Predefined-time adaptive neural tracking control for quadrotor unmanned aerial vehicles

This paper investigates a predefined-time trajectory tracking control problem for quadrotor unmanned aerial vehicles (QUAVs) subject to system uncertainties and external disturbances. A predefined-time adaptive neural control scheme is developed by integrating backstepping design, neural network approximation, and predefined-time stability theory. First, a cascaded control framework is constructed, where the outer-loop position controller generates the desired attitude commands, and the inner-loop attitude controller ensures fast and accurate tracking. Radial basis function neural networks are employed to approximate unknown nonlinear dynamics, and adaptive laws are designed to update the neural parameters online. By introducing a novel predefined-time Lyapunov-based design, all closed-loop signals are guaranteed to be bounded, and the tracking errors are proven to converge to a small neighborhood of the origin within a prescribed time independent of initial conditions. Numerical simulations and real-time experiments on a QUAV platform, including comparative experiments, are conducted to validate the effectiveness of the proposed method. The results demonstrate that the proposed scheme achieves superior tracking accuracy, faster convergence, and stronger robustness compared with conventional finite-time and backstepping control methods.