基于神经网络动态逆的飞艇定点控制及仿真

Fixed-Point Control Of Airship Based On Neural Network Dynamic Inversion

本文在分析飞艇自主定点控制重要性的基础上,采用神经网络动态逆控制设计飞艇水平面内的定点控制律。首先建立了飞艇水平面内的定点控制方程,然后利用神经网络补偿动态逆误差,建立了神经网络动态逆控制律,进行了仿真,并与未加入神经网络的动态逆控制方法进行了比较。仿真结果证明,在被控对象存在不确定性或者有外在干扰的情况下,本文设计的控制律控制律可以控制飞艇回到期望的定点位置,为以后进一步进行自主能力更高的定点控制律设计奠定了基础。

Design fixed-point controller of airship. The paper utilizes neural network dynamic inversion to design fixed-point controller of airship. First establishes the fixed-point model of airship, then designs neural network dynamic inversion controller, carries out computer simulations. Results of the example show that, the controller based on Neural Network dynamic inversion can control airship go back to expected fixed-point even when there is uncertainty and disturbs, which lays the foundation of the future autonomous fixed-point control.