|
4 |
Ultra-Low Power and Integrated Electronic Nose
Developing intelligent, low-power gas sensing technologies is essential for reliable, long-term gas monitoring in IoT and edge-computing environments. Our group addresses this challenge along three complementary directions: low-power MEMS gas sensing platforms with dedicated control electronics that minimize energy consumption at the hardware level; deep learning-based algorithms that overcome the selectivity limitations of conventional sensors to enable accurate, real-time gas identification and quantification; and neuromorphic in-sensor computing architectures, inspired by the biological olfactory system, that integrate sensing and signal processing to further reduce power consumption and hardware complexity. Together, these directions establish a comprehensive framework for next-generation, intelligent gas sensing platforms suited for real-world and edge AI applications.
|
|
3 |
2-1. Low-Power MEMS Gas Sensing Platforms with Integrated Control Modules
We have developed a low-power MEMS gas sensing platform by integrating suspended micro-heaters, micro-LEDs, and a dedicated sensor control module [1,2,3]. By minimizing heat loss through the suspended micro-heaters and enabling efficient optical energy transfer via the micro-LEDs, the platform achieves ultra-low power consumption of less than 10 mW for thermal sensing and approximately 0.1 mW for optical gas detection. To maximize sensing performance, we further developed a multi-channel control module capable of individually driving micro-heaters and micro-LEDs while accurately measuring the wide resistance range of semiconductor metal oxide (SMO) gas sensors through high-speed multiplexed switching. Combined with the excellent sensor-to-sensor uniformity and scalability of our MEMS fabrication technology, this platform provides a reliable, low-power gas sensing solution suitable for long-term IoT deployment in diverse environments.
- [1] Orbe, D. D., Kang, M., Cho, I., Choi, J., Park, J., Gul, O., Ahn, J., Lee, D. -S.* and Park, I*. (2023). Low-Power, Multi-Transduction Nanosensor Array for Accurate Sensing of Flammable and Toxic Gases. Small Methods, 2201352.
- [2] Lee, K., Cho, I., Kang, M., Jeong, J., Choi, M., Woo, K. Y., Yoon, K*., Cho, Y. H*. and Park, I*. (2023). Ultra-low Power E-nose System based on Multi Micro-LED-Integrated, Nanostructured Gas Sensors and Deep Learning. ACS Nano, 17, 539-551.
- [3] Lee, K., Kim, M., Kwon, Y., Park, S., Lim, Y., Kwak, D., Jeong, J., Kim, B., Ahn, J., Kim, J., Cho, Y-H., Kim, I-D*., and Park, I*. Enabling Weather-Independent Gas Detection through Deep Learning on Light-Activated Sensors. ACS Nano, 19, 327-339.
-
Figure 1. Low-Power MEMS Gas Sensing Platforms with Integrated Control Modules: (a) extremely small distance (the thickness of the SiO2 insulation layer is approximately 1μm) between the light source and the gas sensing material minimizes the energy loss and enables ultra-low-power gas detection (I. Park et al., ACS Nano, 2023); (b) synthesis of Bi-doped In2O3 NFs via electrospinning and their integration with μLEDs as a reliable sensing platform (I. Park et al., ACS Nano, 2025).
|
|
2 |
2-2. Deep Learning-Based Intelligent Gas Identification Systems
We have developed deep learning-enabled intelligent electronic nose (e-nose) systems to overcome the selectivity limitations of conventional gas sensors [4–6]. By integrating advanced deep learning algorithms with our sensing platforms, the developed systems enable accurate gas identification and concentration estimation from complex sensor response patterns. Furthermore, our approaches improve sensing reliability and practicality, providing intelligent gas sensing solutions for real-world monitoring applications.
- [4] Kang, M., Cho, I., Park, J., Jeong, J., Lee, K., Lee, B., Orbe, D. D., Yoon, K.* and I. Park.* (2022). High Accuracy Real-Time Multi-Gas Identification by a Batch-Uniform Gas Sensor Array and Deep Learning Algorithm. ACS Sensors, 7, 430–440.
- [5] Cho, I., Lee, K., Sim, Y. C., Jeong, J., Cho, M., Jung, H., Kang, M., Cho, Y. H., Ha, S. C., Yoon K.* and I. Park.* (2023). Deep-learning-based gas identification by time-variant illumination of a single micro-LED-embedded gas sensor. Light: Science & Applications, 12:95
- [6] Lee, B., Kang, M., Lee, K., Chae, Y., Yoon, K., Lee, D.* and I. Park.* (2025). Multi gas Identification by Temperature-Modulated Operation of a Single Anodic Aluminum Oxide Gas Sensor Platform and Deep Learning Algorithm. ACS Sensors, 10, 954–964.
-
Figure 2. Deep Learning-Based Intelligent Gas Identification Systems: (a) real-time classification and regression of target gases by the convolutional neural network (CNN)-based analysis of gas-sensing data (I. Park, K. Yoon et al., ACS Sensors, 2022); (b) temperature-modulated operation of a single gas sensor for multigas identification and application (I. Park et al., ACS Sensors, 2025); (c) deep convolutional neural network (D-CNN) for classifying ve gas species (air, methanol, ethanol, NO2, and acetone) and for quantifying the concentrations of each gas (I. Park, et al., Light: Science & Applications, 2023).
|
|
1 |
2-3. Neuromorphic In-Sensor Computing for Bio-Inspired Olfaction
To overcome the computational bottlenecks of conventional AI-based gas sensing, we have developed neuromorphic in-sensor computing technologies inspired by the biological olfactory system [7,8]. By integrating neuromorphic devices with gas sensors, our platform enables simultaneous sensing and signal processing, significantly reducing power consumption and hardware complexity. Building upon this concept, we have further advanced ultralow-power neuromorphic electronic olfaction systems for efficient and intelligent gas recognition. These technologies provide a promising foundation for next-generation edge AI sensing platforms.
- [7] Han, J-K., Kang, M., Jeong, J., Cho, I., Yu, J-M., Yoon K-J., Park, I.* and Choi, Y-K.* (2022). Artificial Olfactory Neuron for an In-Sensor Neuromorphic Nose. Advanced Science. 2106017
- [8] Kang, M., Han, J-K., Lee, K., Jeong, J., Yoo, C., Jeon, J-W., Park, B., Choi, W., Ahn, J., Yoon, K-J., Hwang, C-S*. and Park, I.* (2025). Neuromorphic olfaction with ultralow- power gas sensors and ovonic threshold switch. Science Advance. eadv9222
-
Figure 3. Neuromorphic In-Sensor Computing for Bio-Inspired Olfaction : (a) artificial olfactory neuron module mimicking a biological olfactory system and the practical application of the artificial olfaction system (I. Park et al., Science Advance, 2025); (b) biological olfactory system composed of various olfactory neurons including olfactory receptors, Glomerulicells, mitralcells, and interneurons in the olfactory bulb (I. Park et al., Advanced Science, 2022).
|
