Edge-Fog-Cloud Real-time Water Quality Sensing and Alert System

With the development of modern fisheries, there is an increasing demand for real-time water quality sensing, particularly for cage aquaculture. The study developed a sensor-equipped small boat powered by solar energy, used to measure various water quality parameters such as PH, conductivity, sea temperature, and dissolved oxygen. We utilized Arduino Mega 2560 (Arduino) and Narrowband Internet of Things (NB-IoT) for data collection and transmission of water quality information. The data transmission employed an Edge-Fog-Cloud architecture, and we also developed a web and a mobile application (App) for users to conveniently access the water quality data at any time. When the water quality data deviates from the normal range, the system will issue an alert to allow users to promptly investigate. Additionally, we obtained meteorological and oceanographic data from seven locations near the Penghu Islands from the Central Weather Bureau (CWB), which can also be viewed on the web and App. Finally, we proposed a predictive model called LSTM for sea temperature prediction. The model exhibited excellent predictive performance with a Mean Absolute Error (MAE) of 0.50, Root Mean Squared Error (RMSE) of 0.65, and R-squared (R²) value of 0.96. These results demonstrate that the LSTM model has outstanding predictive capability and is highly suitable for sea temperature forecasting.