The paper presents a multi-head attention mechanism-based cooperative encirclement strategy for multiple unmanned aerial vehicles (UAVs) under electronic interference. Through the calculation of the virtual centroid of the multi-UAV formation using UAV coordinates, it reduces computational complexity and local communication demands. To address the issue of poor model generalization capability, the paper proposes that the model simultaneously attend to different parts of the input, thus better capturing the relationships and dependencies among UAV formations and improving the model’s robustness. Simulation and experiments demonstrate that the proposed multi-agent recurrent deterministic policy gradient (MARDPG) algorithm based on centroid and multi-head attention mechanism (CM-MADDPG) effectively solves the problem of multi-UAV formation cooperative capture in complex environments.
Multi-UAV Cooperative Encirclement Strategy in Complex Environments
Lect. Notes Electrical Eng.
International Conference on Guidance, Navigation and Control ; 2024 ; Changsha, China August 09, 2024 - August 11, 2024
2025-03-02
10 pages
Article/Chapter (Book)
Electronic Resource
English
Cooperative Multiple Target Encirclement via Platooning
AIAA | 2024
|IEEE | 2022
|Fixed-Time Control for Cooperative Target Encirclement and Tracking by Multi-UAV
Springer Verlag | 2024
|Escaping-target Encirclement and Navigation by Multi-agent Systems
British Library Online Contents | 2012
|