What follows is a blog post by Durham physics undergraduate, Thomas Rudkin.
I recently took part in EEOPS-D, a 3-day opportunity to build and explore the use of drones run by the Durham Institute of Research, Development and Innovation (DIRDI).
When we arrived, we first had some practice flying much smaller drones than the one we would build, to get our eye in with the controls. We then had an introductory talk from the DIRDI team to tell us more about the institute, and to give us some ideas about what drones are being used in innovation, such as for quickly surveying the moisture content of farmland (using the colour of the soil) to help inform irrigation strategies.
We then split into small teams and set about building our drones, attaching motors and the electronics necessary to control the drone to a plastic frame. We then connected the drone to the software (ArduPilot) that we were using to control it. A remote controller was also paired with the drone, so we could control it manually in the air.
On the second day, we took the drone for its first flight, observing how it responded to our steering inputs and coped with the wind. We then set about collecting flight data and running programs in ArduPilot to adapt parameters within the generalised drone software for our specific construction. This resulted in an obvious improvement in the stability and manoeuvrability of the drone. In pursuit of even more stability, we also collected vibrational data to identify the frequencies induced by the motors using a fast Fourier transform program. This allowed us to use functions in the software to attenuate these frequencies.
On the last day, we were able to successfully program the drone with pre-planned flight paths, the drone taking off, executing several direction and altitude changes before landing again. With the drone now being more stable after yesterday’s efforts, a camera was attached to it and we were able to capture clear footage (mostly of ourselves!).
Throughout this challenge, we had to work with our teammates to discover how to use the software, and to overcome technical problems. This was an occasionally frustrating but overall rewarding experience, with a largely autonomous drone to show for it at the end of the 3 days.
Participants in EEOPS-D 2023, with Thomas pictured top left.