The Robotic Right Stuff: What It Takes to Become the First Robot Astronaut

By Eliza Strickland | October 29, 2010 1:23 pm

When the space shuttle Discovery launches on Thursday (weather and technology permitting), it will be ferrying an unusual passenger to the International Space Station: Robonaut 2. This humanoid robot was designed by NASA and General Motors to work alongside astronauts on the space station, and could eventually take over some tedious or dangerous tasks.

Human beings who dream of becoming astronauts acquire things like advanced science degrees or the ability to fly jet planes in hopes of catching NASA’s eye and being chosen as astronaut candidates. If they do become candidates, there’s still scads of training before they can take a flight up to the ISS. But how does a robot qualify for and prepare for that trip to orbit? DISCOVER spoke with Marty Linn, General Motor’s principal engineer of robotics, to find out.


Physical Fitness: Human astronauts have to pass the NASA long-duration space flight physical to prove that they’re healthy, fit, and strong enough for astronaut duties. Robonaut 2 has to be pretty strong, too: Here on Earth, he proved that he can do arm curls with 20-pound free weights. “The limitation is grasp strength,” says Linn. “The weak link is how strong the fingers are.” The robot didn’t have to spend any time on the treadmill, though, because this model doesn’t have lower limbs—it’s simply a torso with arms and a head.

Intelligence: To be honest, R2 (as its buddies call it) isn’t that bright—it can’t make independent decisions. NASA’s top priority for the experimental bot is guaranteeing that it won’t pose a threat to the astronauts or the space station, so for now R2 will be under the strict control of astronauts and ground crew. R2 “isn’t going to go berserk,” Linn stresses, but it’s still nice to have an off switch. He also explains that the robot’s actions can be programmed joint by joint, or it can be controlled by a tele-ops system, in which an astronaut dons the tele-ops gear and puts the robot through its paces by moving her own arms or head.

Vision: NASA has always paid careful attention to the eyesight of its astronaut candidates, and only recently decided that people who have gotten laser surgery to correct their vision can still be considered for the job. R2’s vision is top-notch. It’s equipped with high-resolution digital cameras, can detect motion and distinct objects, and has a 3D mapping tool to allow it to determine where objects are in space. It also has lower resolution cameras for tele-operation, Linn explains, which “allow the operator to see through the eyes of the robot.”

Dexterity: This isn’t a top priority for NASA when selecting human astronauts, but it was a big factor in R2’s selection. Linn explains that R2 is one of the most nimble-fingered robots ever built. “Traditional robots don’t do well in an unstructured environment, where the objects aren’t rigid,” he says. “It’s very difficult for them to grasp a flexible object like a space blanket.” R2 can handle not only flexible objects, but can also deal with the small switches and cables that astronauts routinely work with on the space station.


Microgravity: An important part of astronauts’ training is exposing them to the weightlessness they’ll experience in orbit. NASA handles the prep with a modified jet plane that flies in long parabolic arcs to produce brief, half-minute bursts of weightlessness; fully suited astronauts also train for zero-G spacewalks in a water tank called the Neutral Buoyancy Laboratory. That’s one big gap in R2’s preparations for spaceflight: It has no experience with microgravity. Obviously R2 couldn’t be dunked in the pool, and the parabolic flights wouldn’t provide enough time to do meaningful tests of R2’s capabilities, Linn says. “The best experiment will be on the space station,” says Linn. That’s when the engineers will learn whether the procedures they tried on Earth will translate to orbit.

Health Check: The flight crew for the ill-fated Apollo 13 mission was famously changed at last minute, when pilot Ken Mattingly was grounded due to his exposure to German measles. The final health checks for R2 were a little different. The engineering team wasn’t worried about biological contaminants, but they did need to check for chemical hazards. “The robot did go to White Sands [Test Facility], and it went through a bunch of outgassing tests,” Linn says. “They put him in a chamber to see what kinds of chemicals were coming off this thing, to make sure it was all safe for humans to breathe.”

Spacesuit Fitting: Just like human astronauts, R2 had to get suited up for space. Its original “skin” of Lycra and neoprene didn’t meet the Space Station’s stringent flammability requirements, so it was swapped out for a fire-resistant spacesuit material. The new skin is made of Kevlar, Teflon, and Nomex.

Practice for Mission: Astronauts practice for their missions with shuttle simulators and full-sized mockups of the Space Station’s flight deck and payload bay. Operating on a more modest scale, R2 has practiced with a task board that has “typical switches that are found on the space station, and different kinds of buttons,” says Linn. Since R2 is the first of its kind, it won’t be given any crucial tasks aboard the ISS; once it’s unpacked and set up it will be put into action so that engineers can study its capabilities and limitations in zero-G. Eventually, they may let R2 try its hand at one tedious task: cleaning. “We’ve experimented with various kinds of cleaning tasks, like cleaning air filters and wiping down surfaces,” says Linn.

Preparing for Travel: The astronauts who will fly on mission STS-133 won’t climb aboard the shuttle until Tuesday (assuming the launch date remains the same), but R2 is already stowed safely in the cargo bay. Here’s a time-lapse video from NASA’s Kennedy Space Center showing how R2 got prepped for flight and packed up:

The Future

Upgrades: NASA and General Motors have grand hopes for R2, and say this experimental bot can be upgraded to become more functional. Lower limbs would let the robot move around the space station on its own, and engineers have already experimented with a number of different movement systems. One strong argument for lower limbs: They would provide both stability and maneuverability, which could also allow R2 to take part in spacewalks. Says Linn: “From a time constraint point of view, it’s a lot easier to send the robot outside that for an astronaut to get all suited up and go outside.”

No Return: One thing that NASA didn’t have to worry about when preparing R2 for his trip was psychology. If you told human astronauts that they’d be taking a one-way trip to the space station, they’d be pretty upset. Not so with R2. This robonaut isn’t coming back, but will instead remain on the ISS for as long as the space station is in orbit. But it doesn’t seem alarmed about the prospect. As R2 recently proclaimed on its Twitter feed, “I’m living on the station as long as there is a station. ISS 4 LIFE!”

Related Content:
80beats: Japanese Consortium: We’ll Send a Humanoid Robot to Walk on the Moon
80beats: A New Crew Member for the Space Station: The “Robonaut 2”
80beats: Robonaut 2: Coming Soon to Space Stations and Assembly Lines Near You

Images: NASA

CATEGORIZED UNDER: Space, Technology, Top Posts
  • Anthony Pham


  • Grob Hahn

    Doesn’t it make some sense that we should send dozens of these things to Mars to get set up for human arrivals? Imagine the jobs created here controlling the actions of these on another world. Putting one on the ISS means we maybe don’t repeat the Skylab (apathy/political) disaster. I’m still shaking my head over this 100 year starship crap when we can totally circumvent that with this technology. Technology that is basically Off-the-shelf in 2010!

  • Jake Levee

    I think the idea of sending robots to mars has already been tried…and found marginally successful. From a travel and survival standpoint, robots make a lot of sense. However, our communications time lag is so great that we’d be lucky to have one accomplish minute tasks in a day. I think our hardware is ready, but we really need more work in AI. Once we have the intelligence to go along with the hardware, then it would definitely make sense to send bots to Mars and elsewhere. On that note, I imagine AI isn’t too far off in our future. I agree with the 100 year starship idea being crap. Why send humans, who have all the emotional baggage and physical limitations, when we can send intelligent bots who lack these ‘handicapping’conditions. In another 50 or so years, maybe even less, we’ll be there.

  • Demian Wilshire

    I agree with both the above comments. I wanted to add that it is my hope that in less than a 100 years mankind has established bases on both the Moon and Mars. I understand the increased feasibility of robots as explorers. I think that Robots will be our next step beyond Voyager 2 to exit and explore other star systems. I am hopeful that the experiments with R2 prove to be successful and we see more advanced models begin to take over a large part of our space exploration. Then when things are established at destinations then we can begin to travel to these places in order to establish colonies.

  • sam s.

    Why is R2 not part of the crew pics? I mean after all,it is a crew member(changing filters,cleaning)? And not to mention it is the first of its kind?I feel that if we are going to go in this direction,we should treat A.I.bots with respect and dignity?Someday we may be faced with a civil rights movement from A.I.bots….Mmmmm?

  • http://n/a anonymous

    R2 isn’t part of the crew pics because he’s been packed up since July. They could have used one of the other units for pics though, probably should have.

  • ojak

    As neat as all this is, it’s still too small-scale and uninspiring. Look, we went to the moon in less than a decade and the public will never really give two sh*ts about a “toy” robot. After working on DARPA grants for many years, I can confidently say that the primary problem is that researchers (i.e. scientists) are generally bad at tangible deliverables. It’s always easier for grantees to deliver on loosely specified long-term “theoretical” goals than on concrete, discrete achievement.

    I guess I’m saying that NASA lacks a charismatic visionary to set bold, uncompromising deadlines. We need a true visionary that will put up… otherwise the public will never back the investments necessary to truly excel in this field.

  • michael-savory

    That’s one good looking robot! He’s got the right stuff. By the way, check out this Physics site

  • Earp

    Jake Levee said:
    ‘I imagine AI isn’t too far off in our future.’

    Unfortunately, it is, seeing as there has been virtually no progression in AI in the last 20 years. Making computers better at doing tasks and acting autonomously is not AI.

    We could send robots to mars no problem in 50 years – but at our current state of AI development, they wont be AI. Some type of breakthrough (totally outside improving processors and programming) will be required for this (perhaps simulations of actual brains).

  • Frank

    Here’s a spherical image of Robonaut and its development team in the SSPF at KSC on August 12th. Be sure and look up and down. Enjoy!

  • Science Box

    Great! May I copy this article on my blog?

  • georges

    Sending robots on Mars has already been done folks. Mars Explorer, etc …
    What would be new is a robot with AI on Mars.
    But let’s first do a robot with AI on earth, or even just a computer with true AI on earth!!!

  • ghost

    We sent robots to mars in 1971. Yes, 1971. we lost contact with the first two, but we did succeed in 1976.

  • Folder Printing

    Love the article… thanks for the post!


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!


80beats is DISCOVER's news aggregator, weaving together the choicest tidbits from the best articles covering the day's most compelling topics.

See More

Collapse bottom bar