Neural Network Model of Unsteady, Nonlinear Aerodynamics

Many flight control systems are developed from aerodynamic measurements obtained from static wind tunnel testing. These control systems frequently inadequately handle unsteady, nonlinear flight conditions. Dynamic roll angle measurements made in a wind tunnel have been obtained. This aerodynamic data presents a nonlinear, unsteady dynamical system. The roll angle trajectories have been successfully approximated with multilayer feedforward backpropagation neural networks.