| 摘要: |
| 针对机动飞行器强耦合性、快时变性以及高不确定性的问题，建立了六自由度刚体运动数学模型，按照时间尺度将姿态运动模型分为快变子系统和慢变子系统并分别设计动态逆控制律。然后给出了神经网络自适应控制器的设计方法，对神经网络的逆误差补偿原理、输入的选择、权值矩阵更新规则进行分析。通过算例进行六自由度仿真，证明本文方法的跟踪精度较高。 |
| 关键词: 机动飞行器 动态逆 自适应控制 神经网络 |
| DOI： |
|
| 基金项目: |
|
| An Adaptive Maneuverable Vehicle Control Method Based on Neural Network |
| ZHANG Guangyong,ZHENG Xin,SHI Baolan,Hong Bei |
| (Beijing Institute of Astronautical Systems Engineering) |
| Abstract: |
| Considering the heavily coupling, fast time-varying and high uncertainty of maneuverable vehicle, six degrees of freedom rigid body model was built. The attitude motion model was separated into fast states subsystem and slow states subsystem in the time scale and dynamic inversion control system was designed. Then, an adaptive control method based on neural network was researched in the paper. The principle of neural network compensating dynamic inversion error, the choice of network inputs and the weight matrix update law were analyzed in detail. The validity of Neural Network adaptive control method was elaborated by six DOF simulations. |
| Key words: Maneuverable vehicle Dynamic inversion Adaptive control Neural network |