Flight envelope limiting systems using neural networks

This paper discusses the development of a flight envelope limiting system that uses neural networks to calculate the control deflections which result in a flight envelope limit exceedence. This information is then used to provide tactile cueing through variable force-feel controllers. The system is demonstrated in a simulation of the XV-15 tilt-rotor aircraft The results show the capability of the system to handle multiple envelope limits by providing force cues in multiple control axes. The system has also been applied to provide normal load factor and angle-of-attack limiting on the V-22 tiltrotor aircraft and evaluated in real-time piloted simulation. Results show that the pilots can maneuver along limit boundaries without exceeding the limits, thus the cueing system expands the effective safe maneuvering envelope of the aircraft.