Stop, Think, and Roll: Online Gain Optimization for Resilient Multi-robot Topologies

TL;DR

This paper presents an online distributed optimization method to enhance the resilience of multi-robot network topologies, preventing single points of failure and maintaining connectivity amid dynamic changes.

Contribution

It introduces a novel online distributed control strategy that optimizes robot parameters to ensure resilient, failure-tolerant network topologies in multi-robot systems.

Findings

The method effectively maintains network connectivity during robot failures.

Simulation and real-robot experiments validate the approach's robustness.

The strategy adapts to dynamic network changes in real-time.

Abstract

Efficient networking of many-robot systems is considered one of the grand challenges of robotics. In this article, we address the problem of achieving resilient, dynamic interconnection topologies in multi-robot systems. In scenarios in which the overall network topology is constantly changing, we aim at avoiding the onset of single points of failure, particularly situations in which the failure of a single robot causes the loss of connectivity for the overall network. We propose a method based on the combination of multiple control objectives and we introduce an online distributed optimization strategy that computes the optimal choice of control parameters for each robot. This ensures that the connectivity of the multi-robot system is not only preserved but also made more resilient to failures, as the network topology evolves. We provide simulation results, as well as experiments with real robots to validate theoretical findings and demonstrate the portability to robotic hardware.