最小二乘法在电子鼻动态信息融合中的应用

Application of Least-Square Algorithm to Dynamic Data Fusion of Electronic Nose

基于集成气敏传感器阵列和多传感器信息融合技术的电子鼻系统,可用于气体/气味的定性定量识别。电子鼻走出实验室的关键是高效的识别算法和简化的实验装置。本文提出最小二乘法在简化的电子鼻系统中的应用模型。利用线性最小二乘法拟合传感器阵列的动态响应曲线,建立电子鼻系统的知识库并进行智能识别。利用该方法并结合气体传感器阵列在不同温度条件下的动态响应信号,对雪碧、酷儿和绿茶三种饮料样本进行定性识别,仿真验证结果表明该方法的识别准确率达到100%。

Electronic nose system based on integrated gas sensitive sensors array and multi sensors information fuse technology can be used to identify qualitative and quantitative gas/odor. High effective recognition algorithm and simple device are keys for electronic nose used in practical conditon. Application model of least-square algorithm in simple electronic nose system is introduced. Dynamic response curve of sensor arrays is fitted using linearity least-square method. The electronic nose system forms arepository and do identify intellectually. It can identify different beverages as sprite, Qoo and greentea combining this method with the dynamic responses of electronic nose under different working temperature of sensors. Results show that preciseness ratio of recognition reaches 100%.