iRobot's 510 PackBot

[attachment=17884]

iRobot Corp.

September 11 had quite a different meaning for me before it became 9/11. It's my birthday. And early on that cool blue morning I took my daily run from our Chelsea apartment down the West Side Highway past the World Trade Center to South Ferry and back. I never thought much about the Twin Towers then, except sometimes to remember the convoluted history of how they got built in the first place, or to feel their overwhelming presence in what was then a low-rise part of Manhattan as I breezed past, with no sense of how gigantic their absence might become.

At work later that morning, when the news came in, we first thought "accident." After all, in 1945 a B-25 bomber had crashed into the 79th floor of New York City's Empire State Building. But how could this be on such an acutely clear day? Now, 10 years later, when almost any disaster is tinged with the possibility of terrorism, it's hard to remember a time when these thoughts were not second nature.

Before anyone knew there would be no one to rescue, people from around the United States, and around the world, flocked to the city to help. Among the would-be rescuers who came that day and the ones that followed were robots. They were driven to the scene by engineers and technologists from places like iRobot Corp. and Foster-Miller in the Boston area, and from Robin Murphy's brand-new Center for Robotic-Assisted Search and Rescue, then at the University of South Florida, in Tampa, and now at Texas A&M University.
While they saved no one, these robots were able to traverse some of the vast debris field, going where humans and dogs dared not, demonstrating indisputably that they weren't toys or expensive curiosities but viable machines capable of standing in for humans in dangerous situations. As Murphy has noted, before 9/11 the idea that intelligent robots could help save lives at disaster sites was dismissed as science fiction. But not after.
To be fair, the artificial-intelligence community and the military were skeptical for good reason. While serious work on robot development had picked up in the 1990s, the problems facing robot builders were intractable and expensive: Computer vision was still slow and unreliable; robotic control, remote or autonomous, was tenuous. Sensor-based navigation was demonstrable but fragile. Materials and motor technologies were rudimentary, and sensors weren't small enough or smart enough.
But since 9/11, work and progress in all these areas has accelerated dramatically. Now robots are everywhere, and the demand for more is tremendous. The U.S. military, for one, has seized upon the development of ground robots to detect improvised explosive devices and drones to fly reconnaissance missions, as our August story "Autonomous Robots in the Fog of War" makes clear.