Flight test of a Flying-Wing Type UAV with Partial Wing Loss Using Neural Network Controller

This paper presents flight test result of flying-wing type UAV with partial wing loss using neural network controller. 22% and 33% loss of wing area moment were considered for the damage configuration. A new trim state was obtained for each damaged model with the variation of mass, center of gravity and moment of inertia. A numerical simulation was performed to investigate the changed flight dynamics. It is verified that the damaged UAV shows sluggish roll behavior with unstable longitudinal response. A neural network based adaptive controller combined with feedback linearization was designed in order to compensate for the partial damage. It is verified that the instability caused by partial wing damage can be effectively controlled, and the overall system can be stabilized via neural network controller.