It’s been confirmed that I get my first choice of dissertation subject! I’ve had a meeting with my dissertation supervisor to discuss ideas, and he thinks that we should try and build a quadropter from scratch. It would be one of those quadropters that has the propellors embedded in to a main disc shape, rather than on sticks. This will hopefull reduce the chance of a bird injuring itself. We will also add some small 3D printed arches over the blades to further prevent the bird contacting the blades.
This shape of UAV will be slower than a traditional quadropter (it will not be very aerodynamic), but this is not an issue for this project. It will only need to hover in place, not really moving around much.
I will be having a meeting with some falconers at the center I trained at to discuss the project further. The main design decision will be what sort of payload will be delivered- a lure on a bit of string for the bird to take from the UAV, or a piece of meat dropped from the UAV for the bird to chase. I will also be considering how many cameras (most likely GoPros) I can attach to the UAV depending on the weight of the payload. I will probably be aiming for the UAV to be able to carry around 500g. I will need to look in to how to calculate the carrying capacity based on a design.
It has been recommended to me that I use ArduPilot in this project. This will provide me free software to control the UAV from a laptop, as well as to act as the on-board controller for the UAV. I may use MavLink to send commands to the UAV from the simplified controller I hope to build. This will not deal with advanced functionality. It will only be required to make the UAV go up/down, and release the payload. Advanced functionality could be controlled using the laptop if needed. The control device I build will also have a screen that displays the current height of the UAV, as well as the remaining battery power potentially.