The flight dynamics of flying wing micro air vehicle (MAV) is remarkably nonlinear and unsteady because of MAV' S small size, low speed, special aerodynamic configuration and complex flight environment. The traditional control methods are incompatible with the development of MAV. The dynamic inversions to slow states and fast states are designed using the theory of time-scale separation. On-line neural networks are introduced to compensate the dynamic inversion errors, and pseudo control compensations are used to cancel the interaction between the actuators and the adaptive factors. The adaptive flight control system of flying wing MAV is studied based on the theories above and compared with the flight control system using dynamic inversion-PID. The simulation results demonstrate that the flight control system based on adaptive dynamic inversion is robust and capable of following commands. Compared with the dynamic inversion-PID control system, the adaptive control system is more suited to MAV.%飞翼式微型飞行器由于尺寸小、速度低、气动布局特殊和飞行环境复杂多变,其飞行力学具有显著的非线性和非定常特性,传统的控制方法已不能满足要求.本文运用时标分离理论,设计了快变量和慢变量动态逆,同时引入在线神经网络补偿动态逆误差,并采用伪控制补偿器消除作动器和自适应单元之间的相互影响,在此基础上提出了飞翼式微型飞行器的自适应飞行控制系统,并与采用动态逆-PID控制方法设计的飞行控制系统进行比较.仿真结果表明:基于自适应逆的飞行控制系统,具有较强的鲁棒性和指令跟踪能力,比动态逆-PID飞行控制系统更适合于微型飞行器.
展开▼
