August 15, 2007: The U.S. Army and Marines have more than 5,000 unmanned ground vehicles (UGVs) operating in Iraq and Afghanistan, all but a handful of them unarmed. After several years of effort, robot manufacturer, Foster-Miller, convinced the Department of Defense to send three of its armed (with a 5.56mm machine-guns and 350 rounds of ammo) Talon robots to Iraq for more realistic testing. The Talon IIIB, also known as Swords (Special Weapons Observation Reconnaissance Detecting System) will be used as a 125 pound armed sentry, not a combat droid. So far, the tests have been successful. Swords can also be armed with a 7.62mm machine-gun (and 300 rounds of ammo), a .50 caliber sniper rifle or a 40mm automatic grenade launcher. Swords cost $200,000 each. Part of the high cost comes from the addition weapons safety systems installed, and tested. Swords can now defend itself, or be sent into a particularly dangerous location to kill or wound enemy troops.

Most of these small robots, like the Talon and PackBot, are small combat robots a soldier can carry to a window, throw it in, and have it move around and check the place out. For years, robot manufacturers like iRobot and Foster-Miller have tried to persuade the Department of Defense into allowing battlefield tests of armed robots. This made a lot of people outside the combat zone, especially politicians, nervous.

Once American infantry got their hands on reliable, and portable, UGVs (unmanned ground vehicles), they did two things. First, they used their small robots as much as possible, especially for dangerous jobs like checking for roadside bombs, or bad guys lurking inside buildings or caves. The second thing the troops did was ask the UGV manufacturer to put weapons on the robots. The Department of Defense was reluctant to experiment with proposals to arm these MTRS (Man Transportable Robotic System), because of safety concerns. It's not that the armed robots would just be turned on, and turned lose. They would be controlled by their human operator, but there is a reluctance to having the troops equipped with an armed robot. Such systems are more prone to friendly fire incidents. But the troops want them, and the manufacturers of the robots, spent their own money to develop armed versions of its UGVs.

For combat use, armed robots have limitations. Human troops can move a lot faster, and speed is often more important than having an expendable trooper (the robot) in the lead. Armed with a weapon, say a shotgun for opening a locked door, the robot would be more useful. An armed robot would also be more effective when taking the lead in many urban combat situations. Currently, troops continue to find new uses for the unarmed version, like spotting snipers, standing guard and carrying remotely controlled explosives to targets covered by enemy fire. But to be really useful, the troops want armed droids, even if they are not as alert (to smell, sound and other battlefield sensations), as human troops.