Robots are officially on the battlefield—UAVs like the Predator and Reaper patrol the skies while militarized bomb-disposal robots like the Talon detonate explosives on the ground. But where are the humanoids? Roboticist and author Daniel H. Wilson makes the case for a humanoid robot army.
A humanoid robot is a general-purpose robot that looks a lot like a person, complete with a head, torso, arms and legs. The "total package" humanoid can walk bipedally, like a person, and use its hands to dexterously manipulate objects in the world.
Current prototypes like the Honda ASIMO can deliver tea and politely shake hands with their human masters, but based on some great sci-fi movies, humanoid robots are supposed to be terrors on the battlefield—walking titanium endoskeletons crunching over human skulls and mowing down pesky humans with massive handheld Gatling guns.
Will we ever really see a humanoid robot army? I think so, and here are my top five reasons why.
1. There is a one-to-one mapping between the human and the humanoid body.
Robots aren't yet smart enough to play without supervision. That's why human soldiers control unmanned aerial vehicles from thousands of miles away by twiddling joysticks. It isn't easy, but flying a plane through empty space is child's play compared to maneuvering a ground-based robot through rubble and wreckage. And what if you need to do something more complicated than just stepping over a curb, like defusing a bomb?
It's called telepresence. With telepresence, a person feels as though they are the robot by controlling the robot's body and seeing through its eyes. Human-shaped robots are infinitely easier to manipulate because there is a one-to-one mapping between man and machine. Instead of shoving around a non-intuitive joystick, slide your hands into gloves that map your fingers to robot fingers thousands of miles away. Now put your human expertise to work, without putting your human butt in danger.
2. Humanoid robots take advantage of human environments and equipment.
Nothing beats a tank for crossing the desert, but what about crossing a living room? Every human city is designed for a very specific type of animal: homo sapiens. We humans come in a very specific range of sizes and weights, and our environments tend to have specific temperature, vibration and noise limits—all of which simplify the problem of designing a robot. Humanoids are naturally suited to navigating environments designed for humans; they can walk through doorways, climb steps, and see over counters and furniture.
Along with our cities, most military supplies are designed for use by humans. That means a humanoid robot can wear human body armor, boots and camouflage. In addition, it can fire standard-issue weapons and ammunition, removing a need for specially-designed weaponry. Humanoids could also potentially pilot human vehicles. Rather than creating an autonomous vehicle from scratch, just put a humanoid robot in the driver's seat of a standard vehicle. And when a robot squad is on the go and under fire, it always helps to be able to scavenge enemy weapons and improvise. The infrastructure is there, and humanoid robots exploit it.
3. Humanoid robots are easier to train.
War is largely improvised, and that means learning new tricks on the fly. So, how do you teach a robot comrade how to defuse a new type of coffee-can landmine? Without a degree in engineering, you probably don't. But given a humanoid robot, intuitive training approaches are available to regular soldiers. An easy but tedious method is to physically push the robot's limbs through the proper series of movements. Alternately, take direct control through teleoperation and then perform the activity yourself. The robot then just needs to remember how you did it.
Ideally, however, a robot can be trained just like a person—by watching. Robots who learn by demonstration can be quickly trained by ordinary people who do not speak robot-ese or do any programming. That's because it's how we learn from each other. The trainer simply performs the task (e.g., a flying scissor kick) and the robot watches and intuits how to do it. Humanoids are much better at learning by demonstration, thanks to that one-to-one mapping between its body and yours.
4. Teamwork is easier between humans and humanoids.
It is doubtful that robot armies will operate completely autonomously in the near future. Human-robot teams will likely be the norm, as they are today. Therefore, it's important to make sure that human and robot allies can work together without stepping on each others' toes. And that means they've got to have good communication.
Human combat teams communicate and cooperate using language and gestures, and by paying attention to each other's facial expressions and emotions. Robot warriors that recognize human body language will be able to make fast decisions in loud, hazardous environments. Perhaps even more important, a human soldier should be able to understand what a robot is thinking naturally, by reading its body language instead of looking up an error code in an instruction manual. Using the highly familiar human form-factor creates a natural communication channel that allows humanoids to cooperate with humans in chaotic environments where split-second decisions are the norm.
5. The locals could potentially interact with humanoid robots.
War is becoming less about conventional fighting on a mass scale and more about cultural awareness. Last month, President Obama unveiled plans to send hundreds of "social scientists" along with soldiers to Iraq, to counsel the military on local customs. Relative to the faceless robots currently in use, a humanoid robot provides the opportunity for some kind of natural human interaction with non-combatants. Instead of an impersonal unmanned ground vehicle wrecking through walls or an unmanned aerial vehicle dropping bombs from afar, humanoid robots (armed or unarmed) could patrol areas wearing local garb, speaking the local language, and obeying local customs. How P.C.—or just freaky—is that?
On the other hand, humanoid robots can be horribly terrifying.
Mind games are a part of every battle. During World War II, aviators painted snarling teeth on the noses of their fighter planes. Nowadays (and back then), bombs have funny messages written on them, like "Boom shacka lacka," and "You want fries with that?"
Now imagine the enemy reaction on Robot D-Day, when thousands of super-powered humanoid robots march out of the crashing surf, bullets plinking harmlessly from their razor-sharp gilded breast-plates as death metal blares from their metal mouth speaker grilles.
Terrified yet? Well calm down, sissy; humanoid robots aren't on the battlefield, yet. But they might be soon, thanks to their natural ability to communicate and cooperate with humans, the ease with which they can operate in our environments and use our tools, and the terrible fear that blossoms in the heart of man upon laying eyes on the great and horrifying visage of the humanoid robot war machine.
Machines Behaving Deadly: A week exploring the sometimes difficult relationship between man and technology. Guest writer Daniel H. Wilson earned a PhD in Robotics from Carnegie Mellon University. He is the author of How to Survive a Robot Uprising and its sequel How To Build a Robot Army. To learn more about him, visit www.danielhwilson.com.