Abstract
Rapid industrialization and urban expansion have significantly increased carbon emissions, posing serious threats to environmental sustainability and public health. Conventional carbon monitoring methods rely on manual sampling or static monitoring stations, which lack real-time responsiveness and predictive intelligence. This paper presents the design and implementation of an AI-driven carbon emission monitoring system that integrates Internet of Things (IoT) sensors with machine learning (ML) algorithms for real-time data acquisition, analysis, and prediction of carbon emissions. The proposed system employs distributed gas sensors to continuously monitor CO₂ concentration and related environmental parameters such as temperature and humidity. The collected data is transmitted through a wireless communication framework to a centralized processing unit, where machine learning models perform emission trend analysis, anomaly detection, and short-term emission forecasting. A supervised regression-based learning approach is adopted to predict emission levels with improved accuracy, while a threshold-based alert mechanism enables timely detection of abnormal emission events. Experimental evaluation demonstrates that the proposed system achieves high prediction accuracy with reduced latency compared to traditional monitoring approaches. The integration of AI with IoT enables scalable, energy efficient, and intelligent emission monitoring, making the system suitable for smart cities, industrial zones, and environmental regulatory applications.