Vision-aided Leader-Follower Collaborative Mobile Manipulation with Control Barrier Functions

This paper proposes a method for autonomous cooperation among Unmanned Mobile Manipulators (UMMs) for load transportation tasks. In the proposed approach, the leader UMM utilizes force/torque measurements and carries fiducial markers arranged in a truncated rhombicuboctahedron configuration. The follower UMMs determine their pose relative to the leader using visual information while also employing their own force/torque sensors. The developed framework is fully distributed and operates without requiring any communication between the leader and followers. Safe operation is ensured through the integration of Control Barrier Functions (CBFs) in the UMM control loops, which maintain strictly bounded internal forces during collaboration. Extensive real-world experiments validate the proposed method, demonstrating its performance and generalizability for cooperative manipulation with aerial manipulators and aerial-ground mobile manipulator teams.