This Blind Bipedal Robot Navigates Uneven Terrain By Itself

FYI.

This story is over 5 years old.

Tech

This Blind Bipedal Robot Navigates Uneven Terrain By Itself

Four legs good, two legs better.

It's a challenge getting bipedal robots to walk on their own. But now a group of electrical engineers and computer scientists have created a robot that walks by itself while keeping upright in snowy, uneven, and slippery conditions. And here's the kicker: It can't see.

The group's robot, dubbed MARLO, is a top-heavy box on stilts with no vision that walks in much the same way as a toddler does. But still, these are pretty exciting times as most previous attempts to get bipedal robots to walk unaided have been desperately surreal, or resulted in snail-pace locomotion.

Advertisement

"The robot has no feeling in her tiny feet, but she senses the angles of her joints—for instance, her knee angles, hip angles and the rotation angle of her torso," Jessy Grizzle, professor of electrical engineering and computer science and of mechanical engineering from the University of Michigan, said in a press release.

To move MARLO, the researchers used mathematical modelling to create 15 walking styles that could handle different speeds and types of terrain. They applied these to MARLO, giving commands from an Xbox controller on how quickly and in what direction the robot should walk. However, it's up to MARLO, which has onboard feedback algorithms, to process its environment and adapt its gait accordingly.

For the moment, MARLO can move smoothly on plywood squares and bumpy astroturf as long as it doesn't have to make any sudden turns. Improvements in the algorithms feedback ability are required for greater agility—something the group are currently working on.

What makes this research interesting is the future potential to apply the algorithms developed for MARLO to other robots, and to people with prosthetic legs.

Robert Gregg, an assistant professor of mechanical engineering and bioengineering at the University of Texas-Dallas, for example, has already gotten started: He has adapted the research group's algorithm to control a prosthetic lower leg.

"We hope to encode similar abilities into our robotic prosthetic leg so that lower-limb amputees can just as easily walk about the community without having to think about the terrain," said Gregg, in a press release.