Gas discrimination based on enhanced gas-species related information obtained by a single gas sensor with novel temperature modulation

A portable electronic nose (e-nose) with miniature size and low-power consumption allows us to utilize the information of smells, greatly contributing to industry and daily life. In order to realize such a miniature and lowpower e-nose, acquiring the data including plenty of information of gas species or smells with a reduced number of sensor elements is required. For this issue, this study proposes novel temperature modulation based on heater voltage, which amplitude and frequency periodically change. We first conducted gas discrimination tests based on the data obtained from a single gas sensor. The proposed method enabled to obtain responses under various temperatures and frequencies at one time. Furthermore, the periodical changes in amplitude and frequency caused the hysteresis characteristic to the gas species in the time-dependent response. Based on principal component analysis, we confirmed the improvement in the gas discrimination performance compared with that by conventional temperature modulation with sinusoidal heater voltage. Gas discrimination using machine learning algorithms was also carried out. The discrimination accuracy was revealed to increase with the dimensions of the principal component, reaching 98.8% at the maximum. Consequently, the promising properties of the proposed temperature modulation for obtaining enhanced information of gas species are demonstrated.

Automated summary

This study introduces a novel temperature modulation method based on heater voltage for gas discrimination, which allows for acquiring information of various gas species or smells with a reduced number of sensor elements. Experimental results show that this method enhances gas discrimination performance and can be applied for gas discrimination using machine learning algorithms.