高超声速飞行器神经网络动态逆姿态控制器设计

Neural Network Dynamic Inversion Design for Hypersonic Attitude Control

以一类通用高超声速飞行器纵向模型为研究对象,推导了飞行器纵向姿态运动方程。考虑受到飞行器附加攻角扰动以及气动参数不确定性的影响,对运动方程中速率变化快慢不同的攻角和俯仰速率分别引入精确线性化动态逆反馈,并利用神经网络对系统逆反馈误差进行补偿,有效地抑制了模型参数摄动,实现了对攻角指令的精确跟踪。将设计的控制器对高超声速非线性模型进行系统闭环仿真,仿真结果说明所设计的姿态控制器具有抗参数摄动的性能,能够满足高超声速飞行器复杂飞行条件下的姿态控制要求,具有较强的鲁棒性。

This paper takes a nonlinear longitudinal model of generic hypersonic vehicles as object,and the attitude kinematics equation is deduced.For the excess disturbance of angle of attack and parameters perturbation,the precise linearization using the dynamic inversion to equation of angle of attack and pitch rate is introduced,and the dynamic inversion error is compensated using neural network,the parameters perturbation is restrained,the angle command of attack is traced availably.The close loop simulation between control system and nonlinear hypersonic vehicles model is presented,the result shows the designed attitude controller has counter-disturbance capability,which is satisfied with the need of hypersonic vehicles complex flight condition,and it has good robustness.