Flutter Suppression of Nonlinear Aeroelastic System using Adaptive Control based on Neural Network

This paper presents two adaptive control schemes for the stabilization of a nonlinear prototypical aeroelastic wing section with a trailing-edge and a leading-edge control surface. Using a Lyapunov approach, two types of controllers with high performance are designed based on neural networks. The first scheme accomplishes adaptive compensating control by eliminating uncertainties including structural nonlinearities, but certain parts of the system are required to design a linear feedback controller. Compared to the first scheme, the second does not need to consider internal information of the real system which is globally identified on-line using neural networks, and to improve robustness, approximate error is offset by adding extra terms to adaptive control laws. Results show that despite large uncertainties and unmodelled nonlinearities, the compensating controller is effective for flutter suppression, while the second scheme even have higher flutter speed than the former.