Abstract:

Delivering water quality in campus and semi-urban areas is extremely important for the health, agriculture, and safety of ecosystems. Conventional water quality monitoring suffers from latency, discontinuity, and reliance on fixed infrastructures. This work provides a wireless water quality monitoring system that combines long-range, low-power LoRaWAN sensing with ultra-reliable, high-speed connectivity to deliver distributed edge intelligence in real-time. The sensor nodes send their data to an outdoor LoRaWAN gateway (DLOS8N), which sends it to the cloud using The Things Network. A machine learning pipeline uses Random Forest and BiLSTM models to classify the water status as safe, moderate, or unsafe. A real-time dashboard and alerting system was implemented using Flask and ThingSpeak. The proposed architecture supports scalable deployment of real-time water quality classification and fault detection in resource-limited areas.