采用动态神经网络的多模型自适应重构控制方法

Multiple-model Adaptive Reconfiguration Control Based on Dynamic Neural Network

针对复杂的系统,提出一种基于多模型结构的自适应重构控制方法,使得系统可以在不同的运行环境下跟踪给定的信号,并且对特定的故障情况具有控制重构的能力;首先,由多个线性模型和一个模糊模型构成多模型控制结构,并设计多模型自适应控制器的权值调整规则,以获得当前最佳的控制输入,再引入动态自适应神经网络以保证系统的稳定性,并避免模型切换等噪声干扰;最后,对某型歼击机进行正常和故障状态下的控制仿真,结果表明所提重构控制方法是可行有效的。

For the complex control system, a kind of adaptive reconfiguration control method using multiple models is presented in this paper, in order to make the controlled system track the given signal under different working conditions, and to reconfigure control law under some structural failure condition. First, the multiple--model control structure is formed combine several linear models with one fuzzy model, so the weight-- adjusting regulation of the adaptive controller are obtained in view of the multiple--model structure. Then a dynamic neural network is introduced to stable the whole system and eliminate the influence caused by frequent switching. The simulation results show the control method presented is effective by demonstrating normal flight process and that with failures for an uncertian fighter plane.