Research Information System
Osmaniye Korkut Ata Üniversitesi Fen Bilimleri Enstitüsü Dergisi (Online), vol.6, no.2, pp.1137-1158, 2023 (Peer-Reviewed Journal)
In direct proportion to developments in bandwidth technologies for Internet data transmission networks, the use of the internet in daily life is becoming more common. The concept of the Internet of Things (IoT) refers to the new technological ecosystem consisting of numerous objects that can be added to these technologies. One of the most important visions of the IoT is the concept of a smart city. This concept, which means that every component in cities, from communications to transportation, is connected to the internet and controlled and monitored by artificial intelligence-based computer algorithms, promises to ensure that increasingly crowded cities function in an orderly manner without descending into chaos. This study proposes a decision support system based on time series analysis that monitors traffic density in cities and makes future predictions. The proposed procedure is an Artificial Neural Network (ANN) based Time Series (TS) decision support technique. The study used the number of vehicles passing by three randomly selected junctions every hour as data. The relative effects of vehicle density at the junctions were calculated and traffic flow models were designed. The most appropriate traffic flow model is determined based on the accuracy of the forecast data provided by the models created. When the data are considered stable, predictions can be made with 93% accuracy for the ANN-based TS models J1 and J2 and 66% for J3. For the dynamic model, according to the design of the traffic flow, it was found that the model of serially connected traffic between the junctions has the highest accuracy with a joint mean value of 0.86.