When manipulating an object with multiple effectors such as in multidigit grasping or multiagent collaboration, forces and torques (i.e., wrench) applied to the object at different contact points generally do not fully contribute to the resultant object wrench, but partly compensate each other. The current literature, however, lacks a physically plausible decomposition of the applied wrench into its manipulation and internal components. We formulate the wrench decomposition as a convex optimization problem, minimizing the Euclidean norms of manipulation forces and torques. Physical plausibility in the optimization solution is ensured by constraining the internal and manipulation wrench by the applied wrench. We analyze specific cases of three-fingered grasping and 2-D beam manipulation, and show the applicability of our method to general object manipulation with multiple effectors. The wrench decomposition method is then extended to quantification of measures that are important in evaluating physical human-human and human-robot interaction tasks. We validate our approach via comparison to the state of the art in simulation and via application to a human-human object transport study.

Cooperative dynamic object manipulation increases the manipulation repertoire of multiagent teams. As a first step toward cooperative dynamic object manipulation, we present an energy-based controller for cooperative swinging of two-agent pendulum-like objects. Projection of the complex underactuated mechanism onto an abstract cart-pendulum allows us to separate desired and undesired oscillations. The desired oscillation is excited up to a desired energy level, while an undesired oscillation can be actively damped. Communication between the agents is restricted to force feedback. The controller can render leader and follower agents. The follower actively assists the swinging task by imitating the leader's energy flow. Real-world experiments with a robot and a human swinging complex pendulum-like objects are presented. The experimental results indicate that no simultaneous damping of the undesired oscillation is needed, also because it disturbs the human partner. A successful contribution of the robotic follower to the swing-up effort in interaction with a human leader supports the proposed control approach.

In this paper we propose to track online computed force trajectories to control the energy of various types of pendulum-like objects. The considered pendulum-like objects can be controlled by multiple agents and swing in different oscillation degrees of freedom. Our goal is to excite one specific oscillation, the intended oscillation, while damping all other disturbance oscillations. By approximating the intended oscillation as a simple pendulum oscillation, we can specify a desired force trajectory. Tracking of this force trajectory results in a controlled swing-up of the intended oscillation accompanied by a simultaneous damping of the disturbance oscillations. Simulation experiments with a two-agent trapezoidal pendulum show convincing control performance. A human-robot virtual reality experiment shows the transferability of the control approach to a human interaction partner. The limitations of the approach are discussed based on simulation results obtained for a single-agent double pendulum.

Cooperative dynamic object manipulation can extend the manipulation capabilities of robot-robot and human-robot teams. In order to be able to inject energy into various suspended objects of unknown parameters, in this paper we propose an adaptive controller which combines reinforcement learning with energy based swing-up control. The proposed controller is successfully verified in a single robot and human-robot experimental setup for different types of suspended objects.

Cooperative dynamic object manipulation extends the manipulation capabilities of human-robot dyads. This paper investigates cooperative swinging of rigid objects with the goal of reaching a desired level of energy, i.e. a desired object height. A human-human pilot study indicates that the arm-object-arm system can be approximated by a simple pendulum with two-sided unidirectional pulsed torque actuation. Based on the results of the human-human experiments, a robotic leader and follower controller are synthesized. Multi-body simulations based on human-like parameters successfully replicate the characteristics observed in the human-human experiments.

This paper investigates goal-directed cooperative object swinging as a novel physical human-robot interaction scenario. We develop an energy-based control concept, which enables a robot to cooperate with a human in a goal-directed swing-up task. The robot can be assigned to be a leader or an actively contributing follower. We conduct a virtual reality experiment to compare effort sharing and performance of mixed human-human and human-robot dyads. The leader and the follower controllers yield similar results compared to their human standard.

This paper investigates human-robot cooperative swinging of a complex pendulum-like object. The complexity of the object results in two possible swinging modes. The goal is to excite one mode such that a desired energy level of the pendulum is reached while simultaneously damping the other mode. The energy based control concept relies on the projection of the complex mechanism onto an abstract simple pendulum with two-sided actuation. An actively contributing robot leader and robot follower are implemented. The controller performance is analyzed through simulations. A virtual reality experiment shows the transferability of the control approach to a human interaction partner.