Trajectory Tracking of an Aerial Manipulator based on Adaptive RBF Neural Network

Due to the underactuated characteristics of UAVs and the coupling effects caused by rigid connections with manipulators, robustness and high precision are crucial for UAV controllers. This paper simplifies the dynamic model of a rotor UAV system and designs a Global Fast Terminal Sliding Mode (GFTSM) controller to achieve accurate trajectory tracking under disturbance conditions. To further enhance disturbance rejection performance, an RBF neural network is incorporated into the controller to estimate lumped disturbances (including internal coupling and external disturbances), enabling active compensation and high-precision tracking. Meanwhile, the stability conditions of both the controller and the neural network are derived using Lyapunov theory. Finally, a set of evaluation metrics is established, and comparative simulations with other controllers are conducted. The results demonstrate that the proposed controller significantly improves the robustness and accuracy of the rotor UAV system while ensuring good convergence.