基于全调节RBF神经网络的自适应飞行控制器

设计了一种基于径向基函数神经网络(RBF NN)的飞行控制器结构,运用李雅普诺夫综合法导出稳定的RBF NN参数调节律,以保证整个系统的稳定性.由于能在线地调节RBF NN的全部参数(连接权、高斯函数的中心和宽度),避免了因人为地估计中心和宽度参数而带来的性能损失,因而提高了控制性能.以F8战斗机为控制对象进行了仿真分析,仿真表明,在存在70%的模型误差的情况下,该控制器仍然能实现较好的跟踪控制,表现出很好的鲁棒性,远远优于传统的只调节连接权值的算法.

An Adptive Neuro Flight Controller Based on Fully Tunned RBF NN

A kind of structure of flight controller based on radial basis function neural network(RBF NN) is designed in this paper.Using the Lyapunov synthesis approach,a stable tunning rule for adjusting all parameters of this RBF NN is derived to ensure the stability of the overall system.Because all parameters(center and width of gauss function,weight) of RBF NN can be adjusted on line,and the discount of performance caused by artificially estimating parameters(center and width of gauss function) is prevented,the control performance is improved.A F8 fight aircraft is taken as the control object for performing simulation.Simulation results show that even by 70% model error,this controller manifests excellent control performance and good robustness.Fully tunned law is better than the conventional algorithm which only adjusts weight value of RBF NN.