基于神经网络的自动着陆飞行鲁棒自适应逆控制

Robust adptive inversion automatic landing flight control based on neural network

针对飞机自动着陆飞行提出了基于神经网络的鲁棒自适应非线性动态逆控制器设计方案。首先采用非线性动态逆方法设计着陆飞行的基本控制律,再利用多层感知器神经网络设计适当的权值调整规则使其能够自适应地逼近和补偿逆误差。仿真结果表明,所设计的飞行控制系统是有效的,系统能够克服动态逆误差对着陆飞行控制带来的不利影响。

A adaptive dynamical inversion robust control scheme is proposed based on neural network for automatic landing flight.Firstly,the basic control law of landing flight is developed in nonlinear dynamic inversion;then,multi-layer perception neural networks is implemented to compensate inversion error by design of suitable weights.The simulation results show the effectiveness of the adaptive dynamical inversion robust flight control scheme,and the system overcome the influence of dynamical inversion error.