GA-optimized consensus control and formation transition for multi-UAV systems

Multi-unmanned aerial vehicle (Multi-UAV) formations require flexible formation-transition capabilities during complex mission execution so that they can adapt to variable environments and changing mission requirements. Under a point-mass kinematic model, this paper proposes a formation-structure model based on a virtual leader and consensus control, achieving goal-point guidance, obstacle avoidance, and inter-UAV coordination through a multi-level behavioral control framework. To address the limitations of traditional formation control methods, such as reliance on empirically tuned parameters and low convergence efficiency, the proposed framework integrates virtual-leader-based consensus coordination, multi-behavior task decomposition, and genetic algorithm (GA) joint parameter optimization. A multi-objective fitness function is constructed to jointly optimize consensus convergence performance, multi-behavior coordination, and formation-transition efficiency. On this basis, an adaptive transition mechanism between V-formation and linear formation is designed, enabling the formation to dynamically reconfigure based on the distribution of environmental obstacles. Simulation results show that, in sparse obstacle environments, the GA-optimized algorithm improves consensus convergence efficiency by 30.8% and reduces transition time by 22.7%. In continuous-constraint environments, consensus convergence efficiency is improved by 25.6%, while transition time is reduced by 29.6%. These results confirm that the proposed method improves both convergence speed and formation-transition efficiency for multi-UAV formations in complex environments.