March 7, 2005: The U.S. Army is speeding up its efforts to get more UGVs (unmanned ground vehicles) into the hands of the troops. Over the last few years, hundreds of small UGVs have been used by American troops for checking out caves and buildings. Some of these lightweight (under a hundred pounds) robots are being equipped with weapons. The next class of UGVs will be heavier, weighing 1-3 tons. These include the much anticipated "mule" UGV. This critter will carry the troops equipment, bring up supplies, and move wounded troops. As more gadgets are invented for the troops, the weight they have to carry keeps increasing. One solution is a UGV that can accompany troops, carrying a lot of this load (otherwise, each soldier is going to be carrying about a hundred pounds of gear, which is hardly "fighting weight.") There's one problem, a major one, and that is the building of a sensor/software system that would allow the mule UGV to move along the ground without a human driver. So far, this has proved to be a major obstacle. It is possible to have a remote human operator control a UGV. That keeps people out of harms way, and the military is using operators half way around the world to operate UAVs. This "reachback" technique is possible because of cheaper and more abundant satellite communications. No reason it could not be done for UGVs. And this may be the interim solution until a smart enough UGV navigation system is built. This would be a big help for the troops, as they would have the benefit of these UGV drivers, but these troops (they could even be civilians), would be out of harms way, and would not consume supplies in the combat zone.Once the mules are with the troops and working, and this may happen in the next few years, you will quickly see armed mules, some of them armored and weighing up to ten tons. Without a human crew, a ten ton armored mule would look like a miniature tank, but would be carrying a 25-30mm automatic cannon and 2-4 Javelin missiles. These vehicles would also carry day/night vidcams, thermal imagers, "ears" (acoustic sensors) and a nose (chemical sensors). The combat mules can also talk, using embedded foreign language systems (like the hand held versions troops have been using), or human translators (again, these could be back in the United States, and could even be civilians.) The advantage of these armed mules is that they can destroy enemy armored vehicles, while being harder to detect (as they are about the size of small sports car). No friendly troops are risked when these vehicles are hit, and their sensors and human operators (some of whom are far to the rear) are more alert, over longer periods of time, than human crews in current armored vehicles. You'll be seeing pictures of the prototypes in the next year or two. The laboratory prototypes are not very impressive to look at. The use of a mix of local troops and operators back in the United States will be a big "force multiplier" until software is developed that is smart enough to maneuver the armed mules, and decide when they can use their weapons.
| "Bolo Brigade, reporting for duty!" |
There are already civilian robotic security systems. But their software moves a vehicle around a known course, with limited navigation ability, and only enough smarts to alert a human operator that the droid has encountered something that is not supposed to be there. A combat UGV that will be allowed to decide when and where to fire is possible now. But there is reluctance to turn these loose. The first such autonomous weapons of this class were the landmines developed in the 1930s, and widely used during World War II. Torpedoes and naval mines are the same type of weapon. But at sea, it's much easier to keep track of where you are, where you're going and what the target is. Land warfare is a lot messier, and requires much more "intelligent" software to operate effectively. But it's not a matter of "if" these autonomous war droids are going to appear, but when. It's possible that the next generation of American combat vehicles will be largely unmanned versions.