to do the same, as they expect more, thus avoiding hard braking. “Traffic flow is very specific in the sense that one person can have a global impact on the dynamics of the entire flow. This is found in both micro and macro models and can be understood with a simple example: one person can be a bottleneck and thus affect the traffic of the entire system,” the scientists explain.
Two stages of testing pending
Now a team of scientists is developing a simulator that takes into account more complex road configurations. They suggest that the smoothness benefits will be less impressive, but still present. To then test their results, they plan to conduct a real-world test with about 100 self-driving car drivers.
The idea was to implement two phases of the test. During the first, autonomous cars will be able to act individually. In the second step, they would like to observe the results of “collaboration” between vehicles, in which the observations will be shared on a central server in order to smooth out the overall traffic. It is expected that the tests will take place in the vicinity of Nashville, Tennessee (USA), at the end of 2022. It can also be assumed that there may be positive security implications, although researchers are unwilling to make hasty guesses.