摘要
针对余度伺服系统中的非线性和不确定性 ,进行了神经网络建模的研究。利用网络预测误差的相对平均值对神经网络的泛化能力进行了定量的分析。结合文中提出的动态递归网络 ,通过对三余度伺服系统的仿真 ,验证了本文提出的方法是可行的。
The paper describes a modeling research on neural networks which aims at the non-linearity and uncertainty of redundant servomechanism. Making use of the relative average of network prediction uncertainty, the quantitative analysis is carried out for the ability of generalization. Combining the DRNN offered in this paper and through the simulation , the way described in this paper is proved feasible.
出处
《导弹与航天运载技术》
北大核心
2004年第6期5-8,共4页
Missiles and Space Vehicles
关键词
伺服机构
神经网络
三余度
建模
泛化
Servomechanism
Neural networks
Triplex redundancy
Modeling, Generalization
