Autonomous system parameter identification
One of the challenges inherent in the development of autonomous aerial vehicles, is the tuning of autopilot systems. Tuning procedures typically involve manually flying the aircraft as tuning parameters are varied. These procedures require skillful operators, extended development time, and pose risk to the aircraft. Alternatives to these procedures often require access to sophisticated equipment such as wind-tunnels and state-of-the-art dynamics modeling and analysis software. Further, many tuning procedures result in conservatively tuned autopilot systems, which limit the flight envelope of the aircraft to simplify the tuning process.
To address this shortcoming, the design and implementation of autonomous sytems capable of updating tuning parameters in-flight are actively being researched within the RMRL. This work includes investigations into hybrid attitude determination techniques, adaptive and intelligent control methodologies, and experimental characterisation of such methodologies. To date, an autopilot system capable of autonomous self-tuning for high-wing fixed-wing aircraft has been developed within the RMRL.