We investigate the qualitative and quantitative analysis of gas mixtures using semiconductor metal oxide sensors. This analysis can be performed using multi-sensor systems at a constant temperature or using single-sensor systems at a variable temperature. The latter method is more reliable and efficient, but requires advanced statistical analysis of multidimensional data. Using pre-trained artificial neural networks allowed us not only to recognize analytes with reliability close to 100%, but also to determine their volume concentration with an error not exceeding 15%.
Paper
Full text
Processing Electronic Nose Data Using Artificial Neural Networks
Semantic Scholar · Engineering · 2020
Abstract
We investigate the qualitative and quantitative analysis of gas mixtures using semiconductor metal oxide sensors. This analysis can be performed using multi-sensor systems at a constant temperature or using single-sensor systems at a variable temperature. The latter method is more reliable and efficient, but requires advanced statistical analysis of multidimensional data. Using pre-trained artificial neural networks allowed us not only to recognize analytes with reliability close to 100%, but also to determine their volume concentration with an error not exceeding 15%.