Cooperative Interaction and Manipulation in Human-Robot and Multi-Robot Systems

In the context of physical human–robot interaction, starting from the LuGre friction model, an elastoplastic admittance model is proposed that enables both elastic, reversible motion modifications and persistent offsets of the reference, allowing the operator to intuitively adapt the robot trajectory. The model is initially integrated into a variable-admittance control scheme compatible with non-redundant manipulators. To address collaborative transportation with mobile manipulators, an alternative formulation of the elastoplastic model is presented, along with a whole-body control algorithm based on lexicographic optimization. The controller combines elastoplastic behavior with kinematic constraints and distributes the motion command between the manipulator and the mobile base, prioritizing the manipulator for elastic components and the base for plastic offsets. A trajectory-recovery mechanism, obtained by optimizing an intermediate reference within the lexicographic problem, enables controlled re-entry after a plastic adaptation. The approach is validated in a collaborative transportation task with a human operator. Finally, in the multi-robot domain, a leader–follower control scheme is described in which the leader employs an expanded costmap to represent the environment from the formation perspective. The construction process ensures that the leader planned path is collision-free for the entire formation. A decentralized controller allows followers to maintain formation while enforcing a feasibility constraint that prevents the leader's velocity from exceeding the followers capabilities. The scheme is validated in simulation and on a real system through a collaborative transportation task involving a rigid object.