LittleDog is a four-legged robot designed to learn how to negotiate and traverse unknown rugged and complex terrains. The 5-inch-tall quadruped, manufactured by Boston Dynamics, has been recently upgraded with an impressive set of improved locomotion skills thanks to researchers at the University of Southern California.
The robot uses 12 electric motors (3 for each leg) powered by lithium polymer batteries and a PC-level on-board computer for sensing, actuator control and communications. The sensors measure joint angles, motor currents, body orientation and foot-ground contact.
The latest controller software gives LittleDog greater stability, smoothness, and faster movement. The robot now moves more gracefully and learns where to position optimal footholds and can even pull off special moves for extreme terrain. The video below highlights all of these remarkable skills.
The research, developed at USC's Computational Learning & Motor Control Lab, is part of DARPA's Learning Locomotion Project, which aims to use machine learning techniques to create autonomous control software for a robot quadruped such that it can traverse unknown rugged and complex terrains.
Scientists at leading institutions including MIT, Stanford, Carnegie Mellon, USC, Univ. Pennsylvania and IHMC use LittleDog to probe the fundamental relationships among motor learning, dynamic control, perception of the environment, and rough-terrain locomotion.
The USC researchers say that most of the techniques they've developed in this work are generally applicable to planning and control problems on other systems and are now setting sights on humanoid robots.