Level of Service Classifications within the Smart City Concept Using Artificial Intelligence Tools

In terms of climate change, regulating traffic in cities is very important. The rapid development and emergence of artificial intelligence technologies provide new tools for urban traffic management in the Smart City concept. One way of classifying traffic is to use the standardized criterion of Level of Service (LoS). This criterion is used to assess the quality of traffic based on basic information about traffic density and speed and the specificities of the location. For this reason, the LoS for each location needs to be assessed individually with the assistance of an expert. In this paper, we propose and compare several approaches for LoS classification based on fuzzy sets, neural networks, and high-dimensional random vectors. We aim to minimize the dependence of the LoS determination methodology on local conditions, thus allowing it to be transferable to different locations. Based on the results of our comparison, it appears that all three methods provide results with sufficient accuracy, allowing their further use in combination with meteorological and pollution data.