The project –which incorporates machine learning, teleoperation, and human-robot collaboration-- demonstrates what happens when you combine AUBO Robotics' commitment to building symbiotic alliances among researchers and research institutions with robotics specialists
Robotics Advanced Research Alliance - Universities and Corporations
Peter Farkas | AUBO Robotics
Advanced Robotic Research Alliance between Universities and Corporations have been going on for years and now AUBO Robotics' AUBO-i5 collaborative robot arm is the technology behind exciting new research that's due to be presented at ICRA 2018 (21-25 May 2018; Brisbane, Australia). Held annually, ICRA is the premiere global event for robotics researchers.
The project –which incorporates machine learning, teleoperation, and human-robot collaboration-- demonstrates what happens when you combine AUBO Robotics' commitment to building symbiotic alliances among researchers and research institutions with robotics specialists exploring advanced ideas about human-robot collaboration.
Team lead, Dr. Cong Wang, an assistant professor in the New Jersey Institute of Technology, took some time recently to speak with AUBO Robotics about his research, why the AUBO-i5 was chosen for the project and the key factors that make the AUBO-i5 a “wonderful research robot.”
Researcher On A Mission
Dr. Cong Wang is a researcher on a mission to help bring about a state of "ubiquitous coexistence" between humans and robots.
Inspired in part by the many science fiction movies that show robots and humans casually sharing public and private spaces, Wang's experiments push the boundaries of human-robot interaction and test the limits of current robot capabilities.
In 2017, Wang's team built a bionic hand and motion capture glove designed to teach a robot by demonstration. They then combined this technology with a machine learning system that enabled the robot to learn much like a human would --through practice and self-evaluation.
The robot was able to learn how to flip a nunchuck --a neat trick that involves making a nunchuck spin around the back of your hand and catching it again-- in a matter of hours.
Most importantly, the system developed by Wang and his colleagues is non-task-specific. This means that robots could be trained to perform a wide variety of complex, dynamic motor skills that are beyond traditional robot systems.
Epic Quest To Find A Suitable Research Robot
In order to test their machine learning ideas and showcase some advanced and visually interesting collaborative robot handling skills, Wang's team needed a robot that was versatile, easy to program, and capable of both dynamic maneuvering and precise timing.
They also needed a robot that was affordable.
So, Wang set off on an epic quest to find a suitable robot to work with.
“This is a story I never tire of telling, because I became almost like a storm chaser,” says Wang. “But I was chasing robots. I was going to all the trade shows. Visiting all the industrial and research robot makers. None of them could satisfy my needs.”
The main issue was affordability.
Then Wang discovered the AUBO-i5.
“I realized that for the same money as some industrial and research robots cost, I can buy a dozen or more AUBO-i5s that are physically more capable, more reliable, and rely on more proven technology,” explains Wang.
Open Architecture Saves Time
Affordability confirmed, Wang was soon drawn to the AUBO-i5's other features, especially its open architecture control system, which simplifies integration with other devices --including bespoke robot hands.
“Integration between our robot hand and the AUBO-i5 was really easy,” says Wang. “It took us a year to build the hand, but we integrated it with the AUBO-i5 in a matter of minutes.”
Like many robotics researchers, the “last thing” Wang wants is to work with a “closed system” that offers a limited API or bridgeware.
“We don't want to spend a month trying to figure out how the portal works,” explains Wang. “With the i5, we don't need to worry about coding low level stuff and it takes minimum effort to integrate anything with the robot.'
“The combination of affordability, open architecture and the robot's physical capabilities make the AUBO-i5 “a wonderful research robot,” says Wang.
Looking Forward To ICRA 2018
The team's ground-breaking research will be presented at ICRA 2018, which takes place in Australia in May.
“I can't wait to go to ICRA and tell about people about the advantages of using something like AUBO robots. I want to tell people to stop using other robots. The AUBO-i5 is affordable and does it all,” says Wang.
Now in possession of two AUBO-i5s, Wang's lab is already planning its next research project --a dual arm experiment that goes beyond “pick and place” to demonstrate robots successfully handling different objects in a “ubiquitous and generic” manner.
AUBO Robotics congratulates Wang's team on being accepted to ICRA 2018. We're excited to see how their next project turns out!
Collaborate With AUBO Robotics
For researchers, the AUBO-i5 scores high on four critical factors for research success: affordability, open architecture, high-performance and speed of deployment.
AUBO Robotics looks forward to helping more universities, research centers and companies explore advanced research ideas.
So, whether you're putting together an ICRA-worthy research project about human-robot training or you wish to explore new safety strategies for collaborative manufacturing cells, make sure to talk to the team at AUBO Robotics.
The content & opinions in this article are the author’s and do not necessarily represent the views of RoboticsTomorrow
This post does not have any comments. Be the first to leave a comment below.
Post A Comment
You must be logged in before you can post a comment. Login now.