Investigation of models predicting NO_x level in the sample region and the use of intelligent transportation system

Artificial intelligence (AI), unlike natural intelligence, possesses the ability to problem-solving activities by machines. As AI-based models increasingly provide robust approaches to predicting air pollution, they are becoming more widespread. Intelligent transportation systems (ITS) are poised to be significant solutions for sustainable mobility. These systems, by appropriately enhancing mobility, will prevent the concentration of air pollution in a region through transportation. This study aims to examine AI-based models used in air pollution prediction and demonstrate the effectiveness of intelligent transportation systems in improving transportation-related air pollution. As a sample region, Kocaeli Province, which has highly polluted air, the amounts of transportation-related NO_x pollutants emitted from light and heavy vehicles passing through the Dilovası district were modeled using Adaptive Neuro-Fuzzy (ANFIS) and Artificial Neural Networks (ANN). The results were compared with the outputs of the Calculations of Emissions from Road Transport (COPERT4) program. The evaluations revealed that ANFIS performed better in modeling NO_x pollutants. Based on the prediction results, in case of exceeding the NO_x limit, an intelligent transportation system redirecting vehicles to alternative routes was suggested. For the use of this system, scenarios proposing the redirection of cars in varying proportions, including single-plate, double-plate, and light vehicles, depending on route redirection, were proposed and evaluated. The evaluation of scenario results showed that redirecting a large number of cars to alternative routes with the assistance of ITS resulted in a significant decrease in emissions.