Dark-Room Exchange: Human Supervision of Decentralized Multi-Robot Systems Using Distributed Ledgers and Network Mapping

This paper develops a distributed technique to populate the network graph of a decentralized multi-robot system (MRS) by employing a consensus protocol for extracting the identities and states of each robot's neighbors in the MRS. A dark-room exchange (DRE) technique is proposed wherein each robot uses its on-board 2D LiDAR for range sensing and peer-to-peer communication to identify and track neighboring objects. The resulting information is utilized to build and maintain a distributed ledger populated with the information of the MRS network graph structure to facilitate supervision by human operators. The system is tested in a simulated environment consisting of TurtleBot3 robots scattered in a 2D plane. Using the results of simulation, an analysis of the speed and performance of the DRE technique is conducted that illustrates high reliability and fast response times. The paper concludes with a discussion of the future scope of this research for multi-robot/swarm applications.