NASA's robot event challenges, robots, engineers

12.06.2015
Despite robots that passed by, or even ran over, the objects they were searching for and others that never got off the starting platform, scientists said this week's NASA robotics event is pushing the technology forward.

NASA's 2015 Sample Return Robot Challenge is intended to encourage researchers to develop robotics and autonomy software that could someday be used on Mars or the moon.

The challenge, which involved 16 robotics teams from around the country, requires the robots to navigate without human guidance around a large field and find objects, ranging from a ball to a red rock and a yellow PVC pipe.

The robots then have to scoop up the found object and return it to the starting platform.

The event was held this week on the Worcester Polytechnic Institute campus in Worcester, Mass.

NASA hopes the challenge will provide better technology so robots can work on their own on an asteroid, moon or another planet to find and retrieve objects or caches of objects that other missions leave behind.

"This competition isn't necessarily about the ability of the robot to rove or sense or detect an obstacle," said Sam Ortega, program manager for NASA's challenges project. "Those abilities have been developed. We're really trying to test here the ability to put all of these systems into one platform. Here we're raising the fundamental technology bed for sensing robots and autonomy.

"Anytime you start doing something on the very cutting edge of technology, it's going to be tough," he added.

On Wednesday, the first day of the competition, nine teams competed in Level 1, the lowest level of the challenge, which required the robots to autonomously move off a starting platform in the middle of a field and find and retrieve a single object. The robots, which generally have vision of up to 10 meters, were given a general location of that first object.

Only three of the nine teams made it off the starting platform. None were able to find the first object.

On Thursday, the second day of the challenge, two Level 2 teams, which advanced from Level 1 last year, had a bit more success.

Team Survey, made up of engineers from Silicon Valley, found its first object but then was unable to find any of the others scattered in the field. It also ran over the yellow PVC pipe it was looking for.

A team from West Virginia University won a $100,000 prize from NASA as the only team whose robot was able to find the required three objects.

On Friday, with all of the Level 1 teams competing, none successfully retrieved an object and returned it to the platform, so none advanced to Level 2.

"Sometimes it looks frustrating, and the teams are frustrated, but it actually went really well," said Colleen Shaver, assistant director of WPI's Robotics Research Center and manager of the NASA challenge. "We saw teams detecting samples, teams attempting to pick them up and in several cases actually collecting the sample. It's hard to see just how difficult it is to plan for the weather, like the sun, which can be aimed right into the camera. It's impressive to see how much the teams did today despite the variables."

She noted that some of the teams' robots didn't make it off the starting platform, including a team from MIT, because they were taken down by basic problems.

One Level 1 team whose robot made it off the starting platform and was headed toward its first object before it ran into trouble was made up of students from Schenectady High School in New York.

Unlike the other teams, which are from major universities or made up of professional engineers, the high school teammates worked on the project after school.

"It just shows that this challenge can reach anybody," said Shaver said. "Innovation can be done anywhere, whether it's high school kids or professional engineers. It's how you look at the problem and how much energy you're willing to put into it."

Ken Stafford, associate director of robotics engineering at WPI, noted that what the robotics teams were faced with doing in a short amount of time in the challenge, robots on Mars, for instance, might have a week or more to do.

"There's no rush on the moon or Mars," he added. "We can't give them that much time so this is actually much tougher than it would be in the real world."

(www.computerworld.com)

Sharon Gaudin