We're So Vain: Who Needs a Humanoid Robot Anyway?

On February 15, 2012, man and machine met in space when NASA's Robonaut 2 shook hands with American astronaut Daniel Burbank. The robot then offered a more specific, if silent greeting; it said, "Hello, World" in American Sign Language. At the same time the message appeared on the robot's Twitter feed, which currently has more than 53,300 followers. Robonaut 2 isn't quite on par with HAL9000, the super intelligent computer from 2001: A Space Odyssey that turned on its human crew mates but it does speak volumes about the possible intersection of man and machine in future space exploration.

Robonaut 2, or R2 as it is known within NASA, is a $2.5 million droid developed by the space agency in partnership with General Motors that looks like a human from its would-be waist up. It has a torso, arms, and a head packed with instruments and a sophisticated camera.

R2 arrived on the International Space Station in February 2011 on the last flight of the Discovery orbiter. It's designed to assist astronauts in their more complex chores and generally help with daily upkeep of the space station. For the time being, R2 is fixed to a pedestal, but upgrades will give it legs to walk the halls of the ISS and work outside in the vacuum of space. It's descendants will follow in R2's footsteps. The Robonaut series of dextrous humanoid robots will eventually accompany humans, or go in the place of humans, on deep space missions. They will travel places we can't go and do jobs we can't manage in environments we can't survive in. They will explore alien worlds. Some, like R2's twin on Earth, will be fixed to wheels to facilitate traversing terrain while others will walk.

It's an exciting prospect – humans and their analogues exploring space together far better than either could alone. But it begs the question of whether or not robots are the best option for future space exploration.

As advertised, the Robonauts display incredible dexterity. Their humanoid arms and hands have the added advantage of streamlining components; they can share tools with astronauts. Their human shape also means they'll fit nicely in the same spaces as their fleshy companions. So far, the rational of humanoid robots is sound but the other part of the Robonaut package, the yet-to-be-seen walking function, is far less practical than dexterity.

We use a lot of muscles just to stand, and walking is incredibly balance-intensive. We don't think about it because right from the moment we learn to walk we're training the stabilizing muscles that keep us upright. A robot doesn't have stabilizing muscles. It may be more solid than we are, but mimicking the human gait is a real challenge for robots.

Honda's ASIMO is the most human-like robot out there. In November 2011, the company unveiled its first version with autonomous behavior control technology. This robot can continue moving without input from an operator, interpret human body language, and its improved intelligence and physical proficiency make it the most impressive robot capable of working in a human environment ever built.

If assimilation into, say, an office is the goal, than a humanoid walking robot is a fine goal. But is it necessary in space? Humans are bulky. If a computer can mimic the basic functions of our intelligence, why make the robots bulky rather than compact?

The Mars rovers Spirit and Opportunity are a great example of the power of non-humanoid robots. They are autonomous in that they can make basic determinations about their surroundings without being told, things like "that rock is too big for me to climb over, I'm going to stop and ask for help instead of smashing into it." Sure, they need help from operators and rely on programming to move forward in a mission, but they are basically robotic geologists.

What they don't have is human reasoning—the real problem with robotic exploration. Out present day computing power is impressive but limited. It just isn't at the level where we can anticipate having highly intelligent robots anytime soon. Common sense and an understanding of the world around us is a wonderful human trait that is really hard to program. We know that water is wet and that when we touch a hot stove we get burned. These are things a robot has a hard time "understanding." That human experience is what gives us an edge as explorers, but it's not something easily reproduced.

R2 is certainly a feat of engineering and computing, and might be a fine addition to the ISS, but what about other places in space? A humanoid robot would be fine on Mars, but it would be crushed on Venus; its feet could sink into the methane lakes on Titan. What would happen if it tripped over a crack in Europa's surface and couldn't get back up?

Our robotic ambassadors to other planets need not be in our image. R2 might be a powerful tool – as we humanize it, future generations might be inspired to pursue robotics. But that doesn't make it a practical way forward in robotic space exploration. In light of the recent cuts to NASA's budget, it's more important now than ever that the space agency pursue the simplest and most efficient way forward in its robotic exploration of the Solar System.

Connections

NASA's New Robot Astronaut Can't Walk Yet
Humanoid Space Robot Will Compete for Jobs on the Space Station
First Mission of NASA's Highly Advanced Space Robot Is Tweeting