AI-driven interpretable realitme TN Softsensor(IF=12.4, JCR 1.1%)

Prof. SungKu Heo published a research paper on an interpretable AI-based soft sensing framework for real-time total nitrogen (TN) monitoring in urban rivers in Water Research. The study introduces IRIS-TN (Interpretable Real-time In-situ Soft Sensor for Total Nitrogen), a deep learning framework that combines data-quality enhancement, multi-periodicity-aware temporal representation learning, and explainable artificial intelligence (XAI) to enable accurate and reliable TN estimation under dynamic river conditions. The proposed framework demonstrated superior predictive performance compared with existing deep learning models while achieving real-time edge deployment with an inference latency below 10 ms. By providing interpretable and continuous TN monitoring without relying on expensive online analyzers, IRIS-TN offers a practical solution for intelligent water quality monitoring and real-time decision support in digital water systems. The paper is accessible via the following DOI link: https://doi.org/10.1016/j.watres.2026.126272.