基于神经网络的多模态控制器设计(英文)

Multi-Mode Controller Designing Based on Neural Networks

基于神经网络所具有的定性推理和定量数值并行计算能力 ,以及学习记忆能力 ,集成非线性系统的多个特征模型和控制器 ,实现了控制系统的多模态智能控制 .该方法充分结合系统的定性知识和定量数学描述信息 ,实现了参数大范围变化时变系统的良好控制 .最后用该方法对某型激光制导炸弹设计了一多模态控制器 ,仿真结果表明了该方法的优良性能 .

A method of intelligent multi mode control is introduced. The scheme is based on the neural networks which integrate multiple modules and controllers of nonlinear time varying system, and also the neural networks' abilities of qualitative reasoning, quantitative numeric reckoning, learning and memorizing. Combining the qualitative knowledge and quantitative mathematical information of system, the approach provides good control over the nonlinear time varying system with wide variation of parameters. Finally, we use this method to the control design of laser guidance missile and the result of the simulation shows that the performance is excellent and this approach is sample and easy to realize.