Event Location: GHC 6115

Abstract: Robots with legs have a number of advantages over those with wheels, such as greater ability to navigate through rough terrain, more flexible manipulation of objects, and fully holonomic control of the body. However, leveraging these advantages on a multi-legged system is computationally intensive and often relegates leg usage exclusively to locomotion. It is also important for mobile manipulators to function in human-friendly environments, which are dynamic, poorly modeled, and loosely structured, which imposes further constraints on how a robot interacts with the world.

The proposed work will pursue the use of legs for non-prehensile manipulation tasks, in coordination with locomotion and other behaviors. I will define a framework for describing such manipulation tasks and implement a set of primitives to demonstrate performance of these maneuvers.

Committee:David Touretzky, Chair

Matthew Mason

Anthony Stentz

Kevin Lynch, Northwestern University