At NASA's robot challenge, don't mistake trouble for failure

Teams struggle in autonomous robotics challenge but advances are made

After years of research and late nights, a team of scientists set up their robot Thursday to take on a NASA-funded autonomous robotics challenge.

They switched on their robot, stood back and waited for the machine to begin its two-hour effort.

But the robot failed.

Without ever moving off its platform or moving an inch, the robot's challenge was over for this year.

Members of Team Survey, a group independent researchers in Los Angeles, follow their robot through a field as they show what it can do during NASA's Sample Return Robot Challenge at WPI in Worcester, Mass. (Photo: Sharon Gaudin/Computerworld)

"We knew immediately," said Fred Zyda, a software developer on Team Survey, a group of individual scientists from Los Angeles. "That was quite disappointing. There's a speaker on the robot that announces what it's doing. We turned it on and it said, "Cameras failed." We knew."

Team Survey was participating in NASA's Sample Return Robot Challenge at Worcester Polytechnic Institute (WPI) in Worcester, Mass. In its third year, 18 teams took on the challenge to build a fully autonomous robot that can traverse an area the size of one and a half football fields, find objects and retrieve them.

NASA is looking for technologies that can help build improved robots that would eventually be used on Mars or on asteroids.

Larry Cooper, NASA's program executive for the challenge, said the space agency isn't likely to use any of the technology demonstrated at this week's challenge, but the teams are heading in the right direction.

"I'm sure we won't use any of these on Mars, though some of the people who worked on the robots may put something on Mars some day," he added.

At the end of the day Thursday, the second day of the contest, not one of the teams had successfully completed the challenge. Most of the teams will have a second try at the tasks on Friday, when rain is in the forecast. Their first official attempts were unsuccessful.

Don't mistake these problems for failure, though, said Jascha Little, a mechanical engineer on Team Survey.

"It's hard to appreciate how difficult this is," he told Computerworld. "That's just development. We definitely learned a lot of things."

It's been a long road for Team Survey, which successfully passed through Phase One of the challenge last year and was tackling Phase Two this year. The robot was supposed to be searching for objects, which ranged from a blue rock to a red hockey puck and an orange PVC pipe.

To get ready, the team's seven members worked nights and weekends, totaling nearly 40-hour weeks, on top of their full-time jobs, in the past few months leading up to this week's challenge.

Zyda said he didn't mind the long hours because he loves working on interesting problems. Building an autonomous robot -- the hardware and the software to run it -- qualified.

The team used some off-the-shelf software for the three-wheeled robot, which has nine onboard cameras. For instance, they used the Robot Operating System, also known as ROS, open source software that provides basic robotics functioning.

Team Survey's five software programmers also wrote a lot of their own code. Over the past three years, they've probably written about 10,000 lines of code, estimated Zyda. Most of them used C++ programming language. Little, the mechanical engineer, also wanted to try some of the coding so he learned to use Python, a programming language geared more for beginners.

The programmers, for instance, created custom code for the robot's three independently moving wheels.

Many three-wheeled robots use the back wheel as a castor, meaning that it drags along instead of driving. Team Survey's robot, though, has three wheel pods, each with its own steering motor, which is a wheel motor and two encoders that monitor the machine's steering angle and judge the distance the wheel has rolled.

The programmers needed to write code that would enable the robot to control all of its wheels.

So what caused Team Survey's trouble that knocked them out of the official competition?

It was a Firewire interface for the robot's nine cameras.

"One out of 100 times it doesn't connect when we start it up," Zyda said. "We started it today and the cameras didn't come up. If you can't see, you can't go anywhere."

While the robot failed its challenge, NASA and WPI challenge officials allowed the team to fire it up again and run as a demonstration, but out of contention for any award money this year.

The robot started up on the second try and spent two hours maneuvering around the field, motoring easily and fairly swiftly over the course. Zyda noted that their robot found several objects but was unable to pick up any of them without assistance.

"It shows a lot of promise," added Zyda. "It didn't hit anything. It didn't slip its wheels ... Its motions were a lot smoother this year.

Team Survey plans on competing again next year.

To do that, they'll do more testing and will focus on system integration, one of the biggest challenges for robotics at this level. With so many sensors, cameras and other systems, it's a lot to link together into a seamless operation.

This article, At NASA's robot challenge, don't mistake trouble for failure, was originally published at Computerworld.com.

Sharon Gaudin covers the Internet and Web 2.0, emerging technologies, and desktop and laptop chips for Computerworld. Follow Sharon on Twitter at @sgaudin, on Google+ or subscribe to Sharon's RSS feed. Her email address is sgaudin@computerworld.com.

See more by Sharon Gaudin on Computerworld.com.

Read more about emerging technologies in Computerworld's Emerging Technologies Topic Center.

Tags hardwareNASAEmerging Technologieshardware systemscomputerworldgovernmentGovernment/Industries

More about FredGoogleNASATopic

Comments

Comments are now closed

Amazon vs. Google vs. Windows Azure: Cloud computing speed showdown

READ THIS ARTICLE
DO NOT SHOW THIS BOX AGAIN [ x ]