Robots get real

All day long they scurried around the ninth floor of the BU Photonics Center, humming softly and craning their long metallic necks, searching for snipers and vacuuming floors. These were robots on parade, showing off at From Hollywood to Homes:  Robots for Today, a conference of academic, government, and industry leaders in the artificial-intelligence field, held at the center last Friday, March 31 and put together by the College of Engineering’s department of manufacturing engineering.

The first keynote speaker was Colin Angle, CEO and cofounder of iRobot, which is based in Burlington, Mass., and is the creator of the vacuuming Roomba and the kitchen-floor cleaning Scooba robots, as well as the rugged military/industrial PackBot, about 200 of which are assisting soldiers in Iraq and Afghanistan. Other speakers included Stephen Welby, director of the tactical technology office of the Defense Advanced Research Projects Agency (DARPA), and Pedro Del Nido, chief of cardiac surgery at Boston’s Children’s Hospital, who spoke about robot-assisted surgery.     

Conference attendees discussed the development of robots for household chores, for work in hazardous environments, and even as “personal avatars,” robots that a person could control remotely, seeing and hearing what the robot sees and hears and communicating through the robot, all via wireless Internet connection — “putting your brain into a robot,” as Angle put it.   

While the robotics industry has made great strides in recent years, it is still in its mechanical infancy. According to an article in Computerworld magazine, about $5.5 billion worth of robots were sold in the United States in 2005 for industrial uses such as assembly line work and an additional $1 billion went for robots used for personal services such as education, entertainment, and cleaning homes. Most of the conference speakers focused on how exponential growth in computer processing, data storage, and high-bandwidth networks would propel the industry forward. Several discussed the benefits of using robots in war zones, where, for example, they can search caves in Afghanistan or look for “improvised explosive devices” hidden alongside Iraqi roads. 

“When robots save lives, they get blown up, and the soldiers don’t,” said Glenn Thoren, deputy director of the Photonics Center. Thoren presented the Robot Enhanced Detection Outpost With Lasers, or REDOWL, which can hear, identify, and locate the source of gunfire in milliseconds. Equipped with direction-finding microphones, a 300x zoom camera, a laser rangefinder, an infrared illuminator, and a thermal imager, REDOWL was developed at the Photonics Center over the last 18 months in conjunction with iRobot, Insight Technology of Londonderry, N.H., and the BU startup company BioMimetic Systems. Its first chance to prove itself with soldiers under simulated combat conditions will come this fall during a 10-week Army training exercise at Fort Benning, Ga.

After the conference, Thoren discussed the future of robotics, including REDOWL, with BU Today.

BU Today: The conference speakers talked about many possible uses of robots, from military missions to household chores to space and underwater exploration. Is there one field in which you’d predict the most rapid development of robotics?
 
Thoren: I think the most rapid development will be in the commercial field, because robotics has entered a new era where it is now affordable and acceptable for people to have them take on some of their daily chores. If a robot can do a chore, like vacuum your home, adequately and well and the robot is affordable, then people will buy it. This will soon make robots in the home as acceptable as a DVD player, a personal computer, or a PDA.

In his keynote address, Colin Angle said the limit to the speed at which robots could become integral to our daily lives was not “one of technology, but one of customer acceptance.” Do you agree? 
 
It’s all about trust. One of the key things from my perspective was when Colin [Angle] mentioned customer acceptance in combination with the word trust. Trust is saying that a robotics system can do a task as well as or better than a human being could. Now, in the case of cleaning floors — well, a robot could certainly do it more often. But in the medical world, the accuracy with which the tools can be deployed could exceed the capabilities of a human being, because mechanical robotics systems are exceedingly accurate. With electronics, the placement and motion of medical tools and devices can be more exact for very precise surgical movements that humans are incapable of making. 

In terms of the military, again robots are doing things that humans can’t easily do. They can follow a soldier around and carry cargo, lifting many times what a soldier, or even a dozen soldiers, could manage. They can perform missions under extremely dangerous circumstances, such as retrieving injured solders from the battlefield. They can explore caves or buildings where the space is limited and a soldier becomes an immediate target to an enemy. The other thing about robots is that they can go places soldiers can’t easily go, such as underwater for extended periods of time. They can monitor hilltops or other areas indefinitely.

Of course, there are things that humans can do that robots can’t do, and we’re actually helping to solve some of those issues. For instance, robots need to have the same perceptive capabilities that humans have, such as three-dimensional sound acquisition and identification — being able to perceive where a sound came from and what made the sound. That’s one of the things that make REDOWL unique. On the other hand, robots can have some perceptive capabilities that humans don’t have, such as thermal imaging.

One of the conference speakers spoke of a future with “homes that essentially take care of themselves.” To me this evoked the prediction made many decades back that we’d soon be driving flying cars. In your opinion, where’s the dividing line between what we can expect robots to do for us and where our thoughts are inclined to science fiction?

The evolution of robot technology will be in small, very firm steps, such as taking the place of vacuum cleaners and floor-washing equipment. Your house is an environment you control, and you can put anything in that environment that helps you, whether it’s to carry goods in from your car or mow your lawn, anything that helps you with routine household maintenance. Robotics will become much more integrated into the surroundings that we control, such as our homes, and more networked, so that they can communicate. There are already, for example, stoves that will turn on and start cooking because you tell them via the Internet that you’ll be home in 30 minutes, or a refrigerator that will order food for you when it notices you’re running low on certain things. There are systems right now that allow you to see in your driveway and all around your house through little cameras wirelessly connected to your PalmPilot. So there’s going to be a new level of home security, home maintenance, and home monitoring that has robotics as a part of it.

Your own talk was about REDOWL, developed in a partnership between BU researchers and local tech firms, including iRobot. What’s the next step for REDOWL?

In September, at Fort Benning, Ga., the Army will hold a 10-week field test of new equipment, including the REDOWL. It will be conducted by the Air Assault Expeditionary Force, which is a highly mobile unit, people who have been in battle and understand how to conduct war with the best equipment available. This will be the first time that the REDOWL is put in the hands of the frontline troops, where they can exploit the capabilities of the REDOWL and determine what they like and what they don’t like and how rugged the system actually is.
 
Currently, the robotics industry relies a great deal on universities for research. But as the industry matures and draws in more investors and venture capital, do you see the role of universities increasing or decreasing in future robot research and development?

I see it increasing, for the simple reason that universities have the unique capability of acting as a hub of information transfer. We can hold events like this conference, where people can come together, spend a very condensed amount of time, and learn a lot. It’s a networking opportunity as well as an opportunity to get a snapshot in time of the state of the technology.

That was the thing that most impressed me about the conference. People were extremely open and asking questions and interacting with the speakers across a range of topics, with a broad scope of interest — from starting a robotics company to growing a company to looking for the next technology to bring to a company. Universities have always held that role as a hub of networking activity, communicating between government, industry, and academia. Companies can’t do that as well.