四旋翼飞行器自适应神经网络轨迹跟踪控制器设计

Trajectory Tracking Control Design for a Quadrotor Based on Adaptive Neural Network

针对带有模型不确定性和未知外界干扰的四旋翼飞行器轨迹跟踪控制问题,提出一种自适应RBF神经网络控制策略。该方法利用RBF神经网络在线逼近和补偿系统中的未知非线性函数,减少对数学模型的依赖,提高抗干扰能力;结合Lyapunov方法导出在线调节神经网络权值的自适应律,增强鲁棒性;利用Lyapunov理论证明控制器的稳定性。通过仿真和试验验证该方法的有效性和工程应用价值,结果表明:在时变干扰和参数摄动作用下,所提方法相对于自抗扰控制的调节时间缩短1.1~2.1 s,轨迹跟踪的绝对误差平均值减小38.27%,具有更好的鲁棒性和抗干扰能力。

An adaptive radial basis function(RBF)neural network control strategy was proposed for the trajectory tracking con⁃trol of quadrotor aircraft with model uncertainty and unknown external disturbance.In this method,RBF neural network was used to on⁃line approximate and compensate the unknown nonlinear function in the system to reduce the dependence on the mathematical model and improve the anti-interference ability;combined with Lyapunov method,the adaptive law of on-line adjusting weights of neural net⁃work was derived to enhance its robustness;the stability of the controller was proved by using Lyapunov theory.The effectiveness and engineering application value of the method were verified through simulation and experiment.The results show that under the conditions of time-varying interference and parameter perturbation,the adjustment time of the proposed method is shortened by 1.1~2.1 s com⁃pared with ADRC control,and the average absolute error of trajectory tracking is reduced by 38.27%,the method has better robustness and anti-interference capability.