USE OF NEURAL NETWORKS IN CONTROL OF HIGH-ALPHA MANEUVERS

A method is presented by which an appropriately constructed artificial neural network can be ''trained'' to Predict the force and moment coe fficients of a 70-deg sweep delta wing during a high-angle-of-attack e xcursion. The angle-of-attack time history is a sinusoidal motion from 0 to 90 deg and returning to 0 deg. Experimental data are used to tra in the network, and it is demonstrated that the network has indeed lea rned to model the behavior of the delta wing over a range of frequenci es of this type of angle-of-attack time history. The longitudinal equa tions of motion for a delta wing aircraft are integrated for three sin usoidal angle-of-attack time histories using the predicted network aer odynamic data. This integration generates the longitudinal control def lection time histories required to produce these maneuvers. An explora tion is then made as to whether a second artificial neural network can be trained as a neural stick gearing for such maneuvers. This is inve stigated by attempting to train a network to associate each required c ontrol deflection time history with a specified stick position schedul e.