基于非线性PCA的微气体传感器阵列信号处理

Nonlinear PCA based micro gas sensor array signal processing

在线性叠加模型基础上提出了气体传感器对混合气体的非线性叠加模型,并引入了非线性主成分分析(NonlinearPrincipalComponentAnalysis,NLPCA)法对微传感器阵列的信号进行处理。使用该模型对由4个微热板式气体传感器组成的阵列的信号进行了分析,对照基于线性叠加模型的主成分分析法(PrincipalComponentAnalysis,PCA)的识别结果,说明该方法能够提高对混合气体识别和量化的准确度。

A nonlinear superposition model was proposed based on the common linear additive model the micro gas sensor array signal processing to improve the precision of quantification and identification. According to the nonlinear model, the Nonlinear Principal Component Analysis (NLPCA) was proposed to process the response signals obtained from a 4 Micro-hotplate (MHP) based gas sensor array. Com-pared with the analyzing results obtained from Principal Component Analysis (PCA), which bases on the linear additive model, the accuracy of gas component identification and concentration quantification are improved greatly.