In a speech to the annual conference of the Association for Unmanned Vehicle Systems International (AUVSI) the first week of August, III Armored Corps commander Lt. Gen. Rick Lynch hammered home the need for the increased use of robotics on the battlefield.

“We’re losing so many soldiers in convoys it’s an embarrassment – why does every vehicle have to be occupied by a human being? There’s got to be a sense of urgency,” he told AUVSI, adding he lost 153 soldiers under his command in Southwest Asia and “80 percent of those soldiers didn’t have to die.

“I am so tired of going to demonstrations of technology. The technology is there. We’ve got to get past the demonstrations and into the field. If you’re not fielding, you’re failing.”

From Leonardo da Vinci’s apparently never-built robot warriors to Nazi Germany’s Goliath remote-controlled rolling bombs to Hollywood’s visions of robots run amok, the concept of sending machines into battle – alongside or even instead of humans – has been around for centuries.

In the modern age, the first Gulf War’s Pioneer unmanned aerial vehicle (UAV), while primitive by today’s standards, created a permanent role for robotics in combat. When U.S. forces returned to Iraq a decade later, they not only went with far more advanced – and weaponized – UAVs, but with the first generations of unmanned ground vehicles (UGVs), as well.

Initially used at checkpoints to protect their remote operators from car bombs, they soon took on more dull, dirty or dangerous tasks – the 3 Ds of military robotics. Eventually, that included dealing with the most deadly of all dangers facing Coalition forces – improvised explosive devices (IEDs).

So far, ground robots have had little autonomy – in both size and operation, they more closely resemble commercial high-tech toys than C3PO or R2D2. They carry no weapons and each action is directed by a human controller, usually at short range. But Col. Jim Braden (USMC), Robotic Systems Joint Project Office (RSJPO) program manager, said there is little doubt UGVs will follow their flying cousins in terms of future armament.

“We might use the weaponized robot to kill an enemy threatening a downed soldier or just keep his head down until we can recover the wounded warfighter. But just like all other robots, a weaponized robot would have a human-in-the-loop to decide what the robot does,” he said. “Once we get these developed, we will look for the operators to develop TTPs (techniques, tactics and procedures) allowing them to employ a new system just like any other.”

The real division between military reality and Hollywood fiction, however, is the difference between autonomy and artificial intelligence (AI).

“It is easy to mistake autonomous tasking with AI. None of us is looking for AI in these robots or some kind of learning computer that will make its own decisions,” Braden added. “There is a huge difference between autonomy and AI, especially learning AI. I just don’t see a military application for that.”

Advances have been coming at a fairly rapid pace as both military and civilian labs seek to improve everything from sensors and locomotion to the ability of robots to reach out and manipulate their environment. One of the most important drivers advancing robotic development, however, is feedback from those actually using existing platforms in combat.

“Whenever you put any technology into the field, the operational forces will come up with new and different ways to employ it,” Braden said. “For example, they want to add senses. Right now, it sees well, but they want to know what it can hear because it is up to 300 meters away.

“Some have said it would be helpful to smell, so they can get from the robot the same things they would get using their own senses. They also want whatever sensors we give them to be plug-and-play; we’re not quite there yet, but close. And they want a single lightweight battery and smaller, more agile controllers.”