NASA announces competition to improve Robonaut's sight

NASA's got some of the sharpest minds in the world (actual, you know, rocket scientists), sure, but they'll be the first to tell you that sometimes you've got to look outside for the best solution to a complicated problem. In recent years, that's meant the organization has partnered with the likes of SpaceX to help transfer materials to the International Space Station. The desire to look outside has also taken the form of competitions, which, in the past, have sought to improve the efficiency of solar arrays and help better understand the massive amounts of data collected from various missions over a 30-year period.

This latest competition, a partnership with TopCoder, deals with the unspeakably appealing category of space robots, aiming to improve the vision of NASA's head of menial space station tasks, Robonaut. At present, the 'bot's got the sort of sight problems that would have no doubt barred its fleshier counterparts from making their way through the training program.

"We've already set something up, but it didn't work great, because it was pretty brittle," says NASA robotics engineer Julia Badger of Robonaut's vision. "It wasn't a great solution for being able to do it with the lab lighting conditions on Earth, and then the different cameras and lighting conditions and radiation on the space station. So that's why we're open-sourcing this. We want to know what people want to know out there and how they can solve the problem."

The competition centers around helping the robot determine the state and location of buttons and switches located on a taskboard by identifying their corresponding LEDs. It's a real problem for which NASA has yet to find an elegant solution -- so the organization has decided to look outward. "Whenever you can bring in non-domain experts to a problem that you have, you have a higher likelihood of getting more valuable solutions and approaches that haven't been thought of," explains NASA's Director of Advanced Exploration Systems, Jason Crusan. "If you stick with the normal players -- even if they are experts -- you often stick to your old ways and don't think of new ways to look at the problem."

It's not, he adds, a competition simply for the sake of competition. "We don't run challenges for the purpose of outreach or just to see if we can get a neat solution. If we're going to run a challenge and it's successful, the idea is that it should be easily integrated into the project."

To get involved -- or to just learn more -- please locate and press the source link below.

Show full PR text

TopCoder and NASA Tournament Lab Champion High Value Output-Based Challenges, Now Look to Bring Sight Capability to ISS Robot

TopCoder®, Inc., the world's largest open innovation platform and competitive community of 470,000 digital creators, today announced two new marathon competitions hosted through the NASA Tournament Lab (NTL) that will look to continue delivery of output-based high value returns in the most cost-effective and measurable software development process currently available to government agencies. The competitions will center on NASA's famous space robot Robonaut 2 aboard the International Space Station (ISS). The initial challenge will focus on enabling Robonaut 2 to interact visually with different types of input devices the astronauts use on the Space Station.

Successfully completed projects for NASA on NTL's open and transparent environment for sophisticated technology development include the recent Longeron Shadowing Optimization Challenge, a $30,000 open innovation competition to make the energy-gathering solar arrays of the Space Station more efficient which attracted a pool of more than 4,000 registrants and delivered 459 competitors who produced 2185 unique solutions in less than two weeks, the Space Med Kit competition which delivered a 360 x improvement over the existing NASA algorithm and garnered 2,833 distinct code submissions from 1,095 participants at a cost of $24,000, and the Planetary Data System Idea Challenge, a $13,000 challenge in deployment and future use mode which found novel ideas for new applications to help general users to search, display and understand the more than 100 terabytes of recorded space images, telemetry and models from 30 years of planetary missions. Additionally, the Food Intake Tracker (FIT) iPad App Conceptualization and Voice Command Idea Generation competitions recently finished with architecture competitions beginning soon. The ISS-FIT competition program is creating new dietary tracking applications for use by astronauts.

"We at TopCoder are thrilled to be again working with Harvard and NASA to provide creative and inventive support of NTL's mission to deliver critical solutions while recognizing the need to return high taxpayer value," said Rob Hughes, President and COO of TopCoder, Inc. "We feel that our collective efforts on Robonaut 2, along with our successful Longeron, PDS and FIT programs have exposed the exciting prospect that game changing solutions can be created at reasonable cost by a new generation of engineers, programmers and technologists which is critical to innovation in the public sector and beyond."

About the Robonaut 2 Competitions

Competitors from around the world are welcome to register at http://community.topcoder.com/longcontest/?module=ViewActiveContests . Each phase will run three weeks from start to finish.

This first phase of the two competitions is to enable Robonaut 2 to recognize the state and location of several buttons and switches on a dashboard fitted with LED lights. Contestants will be tasked to work with different sets of imagery from NASA's lab on earth and on the Space Station, as well as in a simulator. The successful algorithm application must work with each of several different camera systems and varying lighting conditions within each environment.

A second contest will ask competitors to use the "seeing" algorithm produced in the first challenge in order to write an algorithm that actually controls the robot's motion. The algorithm will need to "see" an object, recognize it, and correctly operate and interact with it in the most efficient and safe manner possible.

A long term goal is to further develop the capabilities of robots to interact with input devices and move closer to enabling them to one day perform repetitive, monotonous tasks in place of astronauts. Full details are available at http://www.topcoder.com/iss/robonaut/.