机构地区: 西北工业大学自动化学院
出 处: 《传感技术学报》 2004年第2期209-211,184,共4页
摘 要: 基于气体传感器阵列的混合气体分析方法及研究现状 ,对径向基函数神经网络对混合气体浓度预测进行了研究。用气体传感器阵列对由四种不同气体组成的混合气体进行测量并以这些测量数据为样本对径向基函数网络进行训练 ,训练后可使径向基函数网络对各气体成分的浓度预测误差不大于 6 %。 This paper briefly introduces the algorithm of quantitative analysis of gas mixtures and its history, and a radial basis function neural networks (RBFNN) is introduced for predicting concentration of gas mixtures. Gas mixture that consists of four different kinds gas is tested by gas sensor arrays. After RBFNN is trained by these data, RBFNN predicts the error of each concentration no more than 6%.
领 域: [自动化与计算机技术] [自动化与计算机技术]