Title: Collaborative Robots in Action: How Friends Can Help Move Your Furniture
Abstract: Recognition, transport, and manipulation of objects is a fundamental robotics challenge. We apply machine learning techniques to autonomously identify objects, and then we use a coordinated path-planning approach to manipulate and transport objects using multiple robotic agents. In particular, we are interested in solving impedance-controlled manipulation with multiple robots that cannot be solved with one robot. We take a bio-inspired approach that builds on ant studies of collaborative manipulation, as well as building on the existing engineered robotic collaborative transport work. By combining engineered approaches, with well defined guarantees, and robust bio-inspired approaches we hope to develop a system which is reliable and predictable.