Search this site
Embedded Files
Skip to main content
Skip to navigation
Multi-Agent Manipulation via Locomotion using Hierarchical Sim2Real
Home
D'Kitty Coordinate Videos
D'Kitty Push Videos
D'Kitty Avoid Videos
Low-level Videos
Failure Modes
Multi-Agent Manipulation via Locomotion using Hierarchical Sim2Real
Home
D'Kitty Coordinate Videos
D'Kitty Push Videos
D'Kitty Avoid Videos
Low-level Videos
Failure Modes
More
Home
D'Kitty Coordinate Videos
D'Kitty Push Videos
D'Kitty Avoid Videos
Low-level Videos
Failure Modes
D'Kitty Coordinate Videos
A hierarchical policy is trained in sim to coordinate two D'Kitties to move a box to a desired target location and orientation.
The learned behavior transfers to the real world in zero-shot fashion (target given by two white '+' signs on the floor).
Learned behavior generalizes to settings in which one end of the box is heavier than the other.
Google Sites
Report abuse
Page details
Page updated
Google Sites
Report abuse