神经控制器的典型结构

Typical Structures of Neural Control

本文研究了神经控制的典型结构方案，包括神经学习控制、神经直接逆控制、神经自适应控制、神经内模控制、神经预测控制、基于小脑模型联接控制（CMAC）、多层神经控制和递阶神经控制等．文中着重讨论了控制机理，提供了潜在应用基础，对神经控制器结构方案的确定具有借鉴作用．

In this paper some typical structural schemes of neurocontrol have been briefly studied. The schemes include NN-based learning control, NN-based direct inverse control, NN-based adaptive control,CMAC-based control, NN-based internal model control, NN-based predictive control, multilayered NN-based control,and hierarchical NN-based control,etc. This servey focuses on the control mechanism and provides a basis for potential applications. Most of the structural schemes have been used in various control fields.