- The Washington Times - Monday, April 29, 2002

To many people "robot" means a clanking human-looking machine. To others it means a mechanical arm that does things like spray paint in factories.
The first isn't here yet, though labs are working on them. The second has been around for years. But there is a third form of robot, now on the verge of becoming important: the unmanned autonomous or, more likely, semiautonomous vehicle designed for a special purpose.
A good example is a machine called PackBot, being developed with military funding by a company called iRobot(www. irobot.com). It is a small, portable, remotely controlled tracked vehicle, intended for military uses such as reconnaissance in cities. The idea isn't original: Science-fiction novels from the '50s described similar contrivances. Police departments have had less-autonomous but similar bomb-squad robots for some time. What is new is that the technology is reaching the point at which the things actually work.
A few weeks ago in this space I talked about unmanned aircraft and their use in the war in Afghanistan. PackBot and various relatives are the land version. We're going to see lots more robotic weaponry. Similar civilian robots are being designed for police work, going into fires and doing dangerous industrial jobs.
Military uses: A problem in urban warfare is that there are countless ways of getting shot from hiding. Sending a soldier to reconnoiter, perhaps to look around a corner, exposes him to hundreds of potential sniping positions among others, every visible window in every building. If instead you send a small mechanical creepy-crawly with a camera, the most you can lose is a creepy-crawly. For a country that is inclined to back out of wars if there are casualties, this is important. It is also important to the soldier who doesn't get shot.
The idea isn't wild-eyed or blue-sky. Children today have remote-controlled racing cars that work fine for startling unwary adults. Video cameras are cheap. The technology of wireless data links is well understood. To make an effective military crawler, you don't need new magic just good engineering and imagination.
For example, the robot needs to be able to turn itself right-side-up if in gets turned over, and it has to be durable. The company says that PackBot has survived falls from second-story windows, and it can right itself. That sounds practical to me. Another trick, being worked on for PackBot and others, is to make the onboard computers handle as much as possible by themselves so that the soldier at the remote console doesn't have to worry about details.
Once you have the basic vehicle, you can put on it whatever you want for a particular job. Most things you might want night-vision gear, a global-positioning satellite receiver, microphones, radar exist now, even on the civilian market. This the fact that designing a robot means in large part assembling what we already have is likely to speed deployment. In the aftermath of September 11, interest has intensified.
Now, how much will robots affect the military? A lot, I think. In the past, clearing a building of enemy soldiers was extremely bloody and nasty work. If a robot crawler can, in fact, negotiate stairs and obstacles, as increasingly it can, you send it into the building. Put a sizable explosive charge aboard and, if nothing else, send it into a defended room and blow it up. Crawlers are cheap, even if they cost a lot. Troops aren't.
In a time of constrained military budgets, the economics of robotics becomes important. Note that prices of expendable robots can be deceptive. If you used unmanned aircraft carrying explosives to, say, fly into an enemy communication center (not necessarily the best way to use one), the cost of the aircraft might seem high. But compare the price of using a drone to the price of using a fighter-bomber.
The fighter costs $30 million to $50 million, depending. Lose one and you've paid the price of a whole lot of robot planes. In dollar terms alone, pilots are expensive. If you use air-to-air refueling, you're paying for buying and flying tankers. An aircraft carrier costs well into the billions. Operating one is phenomenally expensive.
The actual cost of robots would depend on what war is fought and how. But even in coldblooded money terms, it's likely to be cheaper to send in a small tracked PackBot-style crawler than a scared teen-ager who is susceptible to being killed. To an enemy, how appealing is it to know that the enemy killing you is far off, at a console, with a cup of coffee?



Click to Read More

Click to Hide