Applications of IoT and Artificial Intelligence in Water Quality Monitoring and Prediction: A Review

Currently, internet of things (IoT) devices like environmental sensors are used to capture real-time data that can be viewed and interpreted via a visual format supported by a server computer. However, to facilitate modeling and forecasting, artificial intelligence (AI) techniques are effective in statistically analyzing complex non-linear systems and a large amount of historical data series within a short period. This present review article covers selected research journals published from 2014 to 2020. The findings from previous research indicate that despite the limitations of artificial neural network (ANN) tools, ANN has proved to be useful and powerful techniques that can be used in the field of hydrology. Similarly, ANN tools have the ability to evaluate historical data collected from different river stations and wastewater treatment plants with minimum errors within a short time. Therefore, based on the selected past literature used for this review we found that different types of ANN algorithm such as feed-forward backpropagation (FFBP) algorithm, gradient descent, Broyden-Fletcher-Goldfarb-Shanno (BFGS), conjugate gradient, radial basis function neural networks (RBFNN), neural network fitting (NNF), cascade forward back propagation (CFBP), ensemble ANN (EANN) and single AAN (SANN) have been employed in the prediction and monitoring of water quality parameters with satisfactory outcome. Furthermore, modeling alongside forecasting of water quality parameters would act as a big leap for government agencies and independent organisations in monitoring, decision making and regulating waste discharged into natural water bodies in order to achieve a safe and improved water quality for users.