Development of a Low-Cost, High-Resolution Sensor Array for Real-Time Monitoring of Agricultural Ammonia Emissions

Ammonia (NH₃) is a major agricultural pollutant that contributes to air quality degradation, ecosystem disruption, and human health risks. Despite its significance, accurate monitoring of NH₃ emissions remains challenging due to their pronounced spatial and temporal variability, and the limitations of existing measurement technologies in terms of cost, scalability, and spatial resolution. In this study, we present an innovative, low-cost, high-resolution sensing platform designed for real-time, in-situ NH₃ monitoring in agricultural environments. The system integrates a carefully curated array of metal-oxide (MOX) and electrochemical (EC) sensors, combined with machine learning models to predict atmospheric NH₃ concentrations. Sensor selection was informed by prior literature and validated through controlled laboratory experiments using synthetic gas mixtures of NH₃, CH₄, and VOCs under stable humidity conditions. A Taguchi factorial design and Principal Component Analysis (PCA) were applied to assess sensor performance and identify key contributors to NH₃ signal attribution. While MOX sensors exhibited strong sensitivity, EC sensors provided enhanced selectivity and stability at higher concentrations, supporting the implementation of a dual-model prediction approach. The final instrument incorporates active temperature regulation, environmental compensation, and autonomous data transmission, ensuring suitability for long-term deployment in remote agricultural areas. This work offers a scalable and cost-effective solution for enhanced NH₃ monitoring, with implications for both emission mitigation and regulatory compliance.