Abstract: This paper describes a neural network based direct adaptive control approach to the problem of reconfigurable flight control of a super-maneuvering aircraft . The base control law is designed in nonlinear dynamic inversion, and neural networks are used to cancel the inversion error induced by many reasons, especially by actuator failures. Simulation of two types of actuator failures show that through adaptively canceling the inversion error, neuralnetworks can reconfigure the control instantly when there should be actuator failures. The system will keep steadyand maintain certain handling qualities.

Key words: neural network, super-maneuver, flight control, dynamic inversion, adaptive control