PhD Presentation: Consensus-Based Distributed Connectivity Control in Multi-Robot Systems
In the ever-growing field of robotics, increasing attention is recently being given to multi-robot systems. As technology advances, robots are becoming smaller, cheaper, and more diverse in their applications. Therefore, there is a rise in research in the field of self-organizing networked multi-agent systems, primarily used in large-area sensor networks, collective behavior analysis, and joint formation control. Especially in underwater applications, maintaining a robust and efficient communication network is essential for retaining the desired functionality even under external disturbances and agent failures. We address a distributed connectivity control problem in networked multi-robot systems using a trust-based consensus protocol for the estimation of algebraic connectivity, and a probabilistic model for topology adaptation. The described method does not require the initial communication graph to be connected and is independent of the robot motion control. In this talk, we will present the details of the proposed connectivity maintenance strategy applied to a simulated system and to a group of aMussles – sensor hubs used to monitor the environmental properties of the sea. We will also discuss the preliminary results that show how the system achieves resilience through coordination in case of the varying initial topology of the communication graph, different sizes of a multi-robot system, different deployment scenarios, and robot failure.