Recent Advancements and Applications of Electronic Nose Systems in Environmental Monitoring and Pollution Detection: A Comprehensive Survey

Abstract

This article reviews some of the recent developments in Electronic Nose (E-nose) technology, which is used for environmental monitoring and pollution control and detection. An E-nose model is a sophisticated way of emulating the human sense of smell, which is used to detect VOCs (volatile organic compounds) and hazardous gases mostly used in industrial and urban settings. The use of machine learning models and wireless sensor networks (WSN) has improved their scalability and increased their efficiency for use in real-time pollution control. The survey spans the period from 2020 to 2024 on the available technological advances for the different types of chemical sensors, particularly their sensing materials consisting of metal-oxide semiconductors (MOS) and graphene-based arrays, and their operating software platforms that bring about real-time monitoring using IoT. The study also reviews the experimental findings and the associated theoretical advances in the areas of chemical detection, specifically the study of machine learning techniques such as support vector machines (SVM) and artificial neural networks (ANN) that are observed to enhance the precision of chemical detection and the performance of sensors. This review is derived from 40 peer-reviewed papers, goes through a phase of preliminary selection from different major academic databases, including IEEE Xplore, MDPI, ScienceDirect, and SpringerLink. The literature is divided into major categories based on the thematic approach, including sensor advancements, machine learning integration, and IoT/WSN applications. The reviews also explore the newest technologies, such as energy-harvesting and edge computing, which work for the improvement of the energy efficiency of E-nose systems. The survey reports that sensor technologies, specifically graphene-based sensors and low-power wireless communication protocols, have greatly increased the scalability and deployability of E-nose systems in large and complex environments. The integration of machine learning has also helped in mitigating issues such as sensor drift and improving the system’s ability to classify VOCs. However, challenges still remain, such as the requirement of better energy management and drift compensation mechanisms, among other things. The survey identifies key future research areas including the need for self-calibrating sensors and larger, more labelled datasets for currently used machine learning models that are struggling to adequately generalize their learning to different environments.