In the paper, Khan proposes an approach based on artificial intelligence to predict dissolved oxygen in an urban river environment.
Dissolved oxygen concentration in a water body is the most fundamental indicator of overall aquatic ecosystem health. Having a sophisticated measuring system to assess the health of these precious reserves is essential.
In urban areas, numerous factors, such as land use changes, can contribute to lowering of dissolved oxygen concentration in rivers, which can have devastating effects on ecosystems.
Using a form of machine learning known as a fuzzy neural network, Khan’s model automatically selects the optimum model structure for best performance.
Due to the highly complex and uncertain physical system, a data-driven and fuzzy number-based approach is preferred over traditional methods.
The model output is used to create a risk analysis tool that municipalities can use to identify the risk of low dissolved oxygen under different circumstances. This will allow cities to better monitor and anticipate the health of their rivers.
The full paper can be found here.