Linda Spaa, van der, Jens Kober, Michael Gienger, "Disagreement-Aware Control for Online Learning Personal Solutions to Intention-Aware Physical Human-Robot Cooperation on Real Hardware", IEEE 2022 ICRA full day workshop: Shared Autonomy in Physical Human-Robot Interaction: Adaptability and Trust, 2022.

In this paper we present a control methodology that allows the robot to be controlled and learn from its observations in a high-level abstract state-action space, without the human having to pay special attention to the robot’s kinematics/ dynamics or limits. A small number of data efficient steps allow quick setup and easy transfer to different setups in similarly structured contexts. The controller is tested on a 7-DoF Franka Emika Panda robot arm.