CO传感器神经网络补偿模型

Carbon Monoxide Sensor Compensating Model Based on Neural Network

摘要: 提出了一种神经网络数学模型补偿方法，采用最大最小距离法确定聚类数目，用聚类算法计算RBF神经网络中心和扩展常数，实验结果表明，补偿模型设计合理，达到预期要求,为煤矿CO传感器准确检测低浓度CO气体浓度提供理论依据。
- CO传感器 /
- RBF神经网络 /
- 补偿模型 /
- 聚类算法
Abstract: In this paper, the mathematical model of a neural network compensating method was presented, the maximum and minimum distance method was used to determine the number of clusters, RBF neural network center and expansion constants were calculated by using the clustering algorithm. The experimental results showed that the compensating model design was reasonable and achieved the desired requirements, which provided theoretical basis for detecting low carbon monoxide gas concentration accurately.
- carbon monoxide sensor /
- RBF neural network /
- compensating model /
- clustering algorithm

[1]	王强,余岳峰,张浩炯.利用人工神经网络实现函数逼近[J].计算机仿真,2002,19(5):44-47.
[2]	田社平,赵阳.基于神经网络模型的传感器非线性校正[J].光学精密工程,2006,14(5) :896-902.
[3]	贺玉凯.基于神经网络的矿用红外CO检测仪研究[D].北京:中国矿业大学(北京),2007.
[4]	飞思科技产品研发中心. 神经网络理论与MATLAB 7实现[M].北京:电子工业出版社,2005.
[5]	李一男,刘永平,王汝琳. 一种矿用红外瓦斯传感器的检测模型[J].工矿自动化,2010 (1):25-28.