Integrating MANETs and Hybrid Deep Learning for Enhanced River Water Quality Monitoring

Rivers play a critical role in supporting ecosystems and human activities. However, water quality in rivers is increasingly threatened by various pollutants. The growing threats to river ecosystems demand immediate and effective solutions to monitor and mitigate pollution. Traditional water quality monitoring methods have several concerns such as high cost and difficulty to deploy in remote areas. The necessary actions are needed to control river pollution. Mobile Ad-hoc Networks (MANETs) offer an effective solution for water quality monitoring in real-time based on their infrastructure-less, selfconfiguring, and scalable nature. In this work, the integrated solution is proposed for water quality monitoring. It includes new clustering techniques and deep models for effective communication prediction. For reliable communication, a new clustering protocol is proposed by considering multiple clustering metrics. The Cluster Head (CH) is selected using the metaheuristic optimization algorithm of the Walrus Optimization Algorithm (WOA). To achieve higher accuracy in prediction, a new hybrid deep learning (DL) model combines Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) models. Simulation results show the proposed approach is better in terms of delay, packet delivery ratio (PDR), and energy consumption when compared to other techniques.