Reconfigurable Control of Morphing Air Vehicle Based on Real-time Dynamic Modeling

This paper describes a reconfigurable flight control system for morphing air vehicles. The approach combines real-time physical model identification with Nonlinear Dynamic Inversion (NDI). It could solve the problem such as severe changes of aerodynamic characteristics or influences of atmospheric disturbances during the primary stage of flight, and fulfill the high robustness requirements of flight control systems. The aerodynamic characteristics are identified by the two step method which consists of a separate state estimation step and an aerodynamic coefficients identification step instead of the one step method. The two step method is based on iterated extended Kalman filter and fading memory least squares method which is easy to achieve on-line. The sensitivity of NDI to modeling errors is eliminated by making use of an on-line updated reference model of the air vehicle, then reconfigures for the changes in real time. Furthermore, the effectiveness and robustness of the designed flight controller has been validated by six Degree-of-freedom (DOF) nonlinear flight simulations with severe changes of aerodynamic characteristics on the morphing air vehicle and unknown atmospheric disturbances.