Iuliia Yamnenko

Fellowship for Ukrainian Scientists


National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”

TUM School of Engineering and Design

Prof. Constantinos Antoniou

Video Portrait


The project is called “Machine learning methods for traffic demand and supply prediction.” We are working on the management of road traffic. Data about traffic flow and volume is to be collected and then processed to provide efficient monitoring, prediction, and control.
Construction and validation of traffic simulation models is one of the basic activities in transportation system engineering. Thus, quality of data is a crucial element. We see an opportunity to use the prediction models to infer unknown data using readily available data. Diverse data could be obtained from open sources such as OpenStreetMap, open data portals, and publicly available data. For instance, open data platforms for Paris and Madrid provide archived as well as real-time traffic data. Private companies such as Uber, TomTom, and Google provide limited access to historical or even real-time traffic ­information.
There is a huge potential for machine learning and deep learning here due to the scale and diversity of the data. The combination of machine learning methods, deep learning neural networks, and pre- and post-processing of data is a new and promising approach for ­traffic management.