Neural-network augmented model inversion control is used to provide a civil
ian tilt-rotor aircraft with consistent response characteristics throughout
its operating envelope, including conversion flight. The implemented respo
nse type is Attitude Command Attitude Hold in the longitudinal channel. Con
ventional methods require extensive gain scheduling with tilt-rotor nacelle
angle and speed. A control architecture that can alleviate this requiremen
t, and thus has the potential to reduce development time and cost, is devel
oped. This architecture also facilitates the implementation of desired hand
ling qualities and permits compensation for partial failures. One of the po
werful aspects of the controller architecture is the accommodation of uncer
tainty in control as well as in the states. It includes an online, i.e., le
arning-while-controlling, neural network that is initialized with all weigh
ts equal to zero. Lyapunov analysis guarantees the boundedness of tracking
errors and network parameters. Performance of the controller is demonstrate
d using a nonlinear generic tilt-rotor simulation code.