Her intelligent robots help search for victims of disaster
This profile is part of IEEE Spectrum’s Special Report on Dream Jobs 2009.
Robot Menagerie: Robin Murphy, an IEEE member, shows off her robots at a train wreck staged for rescue drills. Photo: Matthew Mahon
1 September 2001 was a proud day for robotics expert Robin Murphy. It marked the opening of the Center for Robotic-Assisted Search and Rescue, which was aimed at promoting an idea that Murphy and one of her former graduate students had been pushing for six years: that intelligent robots could help save lives at disaster sites.
Ten days later came the attacks of 9/11.
Murphy and three students immediately packed up their robotic gear and drove 18 hours north from Tampa, Fla., to Manhattan, where they worked alongside rescuers at Ground Zero for the next 12 days. They rarely slept or ate, and the experience left them drained. The purplish-blue light that illuminated the rubble pile at night cast a surreal glow. “It looked like something out of The Terminator,” she recalls. For Murphy, and for the technology she had been espousing, it was a baptism by fire.
Since then, she has used robots to help search for victims in the aftermath of eight other disasters, including the 2005 mud slides in La Conchita, Calif., Hurricane Katrina, and the 2007 Crandall Canyon Mine collapse in Utah. When she’s not promoting the use of robots among rescue professionals, she works closely with robotmakers to tailor their designs to search-and-rescue missions. And she’s educating and training a new breed of roboticists to think not just about the detailed engineering of their machines but also about how these devices might function within the larger world of emergency response.
Murphy’s interest in technology came largely through the example of her father, who was head of engineering for a huge chicken-processing plant in the small town of Douglas, Ga. After earning a bachelor’s degree in mechanical engineering and a Ph.D. in computer science, both from Georgia Tech, Murphy landed a faculty position at the Colorado School of Mines, in Golden. The school had strong links with Colorado’s space-science community, and as a consequence, she says, “pretty much everybody in artificial intelligence was looking at planetary robots.” But the 1995 bombing of the Alfred P. Murrah Federal Building in Oklahoma City made her rethink her research. One of her grad students, John Blitch, a major in the U.S. Army, went to Oklahoma City to assist in the rescue efforts. That harrowing experience moved both him and Murphy to apply their knowledge of robotics to aiding the victims of similar disasters. “I made a commitment to myself to focus on rescue robots,” she says.
And that’s exactly what she did—at no small risk to her academic career. At the time, researchers in artificial intelligence considered search-and-rescue robots only a vague futuristic possibility, and the chair of her department wasn’t supportive of the direction her research was taking. But Murphy persevered, eventually landing a tenured position at the University of South Florida, where her faculty colleagues supported her then-fanciful idea that robots could usefully assist rescuers. Elsewhere, though, most AI researchers were skeptical. Then “9/11 happened, and all of a sudden I became brilliant,” she quips.
Although Murphy’s work at USF mostly entailed teaching and scholarly research, she kept her feet firmly planted in the practical realities of rescue operations. She and some of her grad students joined Florida Task Force 3, a state-sponsored urban search-and-rescue team that specializes in locating and extricating victims trapped in collapsed structures or other confined spaces. Working elbow to elbow with safety professionals during training exercises and actual disasters proved invaluable in shaping Murphy’s research agenda. “I love fieldwork,” she says. “Fieldwork is where you learn something.”
At the World Trade Center, for example, she discovered that bigger robots are a lot less useful than smaller ones in probing a collapse—for the simple reason that if an opening is large enough for a person to wriggle into, human rescuers won’t hesitate to do so. They may be risking their lives, but they will almost certainly do a better job than any robot. “It’s kind of ludicrous to think that you can replace the complexity of a human” with a machine, she says. “Also, firefighters don’t want to be replaced.”
But firefighters do want tools to do things they can’t do—like squeezing through shoebox-size openings into a deep pile of rubble to see whether anyone is trapped inside. Although robots have yet to actually rescue anyone in the aftermath of a calamity, they have proved their worth by helping determine whether a partially destroyed structure is safe for people to enter.
Rescue robots need to be reasonably smart, Murphy says. They don’t have to be like Lt. Cmdr. Data of “Star Trek” fame, but they should have something like the brains of a horse, she says. That is, the operator should be able to give the robot high-level commands and not have to specify its every move.
In mid-2008 Murphy left USF for Texas A&M University, where she is the Raytheon Professor of Computer Science. One of the attractions was that A&M hosts the Texas Engineering Extension Service, which has been training firefighters for many decades and boasts a 21-hectare site known as Disaster City, with facilities for simulating just about any sort of catastrophe you can imagine, including downed aircraft, collapsed buildings, and wrecked trains.
Murphy regularly takes her robots out to Disaster City and tests them alongside human emergency responders. Back at the lab, she equips these sophisticated machines with ever better sensors and intelligence, and she teaches her students the principles of search-and-rescue robotics.
The work is so rewarding that Murphy feels no pangs about forgoing what would probably have been a more lucrative career in industry. Indeed, she’s delighted with how her early commitment to rescue robotics has turned out, and she’s confident that the day is close at hand when a search-and-rescue robot will make a critical difference after disaster strikes. “We’re really, really close to helping responders save a life,” she says.
To Probe Further
For more articles and special features, go to Dream Jobs 2009.