基于RBF神经网络的导弹智能控制系统设计

为智能化导弹所设计的导弹智能控制系统应能够充分利用战场信息,自主而准确地生成控制指令完成目标打击。首先建立导弹控制系统模型,并在特征点处设计符合性能要求的PID控制器。在深入分析径向基函数(RBF)网络的结构与训练方法的基础上,通过大量仿真数据对RBF网络进行离线训练,将其训练结果直接作为俯仰与偏航通道的控制器。而滚转通道为典型的2阶系统,可采用滑模控制律,并利用RBF网络实时逼近外界非线性干扰项以提高滑模控制器的性能。通过某型倾斜转弯导弹六自由度仿真说明了本文所设计的智能控制系统的有效性。

Design of Missile Intelligent Control System based on RBF Neural Network

The missile intelligent control system designed for intelligent missile should be able to make full use of battlefield information and generate control commands automatically and accurately to achieve the target attack mission. In this paper, a model of missile control system is established and a PID controller is designed at the feature point. On the basis of in-depth analysis of the structure and training methods of Radial Basis Function(RBF) network, RBF network is trained offline through a large amount of flight data, and its training results are directly used as the controller of pitch and yaw channels. The rolling channel is a typical second-order system, which can adopt the sliding mode control law and use the RBF network to approximate the external nonlinear interference term in real time to improve the performance of the sliding mode controller. The effectiveness of the intelligent control system designed in this paper is demonstrated by the simulation of a certain type of STT missile with six degrees of freedom.