Home /Research /Kinematic multi-robot manipulation with no communication using force feedback
SWARM

Kinematic multi-robot manipulation with no communication using force feedback

Zijian Wang, Mac Schwager

Year
2016
Citations
115

Abstract

This paper proposes a novel decentralized algorithm that coordinates the forces of a group of robots during a cooperative manipulation task. The highlight of our approach is that no communication is needed between any two robots. Our underlying intuition is that every follower robot can measure the direction of the movement of the object and then applies its force along that direction to reinforce the movement. We prove that using our algorithm, all followers' forces will synchronize to the direction of the force applied by one leader robot, who guides the robotic fleet to its destination. We first verify our algorithm by simulation in a physics engine, where 20 robots transport a chair collectively. We then validate our algorithm in hardware experiments by building four low-cost robots, equipped with force and velocity sensors, to transport a cardboard box in a laboratory environment. In addition, our algorithm allows the leader to be a human, and we also demonstrate the human-swarm cooperation in our manipulation experiments.

Keywords

RobotKinematicsPhysics engineIntuitionComputer scienceSimulationRobot kinematicsSwarm roboticsControl engineeringTask (project management)

Related papers

Browse all SWARM papers