Swarm Intelligence Algorithms for Robust 3D Gas Source Localization Using UAV Swarms

Abstract

: The detection and localization of toxic gas sources in three-dimensional environments is a critical task for public safety during disaster scenarios and industrial accidents. Unmanned Aerial Vehicles (UAVs) equipped with gas sensors provide a safe and efficient means of monitoring contaminated areas. This paper presents a comparative study of several swarm intelligence algorithms for gas source localization (GSL) in a 3D environment using multiple UAVs. The algorithms under investigation include Adaptive Particle Swarm Optimization (APSO), Ant Colony Optimization (ACO), and Grey Wolf Optimizer (GWO). Their performance is evaluated in terms of success rate, convergence time, and mean localization error. To assess robustness under realistic conditions, the algorithms are tested under varying levels of measurement uncertainty. Simulation experiments are conducted over multiple independent runs to ensure statistical validity. The results highlight the relative strengths and weaknesses of each method. The findings provide practical guidance for selecting appropriate swarm intelligence techniques for UAV-based gas source localization in real-world scenarios.