Three-Dimensional Formation Control for Robot Swarms

Abstract

This article addresses a distributed 3D algorithm for coordinating a swarm of autonomous robots (ARs) to spatially self-aggregate into an arbitrary shape based on only local interactions. Each AR has local proximity sensors for measuring the relative coordinates of its nearby AR. We assume that only a single AR, called the leader, can localize itself in global coordinate systems, while accessing the arbitrary user-specified 3D shape. We further assume that the leader has a communication ability superior to all other ARs so that the leader can directly send a communication signal to any other AR inside the shape. Our aim is to make the ARs maneuver while maintaining a user-specified 3D formation such that the network connection of all ARs is maintained during the maneuver. Our approach results in a 3D formation shape and does not need the global localization of an AR, except for the leader. To the best of our knowledge, our study is novel in the construction of a 3D formation for covering an arbitrary shape such that network connection is maintained while ARs maneuver based on local communication. We further show that the proposed control yields reliable accuracy against significant AR failures and movement error. Utilizing MATLAB simulations, we demonstrate the outperformance of the proposed formation controls.

Keywords

• distributed network systems
• 3D formation control
• mobile sensor network
• network connection
• user-specified 3D shape