Road Bots: Lionel Robert M '97
Lionel Robert poses among equipment at an on-campus robotics lab at the University of Michigan.
Robert is looking for answers to all of the questions surrounding autonomous vehicle technology.
LIONEL ROBERT is an associate professor of information at the University of Michigan, where he researches and collaborates with students on the relationship between technology and teamwork in modern society. Robotics and autonomous vehicles — AVs for short — are his specialty, which is why, when an AV struck and killed a pedestrian in Tempe, Arizona, on March 18, Robert suddenly found himself in high demand by the media.
The fatality occurred when a woman stepped out in front of a self-driving Uber. The AV failed to stop even with a safety driver, and Robert confirms that the AV’s lidar technology (like radar but using light rather than radio for detection and ranging) should have been able to detect the moving person and respond. But he also stresses that this tragedy is a result of multiple factors.
“We have to be careful to not oversell the technology,” he says. “Look, an autonomous vehicle is a vehicle. A vehicle weighs one or two tons and, moving at x amount of speed, cannot stop on a dime.”
There’s a widespread perception that autonomous vehicles will all but eradicate traffic problems and accidents on the road once they become commonplace. But Robert explains that figuring out how to program AVs and how to integrate them into society brings up one issue after another, many of which are counterintuitive.
“When we first started with the problem of autonomous driving, we thought it would be easy, and the reason why is because driving is a pretty explicit activity,” he says. “There are laws. There are fixed lines. It seems to be made for an algorithm, for artificial intelligence. But it turns out that driving is an incredibly social activity. No one follows the law when they drive.”
Should AVs be programmed to break traffic laws in order to avoid accidents or react to other drivers? Should urban infrastructure and roads be redesigned to accommodate AVs? Questions like these continue to arise as AVs become more of a reality. Robert believes it will take a lot of education and engagement of the general public to move forward with this kind of technology.
In the meantime, Robert is focusing on research with his students. One particular study they’re conducting uses virtual scenarios to explore the ways AVs might communicate with pedestrians.
“The thing about this research is that we’re doing something that people don’t know,” Robert says. “I tell students the answer isn’t in the back of the book; we’ve got to find out together.”