Sign In

Communications of the ACM

ACM TechNews

Teaching Bipedal Robots to Step Across Discrete Terrain

ATRIAS bipedal robot

There are no actuators at the ankles of the ATRIAS bipedal robot.

Credit: UC Berkeley

As scientists work to improve robots' ability to walk on challenging terrain, University of California-Berkeley (UC Berkeley) and Carnegie Mellon University laboratories have enabled a robot to walk dynamically on stepping stones, using recent advances in optimal and nonlinear control systems.

The Hybrid Robotics Group at UC Berkeley, formerly at Carnegie Mellon, believes this is the first time that dynamic walking on stepping stones with simultaneous variation in step length and step height has been successfully demonstrated on a bipedal robot. In experimental tests of their control algorithms on the ATRIAS bipedal robot platform, the researchers achieved dynamic walking over stochastically varying discrete terrain with step lengths between 30 and 65 centimeters and step heights of up to 22 centimeters. The researchers plan as part of the next phase to give the robots information about their surroundings by incorporating computer vision algorithms into the controllers.

The research has many potential applications, including exploration of unmapped areas and search and rescue missions. In addition, the methods could be used in robotic devices that augment humans, such as lower limb exoskeletons.

From IEEE Spectrum
View Full Article

Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA


No entries found