Abstract: A discussion is devoted to the multiple model control method for the new generation fighter in this paper. The full supermaneuverable flight envelope of the new generation fighter is approximated by the multi-nonlinear models at the different equilibrium points. The dynamic inversion controller is designed for every nonlinear sub-model. The switching rule of the controllers is given by the fuzzy neural network. Meanwhile, the inputs and outputs of these dynamic inversion controllers are used to train the neural network. Finally, a uniform neural network controller is given for the full supermaneuverable flight envelope. To demonstrate the availability of the designed robust flight control system the high angle attack flight simulation is carried out. The simulation results suggest that the uniform neural network controller is valid.

Key words: new generation fighter, dynamic inversion control, multiple model control, robust flight control