Algorithm-Driven Multi-User Platform for Decentralized Coordination in Self-Organizing UAV Swarms

Abstract

This paper introduces a comprehensive software platform designed to coordinate self-organizing UAV swarms through a secure and modular client-server system. Developed with multi-user collaboration in mind, the platform features an intuitive, cross-platform interface that allows users to define mission tasks, construct navigation graphs, and monitor swarm activity in real time.

At the heart of the system is a robust path-planning algorithm based on the rotor-router model with loop reversibility, which enables reliable and evenly distributed task coverage without relying on randomness. To enhance fault tolerance and ensure resilience in communication-limited environments, the platform employs a gossip-based broadcast algorithm. This allows swarm members to share information efficiently and maintain coordinated behaviour, even when some nodes experience failures or connectivity issues.

A built-in simulation module enables users to test and refine swarm coordination strategies before deployment, reducing operational risk and improving mission reliability. By simulating various environmental conditions and mission scenarios, users can evaluate system behaviour and optimize task execution. In parallel, the platform supports real-time 3D panorama generation from UAV-captured images, providing rich visual context and enabling more effective post-mission analysis.

Taken together, these features form a scalable, secure, and highly flexible system for managing decentralized drone swarms. The platform is well-suited for applications that demand coordination across multiple agents, including environmental monitoring, search and rescue, infrastructure inspection, and autonomous exploration. It bridges theoretical rigor with practical usability, offering a reliable toolset for both researchers and mission operators. Our work builds on earlier systems, introducing hybrid rotor-router initialization, algorithmic no-fly zone enforcement, and dual-toolchain image stitching.