论文部分内容阅读
结合反馈线性化和神经网络提出了一种新的飞行控制系统设计方法，通过反馈线性化将非线性耦合系统等效转换为线性解耦系统，利用具有在线学习能力的神经网络补偿反馈线性化的误差，建立了基于自适应神经网络的控制结构，并利用李亚普诺夫函数导出了网络权值的自适应调整规则。在巡航弹地形跟踪应用中的结果表明，该控制系统不仅具有精确的跟踪性能而且具有良好的鲁棒性。
A new design method of flight control system is proposed based on feedback linearization and neural network. The feedback linearization is used to convert the nonlinear coupling system into a linear decoupling system. The neural network with online learning ability is used to compensate the feedback linearization Error, the control structure based on adaptive neural network is established, and adaptive adjustment rules of network weights are derived by Lyapunov function. The results of cruise missile topography tracking application show that the control system not only has accurate tracking performance but also has good robustness.