Monday, January 9, 2017

Sharing the road could get dicey

Things could get dicey when human drivers start mixing with self-driving vehicles, at least according to a report from the University of Michigan Transportation Research Institute.

“During the transition period when conventional and self-driving vehicles would share the road,” the report says, “safety might actually worsen, at least for the conventional vehicles.”

That’s not what I expected to read.

Written by researchers Michael Sivak and Brandon Schoettle, the report is dated January 2015. I found it on a Research Institute web site under the heading “Recent Publications.”  

2015? Recent?

I called the university for clarification. Turns out the website isn’t updated as often as it should be.

Anyhow, the two-year-old report also notes that experienced drivers may be just as safe as self-driving vehicles. “It is not a foregone conclusion that a self-driving vehicle would ever perform more safely than an experienced, middle-aged driver,” the authors say.

I hadn’t expected to read that either. Yet of all the studies and reports I have read about autonomous vehicles, this makes the most sense and is the easiest to believe.

So if I read it two years ago, I think I would remember. But I don’t. Neither could I find any reference to it in the online archives of the trucking media.

Maybe it’s old news, maybe it isn’t. One way or the other, there’s interesting stuff in the report with the catchy title: “Road Safety with Self-Driving Vehicles: General Limitations and Road Sharing with Conventional Vehicles.”

For instance, the authors, Sivak and Schoettle, explain that middle-aged drivers who tend to have fewer accidents than younger drivers are most likely to compete with the safety of self-driving vehicles. Experience is the key, of course, but specifically in this case, it is “the predictive knowledge about the likely intentions of other road users.” This is knowledge mature drivers are more likely to have.

Predictive knowledge gained through experience would be difficult if not impossible to program into self-driving vehicles, say the authors. Therefore, it is unclear if the computer power of a self-driving vehicle “would trump the predictive experience of middle-aged drivers.”

All you middle-aged drivers out there feel free to puff out your chests.

On the other hand, that predictive experience may not be quite as valuable when sharing the road with self-driving vehicles. “In many current situations, interacting drivers of conventional vehicles make eye contact and proceed according to the feedback received,” say Sivak and Schoettle.

You can’t have eye contact with a self-driving vehicle, of course.

It seems the road-sharing transition from human drivers to all autonomous highways will take a long time

“In 2002, 13.3 percent of all light trucks sold 25 years earlier were still on the road, with a corresponding percentage for cars of 2.3 percent. As a consequence, there will likely be at least a several-decade-long period during which conventional and self-driving vehicles would need to interact,” say Sivak and Schoettle.

“Furthermore,” they add, “to the extent that some people may want to drive only conventional vehicles, this overlapping period might last indefinitely.”

Indefinitely is a long time. But even when we finally get there, the nirvana of all self-driving vehicles may not be as perfect as we’ve been led to believe. Here the authors deliver what seems another reality check, taking issue with the extravagant claims of many enthusiastic autonomous vehicle boosters.

According to Sivak and Schoettle, “The expectation of zero fatalities with self-driving vehicles is not realistic.”

Wish I had read it two years ago, but I’m glad to read it now because it’s as telling about the challenges of autonomous vehicles now as it was then.