Adaptive nonlinear optimal control for active suppression of airfoil flutter via a novel neural-network-based controller
This paper proposes a novel adaptive nonlinear controller based on neural-networks (NNs) for active suppression of airfoil flutter (ASAF) from the optimal control perspective. Optimal control laws for locally nonlinear systems are synthesized in real time by solving the Hamilton–Jacobi–Bellman equation online with a proposed new form of NN-based value function approximation (VFA) and an extended Kalman filter. A systematic procedure based on linear matrix inequalities is further proposed for designing a scheduled parameter matrix that generalizes the new form of VFA to globally nonlinear systems to suit ASAF applications. Un-modeled dynamics are captured using an NN identifier. Comparisons drawn with a linear-parameter-varying optimal controller in wind-tunnel experiments confirm the effectiveness and validity of the proposed control scheme.