Flight test of flying-wing type unmanned aerial vehicle with partial wing-loss

Abstract

This paper investigates experimental evaluation via flight tests for applying adaptive neural network controller to a flying-wing type unmanned aerial vehicle experiencing partial wing-loss. For this, six-degree-of-freedom numerical model is constructed taking into account damage-induced changes to the unmanned aerial vehicle in aerodynamic coefficients, mass, center of gravity, and moments of inertia. Numerical simulations are performed to investigate the flight dynamics change and to verify the performance of the neural network based controller. During the flight test, main wing-loss is artificially generated by 22% or 33% area moment. The flight test verifies that the damaged unmanned aerial vehicle shows drastic roll behavior with the unstable longitudinal response, and the neural network based adaptive controller combined with feedback linearization successfully compensates for the wing damage.