Engineers in the Intelligent Robotics and Autonomous Systems Lab at the University of Cincinnati have developed an artificial intelligence (AI) to train robots to collaboratively move big objects. The team demonstrated the robots’ ability to move a long rod around obstacles and a narrow door in computer simulations.
According to Andrew Barth, a doctoral student in UC’s College of Engineering and Applied Science, the task of moving such objects is a perfect fit for robots.
“We made it a little more difficult on ourselves. We want to accomplish the task with as little communication as possible among the robots,” Barth said.
Barth is lead author of the study published in the journal Intelligent Service Robotics. The co-authors of the research included Professor Ma, UC doctoral student Yufeng Sun and UC senior research associate Lin Zhang.
Testing the System
In the tests, neither of the robots directed the other, and they didn’t share the strategy before carrying out the task. The robots relied on genetic fuzzy logic, which is a type of AI that is an intelligent control technique where the system mimics human reasoning. It does this by replacing a simple binary classification with degrees of right and wrong, and it modifies individual solutions to “learn” from past results.
“Ultimately, we want to expand this to 10 or more robots working cooperatively on a project,” Barth said. “If you want to build a gigantic habitat in space, say, you’ll need a lot of robots working together. But if you were relying on a communications network and it goes down, then your whole project is done.”
According to Barth, independent robots mean the loss of one will not result in a task failure since the others can compensate.
The task given to the robots involved carrying the virtual couch around two obstacles and through a narrow door, and they completed it 95% of the time in simulations. The robot work partners also demonstrated a 93% success rate in a new scenario where there were two unfamiliar objects and a target door in a different location. The robots demonstrated a near equal accuracy rate and didn’t require retraining. Even more impressively, they demonstrated the same when various factors such as the size of the object were altered.
“If you can train robots to work semi-independently with as little information as possible, then you made your system more robust to that failure and made it easier for large groups to collaborate,” Barth said.
“Our long-term goal is for multiple robots to be able to cooperate to perform difficult tasks — like moving furniture,” Ma said.
Creating New Opportunities
According to Ma, collaborative robots could create big new opportunities in various fields, and new safety features could improve robotic safety.
“There are a host of applications. Any place you have jobs that multiple people are doing in the future, you could have multiple robots doing,” Ma said. “Currently, most robots work alone. But in the future we’ll need multiple robots working together just like people do now.”
The team is developing a control system that is scalable, which enables multiple robots to complete a task.
“And you don’t need to retrain them if suddenly it’s just four or six,” he said. “If one or two fail, the rest can carry on. That’s the key.”