Connected Self-Driving Cars Can Have A Positive Impact On Traffic Flow

Researchers from the University of Cambridge suggest that a fleet of self-driving autos capable of interacting with each other can have a positive impact on overall traffic flow by at least 35%.

An experiment was set up with some autopilot cars driving on a track with multiple lanes. The researchers then examined what happened when one of the vehicles stopped moving.

When the vehicles were not moving cooperatively, any vehicles following the stopped one were forced to halt or to move slowly and had to wait for a gap. This is something that would happen in real life. Behind the stopped car, a line quickly formed, and the traffic stream was slowed down.

In any case, when all of the vehicles were communicating and driving cooperatively, right at the moment one vehicle stopped moving, a signal was sent to the rest of the fleet. Vehicles driving in the other lanes that were closest to the stopped vehicle moved slower in order to let vehicles on the inner lane pass the stopped vehicle quickly without having to slow down or stop.

Co-author Michael He from St John's College who developed the algorithms said: "Self-driving cars could fix lots of problems relating to driving in cities, but there needs to be a way for them to work together."

Also, when a vehicle driven by a human was on the “lane” with the self-driving vehicles and started driving around aggressively, the rest of the fleet could avoid the driver, thus improving safety.

“If distinctive car makers are all willing to develop their own self-driving cars with their own program, those vehicles really need to have the ability to interact with each other in an effective way”, said co-author Nicholas Hyldmar from Downing College who was responsible for the experiment's hardware.

For the research, in order to make the vehicle interact through Wi-Fi, the researchers used motion capture sensors and a Raspberry Pi.

The discoveries were introduced at the International Conference on Robotics and Automation (ICRA) in Montreal, Canada.