Cluster-Based Control of UAV Swarms for Cooperative Cargo Transportation Using Virtual Leaders

This paper addresses the problem of cooperative cargo transportation by a distributed system of unmanned aerial vehicles (UAVs) organized into clusters. A cluster-based control framework is proposed, combining consensus dynamics with a virtual leader approach to ensure coordinated motion of multiple UAV groups. Each cluster is modeled as a graph-based multi-agent system, where agents interact locally while the cluster centroid represents the transported payload. A distributed control law is developed, consisting of a consensus term and a tracking term driving the system toward a reference trajectory defined by a virtual leader. The stability of the proposed approach is analyzed using Lyapunov methods. It is shown that, in nominal conditions, the tracking error converges asymptotically to zero, while in the presence of bounded wind disturbances, the system exhibits practical stability with uniformly ultimately bounded error. A disturbance compensation mechanism based on a first-order observer is introduced to improve robustness. Numerical simulations with 30 UAVs demonstrate that the proposed method outperforms decentralized and leaderfollower strategies in terms of tracking accuracy, convergence, and control effort. The virtual leader approach achieves lower RMSE and cost functional values while maintaining robustness to disturbances and avoiding single points of failure.