Drone Pilots May Need Distractions

LiveScience.com
Drone Pilots May Need Distractions
.

View photo

A U.S. Customs and Border Protection unmanned aircraft arrives in Corpus Christi, Texas to prepare for …

Distractions can be lethal while driving, but they may help when flying drones, researchers say.

Operating a drone such as the Predator may seem like playing a video game, involving a joystick to steer the unmanned aerial vehicle and firing missiles at targets seen on computer screens via a camera on the UAV. However, such adrenaline-charged moments are much rarer in real life than in video games. Drone operators, many of them seasoned fighter pilots, typically spend most of their shift just watching and waiting while automated systems keep the droid running. A shift for the operator of a Predator can involve up to 12 hours of such boredom.

"You might park a UAV over a house, waiting for someone to come in or come out, and that's where the boredom comes in," said researcher Missy Cummings, a systems engineer at MIT. "It turns out it's a much bigger problem in any system where a human is effectively baby-sitting the automation."

Such mind-numbing work can impair performance by making it difficult for an operator to leap into action when intervention is necessary. Cummings and her colleagues have been looking for ways to keep drone operators alert during tedious downtimes.

[Military Test: Drones Could Refuel Themselves Mid-Air]

"We need to accept that we're automating the world more and more, and a side effect of this increased automation is that people are going to be bored monitoring those systems," Cummings said. "But we still need smart people to monitor those systems when they fail; we need smart people to intervene when complex systems go wrong."

The researchers have found that most pilots operating a UAV simulation are less bored and perform better when they have some distractions, such as checking their cellphones, reading a book, or getting up to snack.

"We should think about sterile environments, workplaces where we tell people they can't play 'Angry Birds' on their iPhones during really dull and boring moments. When we're enforcing sterile environments, we're almost setting people up to fail," Cummings told LiveScience.

The scientists enrolled 30 volunteers to interact with a UAV simulation in four-hour shifts. Participants each had to monitor the activity of four drones and create "search tasks," or locations UAVs had to investigate. Once a UAV identified a target, volunteers labeled it hostile or friendly based on a color-coded system. The participants ordered UAVs to fire on hostile targets, destroying them and earning points in the simulation. The volunteers were videotaped to see when they were paying attention.

The volunteer who scored the highest during the experiment also was the one who paid the most attention. "She's the person we'd like to clone for a boring, low-workload environment," Cummings said.

However, the volunteers with the next-highest scores performed nearly as well even though they were distracted 30 percent of the time.

"We get bombarded in the news with studies that distractions are bad. Certainly for real-time control tasks such as driving, we're all in agreement, distractions are bad," Cummings said. "However, one of my issues are things like the Federal Aviation Administration's recent decision to fire an air traffic controller for watching movies at 2 a.m. to stay awake while waiting for something to happen. We want to think about how to keep people who work in low-task-load environments engaged so that when something does happen, they can respond appropriately."

While their simulation required human input only 5 percent of the time, the researchers discovered that most of the volunteer operators tried to work 11 percent of the time, showing they wanted more to do to keep from getting bored. This suggests that distractions or busywork once in a while may be good for productivity, by keeping operators engaged when they might otherwise lose focus.

"We know that pilots aren't always looking out the window, and we know that people don't always pay attention in whatever they're doing," Cummings said. "The question is, can you get people to pay attention enough, at the right time, to keep the system performing at a high degree?"

Personality may be another key to concentrating on drone work. Personality surveys of the volunteers  ranked them in five categories — extroversion, agreeableness, conscientiousness, neuroticism and openness to experience — and found that conscientiousness was a common trait among top performers.

Such conscientiousness may be a two-edged sword. Conscientious people may perform well in environments with relatively light workloads such as UAV operation, but they may hesitate when the time comes to fire a weapon.

"You could have a Catch-22," Cummings said. "If you're high on conscientiousness, you might be good to watch a nuclear reactor, but whether these same people would be effective in such military settings is unclear."

The researchers are continuing experiments to see what conditions might best keep boredom at bay. For instance, regular gentle reminders might help people keep alert. The scientists also are looking into the best length for shifts and the best time of day to hold them.

"We need people who can monitor these systems and intervene, but that might not be very often," Cummings said.

The research might have larger implications, such as for operating automated vehicles like the Google Driverless Car.

"People are already bored when they're driving, and they're going to be really bored when automation drives the car," Cummings said.

The scientists detailed their findings Nov. 14. Their research will appear in the journal Interacting with Computers.

This story was provided by TechNewsDaily, a sister site to LiveScience. Follow TechNewsDaily on Twitter @TechNewsDaily. We're also on Facebook & Google+.

Copyright 2012 LiveScience, a TechMediaNetwork company. All rights reserved. This material may not be published, broadcast, rewritten or redistributed.
View Comments