On the Problem of UAV Collision Avoidance based on Adaptive Load Factor Control

Autonomous collision avoidance for small unmanned aerial vehicles (UAVs) is an important issue covered in many papers. Nevertheless, many questions remain regarding the implementation of adaptive obstacle bypass control with trajectory optimization. This is especially relevant for fixed-wing UAVs due to their limited maneuverability. This paper proposes a two-step solution. Step 1 is to modify the standard autopilot by adding a load factor control loop; Step 2 is to implement fuzzy model reference adaptive control (fuzzy MRAC) to preserve the quality of obstacle bypass trajectories. Simulations on full 6-DoF, 12-state nonlinear small-UAV models show that this approach is usable and effective. Thus, it enables more flexible control when using UAV collision avoidance algorithms compared to the standard altitude state machine.