Results from experiment on traffic flow

Compositional Systems Lab

Results from experiment on traffic flow

Research team including Dan Work, Miles Churchill, Nate Hamilton, Shumo Cui, R'mani Haulcy, Rahul Bhadani, Rafael Stern, Benjamin Seibold, Matt Bunting, Maria Laura Delle Monache, and Jonathan Sprinkle

In July 2016, the lab took part in a groundbreaking experiment in which we showed that self-driving cars can benefit the drivers of cars around them, by controlling their velocity to dissipate traffic waves. In addition to improving traffic flow, we show up to 40% improvement in fuel economy, and 90%+ reduction in braking events during controlled flow, compared to the span of time when the traffic waves were pronoonced. The drivers in our experiment received no special training other than instructions to keep up with the car in front of them. 

Ours is the first result in the world to show this on real vehicles, with real human drivers, not through simulation. This is the beginning of a new era of control "within the flow" for complex emergent phenomena like those frequently seen in traffic. 

If you want to read more about it, a link to a preprint of our technical paper on arXiv is located here. The work is funded by NSF Cyber-Physical Systems, and a joint effort with Illinois, Temple, and Rutgers.

YouTube video:

Press release from Illinois:


Self-driving cars experiment demonstrates dramatic improvements in traffic flow