Intelligent IoT-Based Water Quality Monitoring System Using TD-CNN-LSTM Approach
M. Navaneetha Velammal, Vittam Rakesh, K. Kartheeban, Khalid Ali Qidwai, G. Shyamala, T Thulasimani
Abstract
Today, water contamination affects a large portion of the world's population. Manually collecting water samples from various sites and analyzing them in a lab as part of the routine monitoring technique. These approaches used to be effective, but they are now too slow and laborious to be practical. In addition, cutting-edge methods consider a wide range of physical and chemical elements. The standard method of quality monitoring and data sharing is time-consuming, unreliable, and costly. This emphasizes the need for constant, around-the-clock water quality monitoring. Basic Data Preprocessing, feature selection, and model training are the three main pillars of the proposed approach. The proposed method employs the preprocessing tools. The proposed system used the SOFM method (Structure-From-Function-Movement) to identify and eliminate unnecessary features selection. A model's accuracy can be measured with the help of prediction tools like utilizing CNN, LSTM, and TD-CNN-LSTM, models are trained. The proposed model outperforms two current top-tier contenders. The proposed method achieves higher accuracy (about 98.37%) when compared to other approaches like CNN and LSTM.