The demand for gas sensors is increasing as interests in air quality monitoring related to environmental pollution and industrial safety grow. The semiconductor metal oxide (SMO) type sensor is preferred for its low cost, high sensitivity, mass production, and small size, but it suffers from poor selectivity. To solve this issue, an ultra-low-power electronic nose (e-nose) system was developed using ultraviolet (UV) micro-LED (μLED) gas sensors and a convolutional neural network (CNN). This e-nose system was highly selective, with a gas classification accuracy of 99.32%, and had a gas concentration regression error of 13.82% for five different gases. The μLED-based e-nose system is battery-driven, has a total power consumption of 0.38 mW, and is expected to be widely used in environmental internet of things (IoT) applications.
|