VTOL Freewing Design and Adaptive Controller Development

The freewing concept, describing a vehicle whose wing pitch attitude is separated from the fuselage and managed by aerodynamic moments, has been minimally studied as both a gust-alleviation and VTOL concept. The VTOL freewing configuration belongs to the class of transition aircraft that have traditionally required gain-scheduled linear controllers or control switching strategies to maintain stable flight in a wide range of operations, including the highly nonlinear transition region. This work implements a unified control strategy for a freewing unmanned aerial system (UAS), including all phases of vertical flight, transition, and forward flight. A neural net-based model reference adaptive controller (MRAC) architecture is implemented to manage both the inner loop attitude state and outer-loop translational state control. Modeling errors output from a dynamic inversion element are corrected with a single hidden layer neural network and techniques to combat actuator saturation are included. Simulation results of the freewing concept include successful transition to and from forward flight with the implemented adaptive controller. The adaptive autopilot is integrated with a commercially available flight control software on a flight testbed UAS with initial hover flight test results presented.