Multi-UAV Uniform Sweep Coverage in Unknown Environments: A Self-organizing Nervous System (SoNS)-Based Random Exploration

TL;DR

This work tackles uniform sweep coverage by a homogeneous UAV swarm in unknown GNSS-denied environments using a Self-Organizing Nervous System (SoNS) framework to form a hierarchical ad-hoc network. It evaluates SoNS-based random walk (SoNS-RW) and a Boustrophedon sweep (SoNS-BS) against several decentralized baselines, demonstrating that SoNS-RW achieves faster completion and better coverage uniformity than non-SoNS approaches, while SoNS-BS provides an idealized upper bound under additional assumptions. The study highlights the strength of SoNS for scalable, fault-tolerant, self-organizing coordination in GPS-denied settings, with practical implications for robust field sampling and environment monitoring. It also discusses the speed-uniformity trade-offs among baselines and suggests future work on adaptive formation shifts and density-control strategies to further enhance performance in realistic conditions.

Abstract

This paper addresses multi-UAV uniform sweep coverage in an unknown convex environment, where a homogeneous UAV swarm must evenly visit every portion of the environment for a sampling task without access to their position and orientation. Random walk exploration is practical in this scenario because it requires no localization and is easy to implement on swarms. We demonstrate that the Self-Organizing Nervous System (SoNS) framework, which enables a robot swarm to self-organize into a hierarchical ad-hoc communication network using local communication, is a promising control approach for random exploration in such environments. To this end, we propose a SoNS-based random walk method in which UAVs self-organize into a line formation and then perform a random walk to cover the environment while maintaining that formation. We evaluate our approach in simulations against several decentralized random walk strategies. Results show that our SoNS-based random walk achieves full coverage faster and with greater coverage uniformity than these benchmark strategies, both globally and in local regions.

Figures (3)

• Figure 1: UAV positions in the formation and network topology for the SoNS-based approaches (left), and SoNS-BS sweep strategy (right).
• Figure 2: Mean coverage completeness of each approach over time.
• Figure 3: Coverage performance results. LDR-R and LDR-RP represent LDR-Random and LDR-Repulsive, respectively.