Self-driving Cars Go with the Flow, and Even Improve It

The idea of autonomous cars is slowly but surely seeping into the consciousness of the general public, even as concepts, one-offs, limited editions and vehicles with some degree of actual self-driving capabilities are being manufactured and marketed.

According to CB Insights, a “tech market intelligent platform” that predicts emerging trends and pinpoints growing industries, “Private companies working in auto tech are attracting record levels of deals and funding, with autonomous driving startups leading the charge.”  CB Insights has identified 44 companies developing roadgoing self-driving vehicles, with projects and partnerships in varying stages, and with names like Audi, Tesla, Nissan/Renault, Nvidia, Hyundai, Honda, Valeo, Volkswagen, Toyota, Alphabet (Waymo) and GM among them. 

But are more cars, of whatever type, what society and its roads and highways, really needs?  What about freeways and city streets already overcrowded, obsolete in some cases, and able to handle traffic at a snail’s pace? Too many cars on the road simultaneously, accidents, and investigating/cleaning up after them, poor road design, and steep grades are among the factors that drive motorists simply crazy, and contribute to unnecessary air pollution.

Self-Driving Cars May Ease Traffic

Now, a research team led by the University of Illinois (U.S.) has evidence that may surprise you (not to mention city planners and highway engineers):

Thanks to autonomous vehicles, traffic jams may become history.

“Our experiments (based on test track results) show that with as few as 5 percent of vehicles being automated and carefully controlled, we can eliminate stop-and-go waves caused by human driving behavior,” said Daniel B. Work, assistant professor at the University of Illinois at Urbana-Champaign, a lead researcher in the study.  The team says that such a vehicular mix can produce overall fuel savings of up to 40 percent.

Mr. Work said the role of autonomous vehicles in regulating traffic flow is the next innovation in the rapidly evolving science of traffic monitoring and control, As an analogy, Mr. Work cited fixed traffic sensors that have been replaced by GPS data in many navigation systems. In kind, self-driving cars are ready to replace commonplace freeway traffic control concepts like variable speed limits. But like any evolving man/machine phenomenon, success will lie with a clearly understood and effected dynamic between self-driving cars and the (often unpredictable) human drivers who operate them and share the road with them.

Why do drivers speed up when a traffic jam clears–only to create a new one up ahead?  Why do drivers brake, accelerate, brake, accelerate, often randomly, causing stops-and-starts and backups all around them?

The researchers discovered this very pattern during an experiment on a simple, circular, one-lane level test track. When 20 human drivers took the wheel track they quickly created their own snarled stop and start situation.

Exploiting Today’s Technology

Where real-life traffic control solutions until now have applied “central control” monitoring via video cameras and traffic sensors (which could help regulate flow through broadcast or internet traffic warnings, speed limit and lane closing alerts, etc.), the Illinois scientists want to bring such oversight into the 21st century. Bringing the tracking data capabilities (GPS apps, e.g.) of the mobile phone to the autonomous car turns the vehicle into a “mobile actuator” itself–creating, in effect, a self-contained traffic monitor that can calculate and control situations in the immediate driving area.

The researchers determined that just one autonomous vehicle among 20 human-driven cars would quickly even out the start-and-stop pattern with a corresponding improvement in fuel efficiency.

The paper describing this work (which was funded by the National Science Foundation’s Cyber-Physical Systems program), titled “Dissipation of stop-and-go waves via control of autonomous vehicles: Field experiments,” is available at https://arxiv.org/abs/1705.01693.

Man and Machine

“The proper design of autonomous vehicles requires a profound understanding of the reaction of humans to them,” says Benjamin Seibold, Associate Professor of Mathematics at Temple University Philadelphia, Pennsylvania, U.S.) and a member of the multi-disciplinary team who participated in the research. “Traffic experiments play a crucial role in understanding this interplay of human and robotic agents.”

Adds team member Benedetto Piccoli, Professor of Mathematics at Rutgers University (New Jersey, U.S), “Fully autonomous vehicles in common traffic may be still far away in the future due to many technological, market and policy constraints, (but) increased communication among vehicles and increased levels of autonomy in human-driven vehicles is in the near future.”

The next step? The scientists plan to test self-drive cars in denser, more complex traffic scenarios, such as those involving lane changes.