Gain scheduled model following control of flight control system based on neural network

Gain scheduling control method is widely used in flight control system, but the designing process is time-consuming. This paper introduces a kind of gain scheduled model following control method based on neural network. Using this method, all the parameters in flight controller can be computed automatically. The structure of controller and the principles of selecting referenced model are presented. At the same time, conditions of perfect model following are given and proved. The controller of a fly by wire system in full flight envelope is realized by using BP neural network. Finally, the closed loop system in several typical states is simulated. The simulation results show that the gain scheduled model following control method can satisfy the requirements of flying qualities in longitudinal short period motion.