Distributed Dynamic Clustering and Self-Organizing Cooperative Control for Heterogeneous UAV Swarms Multi-Target Tracking

In this paper, we propose a distributed dynamic clustering and self-organized cooperative control framework for multi-objective tracking of heterogeneous UAV swarms, which addresses the key challenges of centralized planning delays and insufficient coupling of motion constraints in distributed algorithms in dynamic environments. By integrating the self-organization principle, we construct a clustering optimization objective function under a hybrid architecture and innovatively design a distributed clustering protocol based on gradient descent to achieve globally optimal grouping. The dynamic potential field adjustment mechanism optimizes the intra-cluster trajectories while balancing the energy consumption of obstacle avoidance. The hierarchical optimization architecture decouples global decision-making from local control, and combines an improved fuzzy C-mean algorithm with a distributed auction mechanism to achieve collaborative clustering and task allocation with spatio-temporal constraints. Simulations involving 400 UAVs show satisfactory target tracking success rate, stable velocity distribution, and consistent obstacle avoidance distance, validating the robustness and real-time performance of the framework in complex dynamic scenarios. This work provides a new theoretical architecture and technical approach for distributed intelligent swarm systems.