Adaptive Neural Network Based Control Strategy for Tiltrotor UAV: Transition Phase Longitudinal Control

Tilt-rotor Unmanned Aerial Vehicle have gained significant attention for their ability to rapidly respond to complex environments and execute long-range missions. The control challenges during the transition flight phase are particularly complex, involving highly nonlinear dynamics and intricate coupling between propulsion systems and aerodynamic forces. This study focuses on addressing the longitudinal control problem of TUAVs during their transition from rotor to fixed-wing mode. A novel control strategy is proposed, integrating a comprehensive understanding of vehicle dynamics with an advanced controller design. Specifically, a state feedback controller is combined with a Radial Basis Function Neural Network to enable real-time adaptive estimation of lift forces, thereby enhancing altitude tracking precision. Simulation results demonstrate that the proposed controller significantly improves altitude tracking performance during the transition phase, achieving an 89% reduction in tracking error compared to the traditional PID controller.