为战斗机的过失速超机动飞行设计了一种神经网络控制器。基于时标分离的原则，神经控制器被设计为内环控制器和外环控制器两个部分。网络的训练利用改进的ＢＰ算法，将因子模糊化快速进行。样本点数据则由利用动态逆控制所得到的结果来提供。网络的使用利用联想记忆法则。仿真结果说明所设计的系统具有良好的控制性能及鲁棒性 A neural network controller is designed for flirting over maneuvering of fighter planes. Based on the principle of time-scale separation, the neural controller is designed as an inner-loop controller and an outer-loop controller. The training of the network utilizes the improved BP algorithm to quickly vaguely factor. Sample point data is provided by the results obtained using dynamic inverse control. The use of the network uses associative memory rules. The simulation results show that the designed system has good control performance and robustness