September 30, 2009: U.S. Army combat commanders, and their troops, have enthusiastically adopted UAVs and robots. The remotely controlled robots have been used mainly for dangerous situations (like examining possible roadside bombs and boobytraps) where troops actually operated the droids. In the last nine years, the army went from having a few dozen UAVs, to operating several thousand. Most are five pound Ravens, which combat troops have been using to obtain unprecedented aerial reconnaissance support. In that time, the army also added thousands of UGVs, but these were really remotely controlled robots, very similar to radio controlled cars and trucks, sold as toys for decades. Indeed, when the troops were short of army issued robots, they filled the gap with many of the larger radio controlled toy trucks. But efforts to create UGVs and UAVs that can operate more independently have moved along very slowly. Actually, the Israelis are ahead of the United States in this department, and much can be learned by examining the Israeli experience, and their failure to get an autonomous battle droid into action.

Tech savvy officers and troops know that it's possible to build vidcam equipped robots on the ground (UGVs, or unmanned ground vehicles), as well as UAVs, that can operate largely on their own. These devices would only alert a human operator (who could tend three or more of these robots) when something of interest was detected. Israeli efforts in this area show what can be done.

Four years ago, an Israeli firm produced a robotic vehicle based on the two seater all-terrain "TomCar." Called AvantGuard, the robotic vehicle used sensors and software that enabled it to patrol along planned routes, and was capable of some cross country operation as well. The AvantGuard mounted a remote controlled gun turret equipped with a 7.62mm machine-gun. The vehicle had digital cameras facing every direction, and used pattern recognition to identify potential threats (like people sneaking around where they are not supposed to be), or obstacles on the road. The idea was that a pair of human operators could control a dozen or more AvantGuard vehicles. This system was particularly effective at night, because it had night vision and moved quietly. Weighing only 1.3 tons, the AvantGuard was protected against rifle fire and fragments from shells and smaller roadside bombs. AvantGuard proved adequate for guarding industrial parks, but not the vast stretches of Negev desert, along the border with Gaza. Too many things could go wrong out in the desert (obstacles in the road, hostile action) that AvantGuard could not handle.

Two years ago, building on the AvantGuard technology, a new firm, G-Nius, produced the Guardium. Using the same TomCar vehicle, and remote control turret, the Guardium has better sensors and software. Guardium is pitched as "smart" enough to be used in urban areas, and to serve as an emergency response vehicle. That is, these would be stationed along isolated stretches of border, ready to drive off to deal with any terrorists who had gotten through the fence. The Guardium would thus arrive before a human quick reaction team, which would be stationed farther away.

Guardium was seen as preferable to an earlier proposal; placing remotely controlled turrets in isolated areas, along with security cameras. If you spot some bad guys, the remotely controlled weapon can be used. South Korea and Israel have developed their own remote control weapon systems (SGR-A1 and Samson Jr., respectively). The United States has several existing remote control turrets to choose from, and is concentrating more on the array of sensors, the eyes and ears of the weapons.

South Korea wanted to use the system on it DMZ (Demilitarized Zone) border with North Korea. Israel wants to use them on the border with Gaza, which is often just an open stretch of desert. The U.S. wanted to use the systems for base defense in Iraq, Afghanistan and elsewhere. What has made these systems possible has been digital video analysis software that can detect people without human intervention. When that happens, a system operator is alerted, who decides if the person is hostile, and worth firing on. None of these systems proved entirely successful in practice.

However, these systems are vulnerable to attack and interference, which are the main reasons for not using them. Unless the cameras, and other sensors (sound, heat and seismic) can pick up hostiles far enough away, the remotely controlled weapon can be destroyed, along with many of the sensors, thus blinding the operators.

Meanwhile, both the U.S. and Israel have developed smaller armed robots. The American systems is called Swords (Special Weapons Observation Reconnaissance Detecting System). This is a 125 pound remotely controlled vehicles (they look like miniature tanks), armed with a 5.56mm machine-guns and 350 rounds of ammo). Also known as Talon IIIB, the army spent over a year testing them in the United States before sending some to Iraq. There they found there were many ways to mess with Swords. Many tricks didn't even damage the equipment (like having a child or woman come out and throw a towel or sheet over it).

Israel has a similar system, called Viper, that carries a 9mm machine pistol (an Uzi) and can carry explosives, along with the usual video camera and microphones. Both Swords and Viper do have their uses, like entering very dangerous situations (like a cave or building believed occupied by fanatical gunmen). The droids can also be used for guard duty in dangerous locations (where the enemy might get a shot off, or toss a grenade.) But no matter what you have the battle robots do, the mechanical grunts lack the same degree of situational awareness of a human soldier. The sensors used on droids (mainly visual and acoustic) are getting better, as is the software that can quickly evaluate what the sensors see and hear. But humans can also smell, and feel (on their skin), as well as use superior vision and hearing. Until the sensors get better, the combat robots will always be at a disadvantage. But if used with those disadvantages kept in mind, the robots do have their uses.

Civilians in general, and politicians in particular, are not eager to allow armed robots that operate on their own. They are more tolerant of unarmed robots who roam around on their own. Meanwhile, it's well to remember that smart airborne weapons have been in use for decades. The most common is the cruise missile, which is given a target location, and then flies off to find and destroy the target. Again, not too scary. But a UAV that uses the same technology as smart mines (sensors that find and software that selects, a target to attack) is alarming. What scares people is that they don't trust software. Given the experience most of us have with software, that's a reasonable fear.

But the military operates in a unique environment. Death is an ever-present danger. Friendly fire occurs far more than people realize (or even the military will admit). Combat troops were reluctant to talk about friendly fire (mainly because of guilt and PTSD), even among themselves, and the military had a hard time collecting data on the subject. After making a considerable effort (several times after World War II), it was concluded that up to 20 percent of American casualties were from friendly fire. So military people and civilians have a different attitude towards robotic killing machines. If these smart UAVs bring victory more quickly, then fewer friendly troops will be killed (by friendly or hostile fire). Civilians are more concerned about the unintentional death of civilians, or friendly troops. Civilians don't appreciate, as much as the troops do, the need to use "maximum violence" (a military term) to win the battle as quickly as possible.

The air force has good reason to believe that they can develop reliable software for autonomous armed UAVs. The air force, and the aviation industry in general, has already developed highly complex, and reliable software for operating aircraft. For example, there has been automatic landing software in use for over a decade. Many UAVs regularly, and effectively, employ this automatic takeoff and landing software. Flight control software handles many more mundane functions, like dealing with common in-flight problems. This kind of software makes it possible for difficult (impossible, in the case of the F-117) to fly military aircraft that would otherwise not be controllable by a pilot.

Weapons guidance systems have long used target recognition systems that work with a pattern recognition library that enables many different targets to be identified, and certain ones to be attacked. To air force developers, autonomous armed UAVs that can be trusted to kill enemy troops, and not civilians or friendly ones, are not extraordinary, but the next stop in a long line of software developments.

Army commanders are pushing for autonomous air and ground vehicles that would patrol, and look for potentially hostile patterns (someone carrying weapons, or should not be there, or firing on the robot.) This patrol duty is not only the most dangerous for the troops, but also tedious and likely to bore troops into inattention. This has been found very common with troops patrolling roads looking for roadside bombs. The robots don't get bored, and commanders don't have to write letters to the parents of dead robots.