Adaptive Dynamics with Efficient Contact Handling for Articulated Robots

Russell GayleMing C. Lin, Dinesh Manocha

{rgayle,lin,dm}@cs.unc.edu


Articulated body around a bridge

An articulated body with over 500 rigid links wrapping around a bridge model.
The simulation ran at interactive rates.
 



Abstract
We present a novel adaptive dynamics algorithm with efficient contact handling for articulated robots. Our algorithm automatically computes a fraction of the joints whose motion provides a good approximation to overall body dynamics. We extend Featherstone's Divide-and-Conquer algorithm and are able to efficiently handle all contacts and collisions with the obstacles in the environment. Overall, our approach provides a time-critical collision detection and resolution algorithm for highly articulated bodies and its complexity is sub-linear int he number of degrees-of-freedom. We demonstrate our algorithm on several complex articulated robots consisting of hundreds of joints.
Paper
Adaptive Dynamics with Efficient Contact Handling for Articulated Robots
Russell Gayle, Ming C. Lin, and Dinesh Manocha
Robotics: Science and Systems, 2006
Paper: pdf (360KB)
Poster: ppt (available upon request)



Additional Media

Threading an articulated body
An articulated body being threaded through a series of walls

Articulated pendulum
An articulated toy linkage falling onto several small cylinders
Related Work
Acknowledgements
This work was supported in part by a Department of Energy High-Performance Computer Science Fellowship administered by the Krell Institute, ARO, NSF, AMSO, DARPA, ONR/VIRTE, and Intel Corporation.

DOE-HPCSF ARO NSF ONR DARPA
GAMMA
UNC-CS GAMMA Group
Department of Computer Science
Campus Box 3175
UNC-Chapel Hill
Chapel Hill, NC 27599-3175