摘要
针对无人机动态逆控制器存在的明显缺陷,采用在线神经网络对逆误差进行补偿,设计了神经网络动态逆控制器。构造了一种单隐层神经网络结构,证明了整个闭环系统的稳定性,给出了稳定的在线神经网络权值调整算法。通过仿真验证了该方案的鲁棒性、对故障的适应性以及在作动器纯延时下的响应性能,证明了在线神经网络对动态逆控制器具有明显的改善作用。对于无人机先进飞行控制系统的设计研究具有重要的参考价值。
Aiming at evident shortage of dynamic inversion controller of UAV, dynamic inversion controller based on NN is designed through online NN compensating inversion error. A single hidden layer NN struc-ture is constructed and the stability of the whole closed loop system is proved. Also the stable adjustment arithmetic of online NN weight is proposed. The robustness, the adaptability to fault and the response capa- bility to actor delay time of the scheme are verified by simulation. It is also proved that the online NN has improved the performance of dynamic inversion controller well. It has important reference value for designing advanced flight control systems of UAV.
出处
《航天控制》
CSCD
北大核心
2009年第5期3-6,共4页
Aerospace Control
关键词
无人机
动态逆
神经网络
稳定性
Unmanned aerial vehicle （UAV）
Dynamic inversion
Neural network （NN）
Stability
