A NEURAL NETWORK CONTROLLER FOR SUPPRESSION OF WING ROCK

摘要

Artificial Neural Networks (ANNs) are known to be effective in controlling behaviour of non-linear and uncertain systems. Wing rock is one such highly nonlinear aerodynamic phenomenon seen, at high angles of attack, in fighter-class of aircraft with swept back wings. The dynamic motion manifests itself as a limit cycle roll oscillation. The paper presents the design of a feedforward neural network to suppress wing rock. Data for training the neural network are generated using experiments carried out in the wind tunnel on a slender delta wing model. Numerical results, based upon simulations on an approximate mathematical model of the phenomenon, show the effectiveness of the controller in suppressing the wing rock.