The purpose of this research is to document the design and optimisation of a full suite of guidance, navigation and control (GNC) algorithms for a small unmanned aerial vehicle (UAV), the Skywalker X8. This was performed so as to ll a void in the available literature on the selected airframe, which currently only focuses on aspects such as aerodynamic modelling, advanced controller design, or uses of the airframe to perform higher level tasks. All of these research areas make use of o-the-shelf ight controllers, but these are not always the most appropriate foundations for more advanced work as they are inherently sluggish so as to be broadly applicable to a variety of airframes. Subsequently, the Skywalker X8 airframe was modelled, using existing literature, and then characterised so as to establish what the goals might be for an optimal set of controllers. An autopilot was then designed which was optimised so as to be as close to the identied optimal performance characteristics as possible, with eort being put into ensuring that all non-linearities and disturbances were taken into account. This included advanced modelling of sensors, actuators, the environment, and the system itself. The autopilot design was then extended with a set of guidance and navigation algorithms, also developed as part of this research. This consisted of both path planning and path following algorithms which allowed for the synthesis of general classes of paths useful to the application. With both the autopilot and guidance laws developed, the system could be tested under several atmospheric ight conditions. These took the form of various wind directions and intensity levels being applied to the airframe whilst transitioning between a range of dierent waypoint congurations. The system was subsequently shown to be able to follow a set of waypoints very accurately, even with winds and turbulence with magnitudes of in excess of 60% of the aircraft's nominal airspeed. With a strong autopilot designed and illustrated in a high delity simulation environment, this work can now easily be extended into many elds. All of the tools used for this research are available and well documented, and the processes followed repeatable with all justication available in the text. As such, should a project which aims to extend this work wish to adjust the autopilot design or guidance laws, based on dierent requirements, this is easily accomplished and recommendations of starting points are provided. The system model and autopilot are also made available and are usable exactly as they are should one wish to undertake additional research which does not aim to modify, but to extend this work.

show more