Priorities

By: Michael Laris

Matt McClain/The Washington Post

As guinea pigs go, Mark R. Warner is as good as any.

The senator from Virginia was in the passenger seat of a gold Cadillac SUV on Monday, cruising down a closed-off stretch of Interstate 395 south of the Pentagon with a guy named Luke behind the wheel and an algorithm in control of the car.

And things were about to get a tad hairy.

The hands-free, feet-free future of self-driving cars — or at least semi-self-driving cars — is hurtling ahead as decades of research and advances in computing power make automatic autos a reality. Tesla owners got a software upgrade last week adding features to its “Autopilot,” such as “steer within a lane,” and other manufacturers are touting similar features coming next year. Tech giants such as Google and Uber envision swarms of self-driving vehicles delivering products and humans.

But for all the souped-up technology — and the promise of safer, smarter cars wasting less space, gas and time as they thread efficiently through 21st-century traffic — many of the biggest challenges remain in the human sphere.

A console of a vehicle with connected technology is seen during a press event that followed a test of the connected vehicle. (Matt McClain/The Washington Post)

On Monday, Virginia officials gathered to observe a test of the technology and underscore, as Secretary of Technology Karen Jackson put it, that “Virginia is open for automated and unmanned business.”

A number of states across the country, California, Nevada and Michigan among them, are trying to capitalize on such emerging technologies, and officials from Virginia say doing so is particularly important as the commonwealth seeks to move away from depending quite so heavily on defense spending.

Researchers working on the new technologies say a key challenge will be easing a transition in the way people relate to cars.

One issue out of the gate is the fear factor.

When Warner’s Cadillac — tricked out by Virginia Tech engineers — sensed a state trooper’s cruiser ahead on the side of the road, the car steered itself one lane to the left.

Good, so far.

But when the trooper eased back onto the road as part of the demonstration and sped up until the cruiser was on Warner’s tail, the sleek Cadi’s computer brain yanked hard on the wheel and jerked the car back into the right lane to get out of the way.

Virginia Sen. Mark Warner was among the few who took a ride in an automated car that alerts drivers to emergency vehicles and can change lanes by itself. (WUSA9)

With a jolt of adrenaline, Warner thrust out his hands toward the dashboard, bracing himself like any startled passenger might. Without a human sitting behind the wheel, such moments come with added uncertainty — and an urgent hope that the machine and its programmers actually know what they’re doing.

“There was the look of terror there for a moment,” Warner said afterward.

Luke Neurauter, of Virginia Tech’s Center for Advanced Automotive Research, said researchers are working “to improve that natural feel” during such maneuvers. But he added that human drivers can react sharply, too.

“If that police cruiser was coming up on you and you didn’t see him until the last minute and you really needed to get over, it would be kind of that hard jerk and correction to get in the right lane,” Neurauter said. “It gets the job done.

“Is it exactly how you or I would do it? No. Not yet. . . .Obviously, there’s a consumer acceptance and a feedback component of that, too.”

Eliciting that feedback is a primary part of the Cadillac’s job as researchers focus on the softer, human side of such rides. Beyond jitters, a key focus is on how drivers tune out as cars learn to do more for them.

“Humans are terrible with being bored,” said Zac Doerzaph, an engineer with the Virginia Tech Transportation Institute. He and other researchers, working with auto manufacturers, suppliers and government agencies, spend much time investigating what is known in the field as the “transfer of control” issue.

They put an experimenter in the back seat and essentially start messing with the human “driver” in various driver-less scenarios, since, for now, they need him or her to be ready to take back control at any moment.

They’ll have the car stray from its lane — a realistic problem, given bad weather or degraded lane markings. Then they change around the way the car alerts the driver of a problem, and “maybe even make it so [the car is] unaware it’s failing, so the driver really has to take control without being told they need to.”

“Then we can watch how you take control back,” Doerzaph said. “We’re really here to do the research to ensure these technologies not only improve things, like mobility and efficiency, but to ensure that it’s all done in a way that at the very least doesn’t take away from safety. But ultimately our goal is to drastically improve safety. That’s kind of our fundamental belief — that that’s all possible.”

Research scientist Myra Blanco said there are gradations to what people want and what technology can and will be able to do, and she and her colleagues are trying to see how those line up. And there is a long distance to be worked out from traffic-jam assist technologies, which are here today and keep drivers a certain distance from the car inching in front of them, and hailing a robot car to come take you to a ballgame or the beach.

In her technical lingo, Blanco said a key human-level challenge is the notion of “primary task reversal.”

“You’re primary task, people tell you, is driving. You need to pay attention to it,” Blanco said. “But when you have a system that is telling you, ‘Hey, I got you. It’s okay,’ then people think: ‘If you’re taking care of this, I can do other stuff. Very cool!’ ”

Many intermediary technologies will require drivers to be alert, even when their hands aren’t on the controls.

“Time flies so quickly. . . . You’re allowing the user to do other things, but at the same time you’re telling them, ‘But don’t do it for too long!’ ” Blanco said. Research shows that a person’s “primary task” can shift quickly from driving to writing e-mails or making quick edits in that presentation he or she is finishing.

“We are like that as humans. We get engrossed in a task. We’re goal-oriented,” Blanco said. “How do we, after giving the user control of their life back, pull them back to paying attention?”

Warner is a big booster of the technology, and he wasn’t turned off by one startlingly sharp turn on the way to the future.

“I could get used to that,” he said, since it’s expected to make cars safer, faster and more fuel-efficient. Warner, who made his fortune as a cellular technology pioneer, held up his iPhone in a scuffed brown case and said people were skeptical decades ago of the coming mobile revolution.

With manufacturers, Google, Uber and others “putting literally hundreds of millions, if not billions, of dollars” into automation, he said, change will come much faster than people think.

“This is the next great big disruptive technology,” Warner said.