A reference governor with novel domain marching for robust stall avoidance in fixed-wing UAVs

The rise of fixed wing unmanned aerial drones has raised the need for solutions to pilot training and regulation obeying flight algorithms. Within this work, a numerical implementation of a reference governor framework is introduced, implemented, and validated using simulations and custom hardware. A sample application of creating a constraint obeying flight controller that can be used for novice pilot training is used to test the implementation. An angle of attack constraint is used along with a parallelized numerical MOAS computation algorithm to create a stall avoiding manual flight mode. To improve robustness against constraint breaking scenarios the early implementation versions contained, a novel algorithm named domain marching was introduced. Experimental results demonstrate the functionality of the reference governor implementation on flight ready hardware, with simulations resulting in stall avoidant flight characteristics. Additional tests demonstrate the reference governor can be adjusted to be resilient against hardware failures, adding fault tolerant capability.