Research on Adaptive RBF Neural Network Flight Control for Tilt- Rotor Quadrotor

Traditional multi-rotor drones are underactuated, with thrust and torque tightly coupled, making independent position and attitude control challenging. To enhance the flexibility and maneuverability of drones, this paper examines the tilt-rotor quadrotor, which offers greater agility compared to conventional quadrotors, enabling more precise control over the drone's position and attitude. We develop a mathematical model for the tilt-rotor quadrotor and utilize a control allocation matrix to linearize the nonlinear control allocation. An adaptive neural network disturbance compensation using a Global Fast Terminal Sliding Mode Control algorithm (GFTSMC-RBF) is designed for the dynamics model to estimate and compensate for system uncertainties and coupling effects. The feasibility of this controller is validated through MATLAB and CoppeliaSim simulations.