Topic: Versatile resilient robots with whole-body tactile sensing

In animals and humans, contacts with the environment are not avoided and sometimes even actively sought. We will deploy this inspiration from biology to design truly robust machines with distributed control systems, exploiting whole-body contacts with the environment. The main platforms will be robot manipulators or humanoids completely covered with sensitive electronic skin, learning their controllers in a biomimetic fashion (e.g., emulation of sense of pain, withdrawal reflex). The final scenario will be reaching and grasping objects exploiting whole-body contacts, with no or sparse visual inputs. We will demonstrate that reaching and grasping tasks can be accomplished even if conditions dramatically change, for instance when encoder values are manipulated with large offsets, a joint is completely blocked, or a camera observing the workspace is significantly distorted. 

Group and supervision: Research will be conducted within the Humanoid and Cognitive Robotics Group (https://cyber.felk.cvut.cz/research/groups-teams/humanoids/) under the supervision of Assoc. Prof. Matej Hoffmann. There are several state-of-the-art robot platforms, many with electronic skin, available at the group: iCub humanoid robot with whole-body skin, collaborative robot manipulators (UR10e with Airskin, KUKA LBR iiwa, Kinova Gen3). A motion capture system, professional impact measuring devices, and several robot hands and grippers (e.g. Barrett Hand) are also available. The group has extensive experience in physical HRI, biologically inspired robotics, and tactile sensing. 

Collaboration:   We will collaborate with Prof. Sami Haddadin, TU Munich, on physical human-robot interaction, and with Dr. Michael Zillich, Blue Danube Robotics / Airskin, Vienna, on collaborative robotics.