Grope and Hope Matthew T. Mason Dagstuhl Workshop on Multimodal Manipulation Under Uncertainty October 2015
Thanks!
collaborators and sponsors Robbie Paolini Alberto Rodriguez with thanks to Nikhil Chavan-Dafle Jeff Trinkle for the title Annie Holladay Weiwei Wan Laura Herlant National Science Foundation Siddhartha Srinivasa DARPA Mike Erdmann Army Research Office Marc Raibert ABB Tomas Lozano-Perez Zhiwei Zhang Erol Sahin
Human hand General Human-operated simple hands Toolomorphism Anthropomorphic robot hands Anthropomorphism Factory Specialized grippers Simple Complex
An inspiration: the pickup tool
Pickup tool philosophy • Let the fingers fall where they may • Instead of “put the fingers in the right place”. • Grasp first, ask questions later • Instead of knowing pose in advance, and avoiding object motion during grasp. • Grope and hope
Simplicity, Generality, Robotics Dagstuhl 2009
Most significant grasp types.
37.7 31.4 25.2 18.9 12.6 6.4 0.1 12.6 37.7 25.2 25.2 • SVM 37.7 12.6 50.2 • Gaussian kernel • Minimize misclassifications and maximize margin for correctly classified examples.
Singulation SVM, Gaussian kernel, leave half out cross validation, mean accuracy: 92.9% Orientation K nearest neighbors, leave one out cross-validation, mean error: 8.4°
The MLab Hand’s approach to haptic perception • Skin receptors? No • Intrinsic tactile sensing? [Bicchi & Dario 1987] A generalization
Hyperintrinsic tactile sensing • Exclude unstable poses • Use proprioceptors to get the last few bits • Is wholly intrinsic tactile sensing possible? Determine pose with zero sensory bits?
Recommend
More recommend