| 神经网络在校正传感器非线性度方面的研究 |
| |
| 引用本文： | 黄天戍,倪效勇,张红剑,任清珍.神经网络在校正传感器非线性度方面的研究[J].传感器与微系统,2003,22(12):48-50,53. |
| |
| 作者姓名： | 黄天戍 倪效勇 张红剑 任清珍 |
| |
| 作者单位： | 1. 武汉大学电子信息学院,湖北,武汉,430072
2. 红安电力局,湖北,红安,431500 |
| |
| 摘 要： | BP神经网络及其改进算法虽然在测量准确度方面有所改善,但结构的不确定性、训练时间长的缺点限制了其在现场测量系统的应用。在引入LM算法的同时,通过分段的方法对其缺陷进行了改进。仿真结果证明了该方法具有收敛速度快、实时性强的特点。
|
| 关 键 词： | BP神经网络 LM算法 分段 实时性 |
| 文章编号： | 1000-9787(2003)12-0048-03 |
Study on nonlinearty correction of sensors by neutral network |
| |
| HUANG Tian-shu,NI Xiao-yong,ZHANG Hong-jian,REN Qing-zhen.Study on nonlinearty correction of sensors by neutral network[J].Transducer and Microsystem Technology,2003,22(12):48-50,53. |
| |
| Authors: | HUANG Tian-shu NI Xiao-yong ZHANG Hong-jian REN Qing-zhen |
| |
| Affiliation: | HUANG Tian-shu~1,NI Xiao-yong~1,ZHANG Hong-jian~2,REN Qing-zhen~1 |
| |
| Abstract: | The precision of measurement is improved by BP neutral network and its new algorithm, but uncertain of structure and learning for more time limits its application. LM algorithm for BP neutral network is introduced and its defect is maked up by the subsection . In the end, the result of simulation testifies this method has rapid astringency and better real-time property. |
| |
| Keywords: | BP (back propagation) neural network LM (Levenberg-Marquardt) algorithm subsection real-time property |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
