基于自适应神经网络的非线性飞行控制

 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。

NONLINEAR FLIGHT CONTROL USING ADAPTIVE NEURAL NETWORK

A new design method for nonlinear flight control systems using feedback linearization with neural network is presented. Nonlinear dynamics systems are transformed into an equivalent linear system. Feedback linearization error is compensated by neural network which is capable of on line learning. A control structure is built using adaptive neural network. A stable weights adjustment rule is derived using a lyapunov function. It is applied to cruise missile terrain following. The result shows that the controller exhibits robustness and excellent tracking performance.