WATERNET: A Network for Monitoring and Assessing Water Quality for Drinking and Irrigation Purposes

Water is the fundamental pillar of survival for all living organisms, yet its safety is increasingly threatened by industrial discharge, mining activities, and environmental pollutants.Adhering to World Health Organization (WHO) guidelines, metrics such as the Water Quality Index (WQI) and Irrigation Water Quality Index (IWQI) are critical for evaluating water suitability.However, traditional methods involving manual sampling, transportation, and laboratory analysis are often inefficient and slow.This study introduces WATERNET, a real-time water quality monitoring network that utilizes LoRa (Long Range) technology to provide continuous oversight.The network architecture was developed with careful consideration of land topology.Performance simulations conducted in Radio Mobile identified a partial mesh network as the most effective topology for ensuring reliable data transmission over varied terrain.To overcome the challenges of limited environmental data, custom datasets were generated to train and validate three machine learning models: Random Forest (RF), Logistic Regression (LR), and Support Vector Machine (SVM).These models were tasked with classifying water as suitable for either drinking or irrigation.Experimental results demonstrated that Logistic Regression achieved the highest accuracy for drinking water classification, whereas the Support Vector Machine proved most effective for irrigation assessment.Furthermore, Recursive Feature Elimination (RFE) was employed to isolate the most significant chemical and physical water parameters, optimizing the models' predictive performance.WATERNET offers a robust, automated solution for proactive water resource management.