Document Type
Article
Publication Date
2023
Abstract
The motivation for this research is to investigate an alternative method to reduce traffic congestion in cities different from automated traffic lights by utilizing AI with the anticipation that all vehicles will become autonomous in the future. This paper aims to answer the following research question: How can AI be leveraged to develop a centralized traffic control system (CTCS) and can it outperform standard traffic lights? A CTCS can direct traffic through intersections removing the responsibility from independent AI to coordinate with other cars that could potentially be following different protocols. Faced with a conflict of interest, two autonomous vehicles may cause delays, congestion on roads, or safety concerns. For this paper, an experiment is conducted within a closed simulation of a figure-8 track to compare the efficacy of four traffic control methods based on their throughput of cars, safety (car crashes), efficiency, and qualitative factors. It is anticipated that a CTCS will outperform standard traffic signals and results collected from this study can serve as supportive evidence for the possible implementation of this type of system. The visual simulation can also innovate on typical numerical simulations by incorporating more realistic driving conditions and providing qualitative analysis of different traffic control algorithms. The field of machine learning vehicles and intersection control to reduce congestion is well explored, but in the race for car brands to develop self-driving AI behind closed doors, research into systematic city-wide strategies for reducing congestion is limited. This alternative’s results to solving city congestion may demonstrate the feasibility of a machine learning CTCS that may drastically minimize city congestion, thereby reducing the need for road work, lane expansions, or complex AI communication to counteract increasing traffic. This could save city budgets billions of dollars and assist people to reach their destinations more efficiently.