摘要
自抗扰控制方法是一种新型的非线性设计方法,在自抗扰控制器中主要存在着确定待定参数的问题;非线性扩张状态观测器是自抗扰控制器的核心,在研究非线性扩张观测器中的参数整定问题时,将神经网络的思想引入参数整定,提出了基于神经网络的非线性扩张状态观测器的设计方法,运用该方法可以对任意阶的非线性扩张状态观测器进行参数设计;大量仿真算例表明,设计出的观测器具有良好的鲁棒性,有一定工程应用参考价值。
A new type of nonlinear method is named active disturbance rejection control （ADRC）. The main question exsited in ADRC is to determinate some unknown parameters of that. The core of ADRC is nonlinear extended state observer （ESO）. Based upon the capacity of nonlinear approximation of multiplayer neural networks, a design principle and algorithm for ESO is presented. The proposed neural network based ESO are can exactly track the extended states of nonlinear systems. Simulation results show that good robustness is guaranteed. With this work, practical application of ADRC is possible.
出处
《计算机测量与控制》
CSCD
2006年第8期1099-1102,共4页
Computer Measurement &Control
