Edge-Cloud-assisted Real-Time Approach for Water Quality Management for Smart Cities

Urbanization is a rapid process that causes serious difficulties in managing the quality of water. In such situations, the role of intelligent and real-time monitoring systems can safeguard public health and sustainability of the surrounding environments. The proposed framework in this paper consists of an integrated Edge-Cloud architecture to monitor the water quality in real-time within smart cities by monitoring water parameters like the pH, DO, BOD, nitrate levels and alike, perceived through IoT-based sensors. The Edge Layer calculates the Water Quality Index (WQI) and sends it to the Cloud Layer where it can further be processed. Several Machine Learning models viz. Decision Tree, AdaBoost and Gradient Boosting are tested in terms of accuracy and execution time to analyze the water quality, and the Decision Tree algorithm is found most efficient in the real-time context. On Cloud, Space-Time Autoregressive Integrated Moving Average (STARIMA) model is used to augment the Edge-assisted Water Quality analysis, to keep track of contamination through space and time. STARIMA exhibited a high predictive accuracy in ten monitoring stations whereby it can be used to detect and track the sources of contamination ahead of time. The system enables a user interface, having alerts, dashboards, and mobile notifications, which give power to municipal authorities to act promptly. The efficiency of the proposed architecture, its scalability, and subsequent opportunity to enhance the water quality management in urban conditions are confirmed by experimental results.