Tuesday, October 1, 2013
"A Soft, Amorphous Robotic Skin for Texture Detection, Identification and Localization"
PhD Candidate Dana Hughes, University of Colorado Boulder
ABSTRACT: As autonomous robots move towards working in more complicated environments, the need for soft, amorphous sensing skins becomes critical. Robotic skin is useful approach to compliance, non-visual navigation, gripping and object recognition. Dana will present a prototype skin which mimics the function of the Pacinian corpuscle in human skin for sensing, identifying and localizing textures. The skin implements a sparse network of sensing nodes colocating small microphone sensors with microcontrollers. This approach allows the skin to perform computations within the material, allows the skin to be reprogrammed as needed, and solves issues with scalability and computation and communication bottlenecks. Using this approach, the prototye skin is capable of classification of textures with 71% accuracy, and centimeter accuracy in localization.
Hosted by CS Graduate Social Committee
Open to all Computer Science graduate students