Thursday, May 13, 2021
Home » Leading Edge Robotics News » Borrowing Human Reflexes for a Bot

Borrowing Human Reflexes for a Bot

MIT PhD student Joao Ramos demonstrates a Balance Feedback Interface system on HERMES.
MIT PhD student Joao Ramos demonstrates the Balance Feedback Interface system on HERMES.

Researchers at MIT have developed a robotic system that uses a human operator’s innate balancing ability to stop the robot from falling over after an impact. An exoskeleton worn by the human calculates the person’s movements as forces are unexpectedly applied to the robot and human simultaneously. The robot is programmed to mimic the human’s movements as he balances after each impact. Joao says the system works faster than traditional balancing systems, which rely on visual cues.

“The processing of images is typically very slow, so a robot has difficulty reacting in time,” says Joao Ramos, a PhD student at MIT’s Department of Mechanical Engineering. “Instead, we’d like to use the human’s natural reflexes and coordination. An example is walking, which is just a process of falling and catching yourself. That’s something that feels effortless to us, but it’s challenging to program into a robot to do it both dynamically and efficiently. We want to explore how humans can take over complex actions for the robot.”

The robot, called HERMES, was funded in part by the US Defense Advanced Research Projects Agency. It’s intended to one day be deployed to dangerous disaster sites, where it could be safely guided by a remote human operator.

“We’d eventually have someone wearing a full-body suit and goggles, so he can feel and see everything the robot does, and vice versa,” Ramos says. “We plan to have the robot walk as a quadruped, then stand up on two feet to do difficult manipulation tasks such as open a door or clear an obstacle.”

In one experiment, the robot’s torso was hit with a hammer. Ramos stood on a special platform that applied forces to mimic the strikes to the bot. As he reflexively shifted his weight to balance, so did the bot. In other tests, Ramos mimed actions, such as punching through drywall. As Ramos moved, so did the robot. Mimicking punching through drywall, for example, caused the platform to push forward on Ramos to simulate the force of the drywall pushing back on the robot. Ramos rocked back on his heels to balance, and the bot followed suit.

“We don’t quite understand how humans master such gracious dynamic movements,” Kim says. “This research might provide us an opportunity to learn about how we balance during such complex manipulation tasks.”

A paper on the interface will be presented at the IEEE/RSJ International Conference on Intelligent Robots and Systems in September.