AI-based traffic control and management systems - a synthesis of Macedonian research experiences

Under of influence of advances in Artificial Intelligence and Machine Learning technology, the transport sector is on the brink of a major revolution. Algorithms from these areas have been implemented in various fields, from autonomous vehicles, and innovative tools for gathering and visualizing traffic data for in-depth analysis of traffic flow characteristics, to the next generation of intelligent traffic control and management systems. Through two case studies, this paper aims to present the synthesis of Macedonian research experience gained over a long process of exploring the opportunities and challenges of development and integrating emerging technologies into the transport system. The first case study describes the implementation of a reinforcement learning (RL) algorithm in the design and testing of an adaptive control strategy at urban intersections. The second case study involves the application of a sophisticated AI-based tool for traffic flow data gathering, by video data processing. In both cases, the process of testing and evaluation is obtained in the simulation environment. The obtained simulation results demonstrated the efficiency of the implemented solutions in decreasing traffic congestion, and vehicle emissions, and increasing throughputs which benefits society as a whole.