Autonomous Aerobatic Flight of a Fixed Wing Unmanned Aerial Vehicle

This thesis relates to the successful development of a flight control system to perform a range of aerobatic manoeuvres autonomously. The project is the first to try to extend the flight control capabilities of the Computer and Control group at the University of Stellenbosch. A simplified mathematical aircraft model is developed which encapsulates the important dy- namic characteristics of the airframe. It is demonstrated how computational fluid dynamics software can be used to calculate the stability and control derivatives of a conventional air- frame. A vehicle independent kinematic state estimator is presented and used to obtain the complete aircraft state vector. The estimator makes use of extended Kalman filter theory to combine a series of low quality sensor measurements in an optimal manner. A model predictive control strategy is then used to regulate the aircraft about arbitrary, time variant trajectories. The controller’s architecture is not in any way specific to the aerobatic manoeuvres demonstrated in this project. The avionics and ground station used for the implementation of the estimator and control algorithms are presented. The development of a hardware in the loop simulator is discussed and used to verify the correct implementation of the respective algorithms. Finally, practical results from two days of flight tests are presented.