Review–Modern Data Analysis in Gas Sensors

Abstract

Development in the field of gas sensors has witnessed exponential growth with a multitude of applications. The diversity of the applications has led to unexpected challenges. Recent advances in data science have addressed the challenges such as selectivity, drift, aging, limit of detection, and response time. The incorporation of modern data analysis including machine learning techniques have enabled a self-sustaining gas-sensing infrastructure without human intervention. This article provides a birds-eye view on data enabled technologies in the realm of gas sensors. While elaborating the prior developments in gas-sensing related data analysis, this article is intended as an entrant for enthusiasts in the domain of data science and gas sensors.

Sniffing Like a Wine Taster: Multiple Overlapping Sniffs (MOSS) Strategy Enhances Electronic Nose Odor Recognition Capability

“…[ 17 ] With the development of deep learning, event‐driven artificial intelligence (AI) techniques such as recurrent neural networks (RNN) and convolutional neural networks (CNN) are involved in processing abundant gas sensing data of electronic noses and achieving higher performance in complex odor recognition tasks. [ 18 ] Nowadays, bio‐inspired strategies for enhancing the sensitivity and selectivity of electronic noses have attracted the attention of researchers. [ 19 ] For example, more selective sensitive sensing materials are designed based on biological proteins and bio‐probes to mimic olfactory receptor structure.…”