自适应神经网络四旋翼无人机有限时间轨迹跟踪控制

针对带有模型不确定性和未知外部干扰的四旋翼无人机轨迹跟踪控制问题，提出一种基于径向基(radial basis function, RBF)神经网络的自适应全局快速终端滑模控制方法，确保系统对期望轨迹的有限时间跟踪。该方法考虑到全局快速终端滑模控制在实际应用中的适应性和抖振问题，利用RBF神经网络替代等效控制量，以神经网络的在线学习能力补偿系统内部的不确定性和未知的外部干扰，有效地降低了系统的抖振；根据Lyapunov方法导出的自适应律在线调整神经网络权值，以保证闭环系统的稳定性。通过一系列仿真算例和飞行实验验证了该方法的有效性与可行性，结果表明:该控制方法相对于滑模控制的抖振更小，具有更好的收敛性和抗干扰能力，同时对模型的参数摄动具有更强的鲁棒性。

Finite-time trajectory tracking control based on an adaptive neural network for a quadrotor UAV

Aimed at the trajectory tracking control problem of a quadrotor UAV with model uncertainties and unknown external disturbances, an adaptive global fast terminal sliding mode control method based on the RBF neural network has been proposed herein. The proposed method assists the system in tracking the desired trajectory in finite time. Considering the adaptability of global fast terminal sliding mode control in practical applications and chattering problems, the equivalent control quantity has been replaced by RBF neural networks. The chattering of the system has been reduced effectively by compensating for model uncertainty and unknown external disturbances with online learning of neural networks. According to the adaptive law derived from the Lyapunov method, the weights of neural networks are adjusted online to ensure the stability of the closed-loop system. Through a series of simulation examples and flight experiments, the effectiveness and feasibility of the proposed method have been validated. Results show that the proposed method has less chattering, better convergence, and anti-interference ability. It is also more robust toward model parameter perturbation compared to the sliding mode control.