Decentralized Connectivity-Preserving Deployment of Large-Scale Robot Swarms

We present a decentralized and scalable approach for deployment of a robot swarm. Our approach tackles scenarios in which the swarm must reach multiple spatially distributed targets, and enforce the constraint that the robot network cannot be split. The basic idea behind our work is to construct a logical tree topology over the physical network formed by the robots. The logical tree acts as a backbone used by robots to enforce connectivity constraints. We study and compare two algorithms to form the logical tree: outwards and inwards. These algorithms differ in the order in which the robots join the tree: the outwards algorithm starts at the tree root and grows towards the targets, while the inwards algorithm proceeds in the opposite manner. Both algorithms perform periodic reconfiguration, to prevent suboptimal topologies from halting the growth of the tree. Our contributions are (i) The formulation of the two algorithms; (ii) A comparison of the algorithms in extensive physics-based simulations; (iii) A validation of our findings through real-robot experiments.

💡 Research Summary

This paper addresses a critical challenge in swarm robotics: deploying a large number of robots to multiple, spatially distributed target locations while ensuring the communication network of the entire swarm remains connected at all times. The proposed solution is innovative in that it aims to construct a dynamic communication backbone using only a subset of robots, freeing the rest for other tasks or as replacements.

The core technical approach involves constructing a logical tree topology over the physical network formed by robot-to-robot communication links. This tree acts as a backbone for data routing and provides a framework for robots to enforce connectivity constraints during movement. The paper’s primary contribution is the formulation and comparative analysis of two distinct, decentralized algorithms for building this tree: the Outwards algorithm and the Inwards algorithm.

The Outwards algorithm begins at a single, randomly chosen root robot and grows the tree outward by progressively incorporating nearby connector robots until it reaches the worker robots stationed at the target locations. It may create branches that do not lead to workers, which are periodically identified and pruned. Conversely, the Inwards algorithm starts from the worker robots at the targets and grows the tree inward toward the root. This method inherently creates a sparse tree consisting only of paths necessary to connect the workers to the root.

Both algorithms are implemented using the same high-level finite state machine (FSM), where individual robots can assume roles as Root, Worker, Connector, or Spare. The execution follows a cycle: after initialization, the swarm goes through states for starting a new tree, selecting a parent robot, growing the tree (which involves pruning for Outwards and spare robot insertion for Inwards), and selecting a new root. The root selection process employs a distributed algorithm to estimate the swarm’s centroid and elect the robot closest to it as the new root, thereby balancing the tree and minimizing its depth for efficient communication.

The authors conducted extensive physics-based simulations and real-robot experiments with over 20 robots to evaluate the algorithms. Key performance metrics included mission completion time, the number of robots engaged in the backbone, and algebraic connectivity (a measure of network robustness). Results generally favored the Inwards algorithm, which demonstrated faster deployment and formed backbones with fewer robots, as it grows goal-oriented paths from the targets backward. However, the performance of each algorithm can vary depending on the environment and target distribution. This work provides significant practical insights and a foundational framework for building decentralized, adaptive communication infrastructures using robot swarms, with potential applications in search-and-rescue, exploration, and environmental monitoring.