Self-Adaptive Edge-Intelligent IoT Framework for Real-Time Water Quality Monitoring and Predictive Control

Abstract

Monitoring water quality is important for environmental safety and public health in places with fast-changing or unstable water conditions. Conventional monitoring approaches are slow to respond, not scalable, and lack real-time decision-making ability. To address these limitations, a research framework is proposed based on an intelligent edge-IoT architecture for real-time monitoring and assessment of water quality parameters. The system combines sensors such as temperature, pH, turbidity, and salinity with an ESP32 microcontroller for fast data sensing and processing. Edge computing is used to reduce latency, decrease dependence on the cloud, and improve real-time responsiveness. Experimental validation was conducted in three districts of Basrah, Iraq: Al-Maqal, Shatt Al-Arab, and Al-Karma. A hybrid method that fuses IoT sensor data with laboratory reference measurements was used for validation. Time-series data were collected to track environmental changes. The system was evaluated using accuracy, recall, F1-score, MAE, and RMSE. The proposed system achieved 97.0% accuracy, 96.7% recall, and 95.1% F1-score, with low prediction error. It also performed better than recent methods in classification accuracy and prediction error. The results indicate that combining edge computing and IoT sensing improves the efficiency, accuracy, and reliability of water quality monitoring systems. In contrast to existing cloud-based or single-solution approaches, the proposed system is distinguished by combining edge sensing, statistical anomaly detection, and hybrid predictive control on a single, low-cost home device.