中文摘要：

研究一种新的近空间飞行器鲁棒自适应飞行控制系统设计方案。利用单隐层神经网络的函数逼近能力和被控对象分析模型的有用信息构建一种单隐层神经网络干扰观测器，用以在线估计系统中存在的不确定性，具有自适应调节能力的鲁棒控制器用以克服估计误差。将所得单隐层神经网络干扰观测器与轨迹线性化方法结合形成新的非线性系统鲁棒自适应控制方案，严格的理论证明在给定的自适应调节律作用下闭环系统所有误差最终有界。该控制方案被用于近空间飞行器飞行控制系统设计，高超声速飞行条件下的仿真结果表明该方法不仅有效，而且能够提供高精度、高稳定度的控制性能。

英文摘要：

A robust adaptive flight control scheme is presented for a near - space vehicle （ NSV）. A novel estimator design technique called the single hidden layer neural network disturbance observer （SDO） is developed. The SDO utilizes both the universal approximation property of neural network, and more useful information on the controlled system. A new robust adaptive control algorithm is proposed by integrating the existing trajectory linearization control （TLC） method with the SDO technique. Conditions are derived which guarantee ultimate boundedness of all the errors in the closed-loop system. Finally, the flight control system of the NSV is implemented by using the proposed method. Excellent disturbance attenuation ability and strong robustness of the proposed method are demonstrated.