a team of robots in the land of terps

in robots in education by Rethink Robotics

A common stat noted when exploring schools to enroll in is the student-teacher ratio. In the future it could be equally valuable to look at the researcher to robot ratio, as more and more academic institutions bring in robots for research, as well as teaching engineering and manufacturing skills.

At the University of Maryland that’s happening today where there are currently four Baxter collaborative robots on the College Park campus. Here’s a sample of what they’re up to at UMD.

Autonomy, Robotics and Cognition (ARC) Lab

At the ARC Lab, one of the research teams is working with three Baxter robots, which they’ve renamed as Julia, Alice and Bob. Just as one Baxter looks like the next one, imagine how hard it is for humans to tell the difference between identical triplets. Or better yet, imagine a machine trying to recognize and successfully identify people.

One of the team’s projects involved working with their robots to locate and identify. By using the robot’s embedded camera, along with a speech system, Alice is able to point out Julia.

“My long range aim is to try to achieve human level artificial intelligence. So the Baxter would be like a person, maybe not a full-fledged adult. It can do the things that many people can do. The robot would know a lot about itself and would get better over time.” –Don Perlis, Professor, Department of Computer Science

Institute for Advanced Computer Studies (UMIACS)

One of the robots at UMIACS observes a human mixing a drink and tries to make the same mixture, with the same quantities. Reaching over or around obstacles and selecting the right object to pick up is an easy task for a human, but challenging for a robot. Baxter solves for that through planning in order to successfully choose, grasp and pour the correct bottles.

My goal is to build some kind of a software ecosystem that I can autonomously watch humans in a specific task and then have the robot do it. And create models that are really like plug and play. So this is facilitated by Baxter.” –Cornelia Fermüller, UMIACS Associate Research Scientist

Computer Vision Lab

My research focuses around the problem of active perception, which is basically the perception of biological systems. And I am looking for bio-inspired algorithms to transfer such capabilities into artificial systems, such as robots.”

“With other robots that I recall from the past, there were periods of months that people had to study the manuals and be able to produce programs in very obscure programming languages to get the robot to move one inch. Having Baxter learn how to perform actions by observing humans doing those actions, or by observing humans in videos, doing those actions I believe is going to revolutionize our society.

“Industry is in need of solutions for low volume and high mix robotics in factories, cannot solve these problems. A very interesting way of solving this problem is developing very flexible and sophisticated Baxters…and humans will have more time for more meaningful activities.” –Yiannis Aloimonos, Director of UMD’s Computer Vision Lab

The Baxter cobots are certainly a busy bunch in Terp territory, with groundbreaking progress also being made at UMD’s Neuromotor Control and Learning Laboratory, as well as at the school’s Advanced Manufacturing Lab.

Next, take a look at the amazing work others in research and education are focusing on with their robots in the Rethink Robotics video gallery. Or, if you’d like to configure your own Baxter, head on over to our Build a Bot calculator. And for the latest on the collaborative robot space, visit Cobot Central.


(1 comment)

About the Author

1 comment on this article

androidrobo May 22, 2017 at 12:04 pm

One of the team’s projects involved working with their robots to locate and identify. By using the robot’s embedded camera, along with a speech system, Alice is able to point out Julia.

Leave a Comment

Your email address will not be published. Required fields are marked *