激光制导炸弹的神经网络鲁棒控制

针对激光制导炸弹滚转控制通道的时变特性，基于神经网络具有定性和定量多模态控制能力能够实现多个常规控制器的功能，融合炸弹投放过程中的多种工作状态参数信息，设计了非线性神经网络控制器，给出神经网络控制器与常规控制器的功能等价性分析。该控制器具有鲁棒性，能适应时变系统参数大范围的变化，而且方法简单，实现容易。利用该方法对某型激光制导航空炸弹进行仿真，并与炸弹的变结构控制器相比，从根本上解决了变结构控制器的抖振问题，结果表明，该神经网络控制器具有良好的控制性能。

Robust Control Based on Neural Networks for Laser Guided Bomb

This paper presents a method of neuro\|control for roll channel of laser guided bomb. The scheme is based on the capacities of neural networks that neuro\|controller can achieve multiple mode control with the ability of qualitative and quantitative control. Combining more kinds of information of laser guided bomb in variable conditions, the neuro\|controller can adapt to the wide variation of time\|variable system's parameters, furthermore this approach is sample and easy to be achieved. We also have proved that the neuro\|controller has the same abilities with conventional controllers within certain variation scope of system's parameters. Finally one neuro\|controller for some type laser guided bomb is designed, compared with the variable structure controller, this neuro\|controller can rid the tingle of variable structure controller perfectly, and the simulation results show the effectiveness of the proposed method.