Portable vehicular electronic nose system for detection of automobile exhaust

Qi Wang,Kai Song,Tiandong Guo
DOI: https://doi.org/10.1109/VPPC.2010.5729205
2010-01-01
Abstract:In this paper, we developed a powerful vehicular electronic nose system for detection of automobile exhaust gases such as monoxide and hydrocarbon (CO/HC). Commercially available metal oxide semiconductor (MOS) multi-sensor system and artificial neural network based gas pattern recognition method were used to improve selectivity of gas sensors and accurately discriminate gas components. A single chip was used for sampling and processing sensor response data as well as gave the automobile exhaust detection result. The classification for emitted gases was based on a momentum and adaptive learning rate backpropagation (BP) artificial neural network whose weights and biases were trained in advance and programmed in the microcontroller unit (MCU). Experimental results demonstrate that the system not only could effectively detect the individual components from their mixtures but also could monitor the risk grade of each gas with sufficient accuracy.
What problem does this paper attempt to address?