ROBOT: Welcome to Altair 4, gentlemen.
TOM BEARDEN: From Robbie the Robot in MGM's 1956 sci-fi classic "Forbidden Planet."
ROBOT: They shut down the main area. We'll be destroyed for sure.
TOM BEARDEN: To R2D2 and C3PO in George Lucas' "Star Wars," robots in space are a time-honored Hollywood tradition. The National Aeronautics and Space Administration is trying to help reality catch up with science fiction.
This is Robonaut, one of several NASA research projects pushing the boundaries of space robotics. Robonaut's extraordinary hands were designed to help astronauts with the routine chores of EVA, Extra Vehicular Activity.
Astronaut Nancy Currie worked alongside two Robonauts during ground testing. The plan is to essentially have Robonauts in space assisting human astronauts, saving time and energy.
NANCY CURRIE: We've created it like a human being. And in fact, we've made it better than a human being because, for example, the arms are much longer than like a 95th percentile male arm, because one of the things that we have problems with in space in a suited environment is a reach capability.
So, they intentionally made Robonaut's arms longer so he gives longer reach. But the dexterity on Robonaut is almost equivalent to the human hand, and that makes so that you don't have to make specialized tools for the robot to use because currently some of our manipulators in space have literally, you know, tens or hundreds of tools that they have to use.
TOM BEARDEN: So it becomes an all-purpose machine.
NANCY CURRIE: Absolutely.
TOM BEARDEN: It wasn't easy building hands that are almost as flexible as human hands, according to engineer Bill Bluethmann.
BILL BLUETHMANN: Within the thumb it's got three joints. It pitches, it yaws, and then the base pitches and the distal joint pitches. The same is true for the index and middle fingers they both yaw, the base pitches, and then also the middle joint of the finger moves independently.
And the last thing, which is sort of a subtle feature of the human hand that we designed into this is the palm also cups. So you can see the palm cups, and that gives you really helps you align tools to the palm better.
TOM BEARDEN: This is Ranger, a non-humanoid robot that uses tool attachments instead of hands to accomplish complicated tasks. It's an alternative approach to solving construction problems in space. The latest version is currently being tested at the University of Maryland in a neutral buoyancy tank which simulates weightlessness.
Ranger's developers had hoped it would be used to service the Hubble Space Telescope, and operators are practicing that task. Hubble will fail without such a servicing mission, but the decision whether to send astronauts or robots, or whether to service Hubble at all, remains embroiled in controversy. David Akin designed Ranger.
DAVID AKIN: Hubble was designed to be serviced by astronauts in space suits. It was never designed to be serviced by robots. So it is a bit of a challenge for a robot to use the same interfaces and do the same tasks that astronauts would normally do.
TOM BEARDEN: Working on a mockup of the lower portion of Hubble, Ranger unfolded its arms, then tried to grab another mockup of one of the instrument packages inside the telescope that was scheduled for replacement.
REMOTE OPERATOR: How much clearance do I have right now between here and the door?
TOM BEARDEN: Remote operators in the control room spent nearly two hours just maneuvering Ranger's arm before its gripper tool finally locked onto the attachment point of the camera. The work always goes slowly, partly because work in space is deliberately slow to avoid accidents or damage, and partly because real-life robots are almost always slower than Hollywood robots.
And unlike Gort, the all-powerful robot enforcer of galactic peace in "The Day the Earth Stood Still," who was beyond human control, both Ranger and Robonaut are directly operated by humans. Both devices are remotely controlled, in Ranger's case, with joysticks, while Robonaut's controller uses gloves and virtual reality goggles. But that technique, called tele-operation, won't work when the robot is on another planet.
If the Bush administration's plans to send astronauts back to the moon and to Mars are to be realized, much more sophisticated and independent robots will have to be developed. John Stevens of Lockheed-Martin Space Systems in Denver says it would take too long for control commands to reach a tele-operated robot on Mars.
JOHN STEVENS: Fundamentally, the time distance between Mars and Houston, so long as it's 40 minutes or longer, you can't operate those on Mars from Houston, so they have to be autonomous.
TOM BEARDEN: Lockheed-Martin is one of 11 companies that recently submitted broad conceptual ideas for how expeditions to Mars and a return to the moon might be organized. The first vehicles to land would contain autonomous constructor robots that would join the landing vehicles together into living quarters. Those would have to be functional before humans even took off.
Bulldozer-like robots would construct berms to keep the dust from those later landings from damaging the structures. Lockheed would also land several autonomous exploration vehicles at various places long before the astronauts would arrive. They would gather samples and eventually rendezvous at the astronaut landing site, where scientists would analyze their cargoes.
So those robots have to be autonomous for fairly long periods of time?
JOHN STEVENS: Not only do they have to be autonomous to make sure everything's ready right when they're there, but they also have to operate for long periods of time. In some cases there'll be two to four years of wait time for those robots before they're actually used by the humans.
TOM BEARDEN: Getting robots to be that autonomous for that many years is a formidable challenge. Dave Lavery is in charge of solar system exploration at NASA headquarters. Why is it so difficult to get robots to be smart the same way that people are smart?
DAVE LAVERY: For you and I to look at a red cup sitting on the table, for example, and understand that that is a red cup that I see, is a very easy thing. Computers just don't work that way. We have a very, very hard time understanding how to tell a computer, "Take the thing that you see in the camera, discriminate everything else in the scene and understand that that blob that's in front of you that sort of is the same color, it happens to be red and has certain edges, that that shape that it defines is a cup."
That's what we have to do to the computer, but we have to tell it not just how to recognize a cup, but a pencil, a table, a chair, a rock, a boulder, a pebble. That is an enormous problem.
TOM BEARDEN: But all of those problems will have to be overcome if astronauts are to ever walk on Mars. If some of these plans actually come to fruition, it may finally resolve one of the longest running arguments in planetary exploration. Some scientists have long argued that robots are better explorers than people. They don't require food, air, and water, and aren't as fragile.
Supporters of manned space flight point out that no robot can even begin to approach the flexibility of the human mind in dealing with unexpected problems. The ultimate answer may be that robots and humans will explore space together, each relying on the other strengths.