MIT aids human, robot cooperation with cross-training

Robots learn what humans need by doing their jobs on the manufacturing floor

With robots increasingly being used on manufacturing floors, researchers are looking for ways that humans can work better with their robot coworkers.

Scientists at MIT say the answer is cross-training -- humans and robots switching jobs to learn how they affect each other's work.

"People aren't robots. They don't do things the same way every single time," said Julie Shah, head of the Interactive Robotics Group at MIT, in a statement. "And so there is a mismatch between the way we program robots to perform tasks in exactly the same way each time and what we need them to do if they are going to work in concert with people."

Shah noted that a lot of robotics research focuses on making sure robots and humans can operate safely side by side. However, more work needs to be done to make robots smart and flexible enough to work effectively with people.

Often, the issue is that robots are programmed to do the same job the exact same way time after time. Humans, however, aren't so exacting.

So how do you enable robots to anticipate and react to these differences? According to Shah, at least part of the solution involves having the robots, and humans, learn what it's like to do the other's job.

To get started, she worked with MIT doctoral candidate Stefanos Nikolaidis to build an algorithm that would enable the robots to learn from their role swap. This step involved enabling the robots to gain information through demonstration.

By watching humans perform what are normally robotic jobs, the robots are able to learn how humans want the machines to perform their jobs.

MIT reported that Shah and Nikolaidis found that humans and robots worked together as a team 71% more after cross training. They also noted that the amount of time that people spend waiting for robots to finish a task dropped by 41%.

People also reported that the robots worked with them more efficiently after cross training.

"This is the first evidence that human-robot teamwork is improved when a human and robot train together by switching roles, in a manner similar to effective human team training practices," said Nikolaidis in a statement.

Human and robot collaboration has been a major focus of robotics research.

Last summer, for instance, scientists at Harvard University announced that they were working on an Iron Man-like smart suit that could improve soldiers' endurance in war zones.

The suit, which is expected to include sensors and its own energy source, is being designed to delay the onset of fatigue, enabling soldiers to travel farther in the field, while also supporting the body and protecting it from injuries when the soldier is carrying heavy loads.

In late 2011, Toyota Motor Corp. reported that it planned to release a group of robots that will act as health-care aids by 2013. The robots are being designed to lift patients and to help people suffering from paralysis to walk again.

NASA, which has used a robotic arm to carry astronauts in space and has a humanoid robot working on the International Space Station, said it will need humans and robots to work together to enable future, more industrious space missions.

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 hardwareEmerging Technologieshardware systems

More about GoogleHarvard UniversityInteractiveMITNASATopicToyota

Comments

Comments are now closed

Turnbull warns ACCC against undermining Telstra NBN agreement

READ THIS ARTICLE
MORE IN Business
DO NOT SHOW THIS BOX AGAIN [ x ]