Artificial Sensing: AI-Driven Electronic Nose for Real-Time Gas Leak Detection and Food Spoilage Monitoring

Olfactory dysfunction poses a significant threat to health and safety by hindering the detection of hazards like gas leaks and spoiled food, highlighting the limitations of subjective and non-real-time traditional methods. This study presents Scent Sense, an AI-driven Electronic nose (E-nose) syste...

Full description

Saved in:
Bibliographic Details
Main Authors: Lubna Aziz, Hassan Adil, Raheel Sarwar
Format: Article
Language:English
Published: Sir Syed University of Engineering and Technology, Karachi. 2025-06-01
Series:Sir Syed University Research Journal of Engineering and Technology
Online Access:https://sirsyeduniversity.edu.pk/ssurj/rj/index.php/ssurj/article/view/680
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Olfactory dysfunction poses a significant threat to health and safety by hindering the detection of hazards like gas leaks and spoiled food, highlighting the limitations of subjective and non-real-time traditional methods. This study presents Scent Sense, an AI-driven Electronic nose (E-nose) system designed for real-time odor classification and food quality assessment. The system integrates multi-channel MQ-series gas sensors with a Seeed Studio Wio Terminal, leveraging near-sensor computing and machine learning for accurate detection. The development followed a structured, multi-phase approach: (1) gathering requirements from industry experts and individuals with olfactory impairments; (2) designing and integrating hardware and software components; (3) implementing machine learning models (Support Vector Machine, Random Forest, Artificial Neural Network) in Python and hardware interfacing in C++; (4) conducting rigorous testing across unit, integration, and real-world scenarios; and (5) deploying trained neural networks on the Edge Impulse platform for real-time inference. Sensor data underwent preprocessing, feature extraction, and exploratory data analysis before training and evaluation. Experimental results demonstrate 100% accuracy in gas leak detection and 99% accuracy in food spoilage classification, with sensitivity levels reaching 0.01 ppm for H₂S and NH₃ detection. Additionally, the system provides continuous real-time freshness tracking, ensuring reliable food monitoring. Future work will focus on enhancing sensor diversity, improving model generalization, and extending applications to industrial safety, environmental monitoring, and smart home automation.
ISSN:1997-0641
2415-2048