Real-time optimization of traffic lights using artificial intelligence.

Abstract

Urban traffic congestion is a major challenge, reducing the quality of life for drivers and commuters. Traditional traffic lights worsen the issue due to their inability to adapt to real-time conditions, causing delays and inefficiencies. This thesis explores the application of artificial intelligence, specifically reinforcement learning, to create a smarter traffic light system that can dynamically adjust to varying traffic conditions. By making optimal decisions in different scenarios, this approach was implemented in a region of Tlemcen, where traffic congestion has been a notable issue. This real-world application demonstrates how the AI-based system can adapt to varying traffic conditions specific to the region, enhancing traffic flow, reducing congestion, and minimizing vehicle wait times at intersections. The use of AI in this context provides a more efficient and responsive solution compared to traditional traffic light systems.