ECE Seminar - Smart Environmental Monitoring by Low-Power / Self-Powered Gas Sensors and Artificial Intelligence
 

ABSTRACT

With the advent of the 4th industrial revolution, there has been a significant rise in the use of Internet of Things (IoT) and advanced sensors in several sectors, including but not limited to smart factories, healthcare, environment, and entertainment. However, the deployment of an increasing number of sensors in IoT systems has made it crucially important but challenging to reduce their form factors and power consumption. Consequently, there is a growing need for sensors that are miniaturized, low-power, or self-powered. In particular, when it comes to environmental IoT, it is essential to decrease power consumption in gas sensors while simultaneously enhancing sensing performance factors, such as sensitivity, selectivity, and response speed. In this talk, we present some recent advancements in the development of environmental sensors that are miniaturized, low-power or self-powered. These sensors utilize functional nanomaterials, microfabricated sensor structures, and photo-activated sensing mechanisms, which were developed by our research group at KAIST. Additionally, we describe the implementation of deep learning-based signal processing techniques for ultra-low power environmental sensors. The following topics are covered in depth: (a) Low-power semiconductor metal oxide (SMO) based chemoresistive gas sensor array using localized hydrothermal synthesis, (b) Ultra-low-power SMO gas sensors monolithically integrated on a micro LED platform, (c) Self-powered gas sensors using chemo-optically modulating sensing films and photovoltaic cells, (d) Machine learning of gas sensor array for high performance electronic nose (e-nose) systems.