Decentralized Control of Large Collaborative Swarms Using Random Finite Set Theory

Controlling large swarms of robotic agents presents many challenges including but not limited to computational complexity due a number agents, uncertainty in the functionality each agent swarm, and swarm's configuration. The contribution this work is decentralize random finite set (RFS)-based control collaborative for controlling individual agents. RFS-based formulation assumes Gaussian mixture probability hypothesis density approximation complete topology centralized swarm control. To generalize localized or decentralized manner, sparse linear quadratic regulator (LQR) used sparsify gain matrix obtained using iterative LQR. This allows use information near other (localized topology) only agent's own (decentralized make decision. Sparsity performance are compared different degrees localization feedback gains which show that stability do degrade significantly providing swarms.