Discrimination of gases with a single chemiresistive multi-gas sensor using temperature sweeping and machine learning

Selective gas leakage detection is crucial due to the adverse effects of harmful gases present in the air on human health. However, existing high-temperature metal oxide gas sensors, which consume more power, are nonspecific and more sensors are required to discriminate different gases. On the other hand, room temperature gas sensors suffer from slow response and poor reliability despite their low power consumption. We proposed a method to discriminate three gases at relatively low power consumption with a single metal oxide gas sensor using temperature sweeping to address these issues. The machine learning classification algorithms in conjunction with the ternary logic of the sensor response at different temperatures were utilized to discriminate the gases. As the proposed approach requires just one metal oxide gas sensor instead of an array consisting of at least three nonspecific sensors to discriminate the multiple gases, the power consumption can be reduced significantly. As a feasible solution to the existing issues of conventional sensors, this novel methodology paves the way for the widespread use of the sensors in practical applications where reduction of power consumption is necessary.