Airspeed Forward-Invariance for Unpowered Fixed-Wing Aircraft

Researchers Huseyin Emre Tekaslan and Ella M. Atkins from the University of Michigan have developed a new algorithm to ensure the safety of unpowered fixed-wing aircraft. The paper, titled "Airspeed Forward-Invariance for Unpowered Fixed-Wing Aircraft," addresses a critical challenge in autonomous flight: when a drone loses power, its airspeed can only be controlled by exchanging potential energy (altitude) for kinetic energy (speed). This makes the aircraft highly sensitive to guidance commands, especially in windy conditions.

The team uses Nagumo's tangency condition to derive a closed-form, wind-dependent characterization of admissible guidance commands that guarantee the aircraft stays within a safe airspeed envelope. This condition is then embedded in an offline quadratic programming framework to certify safe maneuver primitives for non-ascending flight. The approach was validated using a high-fidelity model on gliding trajectories, demonstrating strict airspeed boundedness. This work is a significant step towards making autonomous flight safer and more reliable, particularly for small drones and electric air taxis that may need to glide to safety after a power failure.