Hybrid formation control scheme for multiple unmanned surface vehicles based on RGT-APF-RBF

Model uncertainties and disturbances challenge multi-USV formation control. This paper proposes a hybrid strategy integrating rigidity graph theory (RGT), artificial potential fields (APF), and a radial basis function neural network (RBFNN). Within a leader-follower framework, RGT maintains the desired formation topology, while APF handles real-time obstacle avoidance. The control design uses the known model for a baseline law; an RBFNN adaptively approximates unknown dynamics and external disturbances, and an adaptive compensator suppresses reconstruction errors and external disturbances. Lyapunov stability analysis proves global asymptotic stability and provides adaptive update laws for the neural network weights and control parameters. Comparative simulations demonstrate that the proposed RGT-APF-RBF controller surpasses conventional methods in tracking accuracy, formation keeping, and disturbance rejection. Physical experiments validate its real-world effectiveness and superiority.