IoT and AI-Driven Dynamic Traffic Signal Manipulation using V2I Communication for Advanced Urban Transportation

Urban traffic congestion significantly impacts mobility, fuel efficiency, and emergency response times, necessitating intelligent and adaptive traffic management solutions. This paper proposes a Vehicle-to-Infrastructure (V2I) enabled dynamic traffic signal control system that leverages AI-driven Adaptive Control Logic (ACL) and Q-learning to optimize signal timings within geofenced intersections. The system employs Raspberry Pi-based edge controllers, GPS-integrated vehicular nodes, and MQTT protocol over LTE networks to ensure real-time, low-latency data exchange. Q-learning facilitates autonomous policy optimization by learning from historical traffic patterns, while ACL dynamically adjusts signal durations based on vehicle density and priority metrics. By integrating reward-based signal adjustments, the proposed system achieves efficient traffic flow regulation, emergency vehicle prioritization, and adaptive congestion mitigation. Simulation results demonstrate significant reductions in average vehicle wait times and enhanced network scalability, positioning this approach as a robust solution for next-generation smart transportation infrastructures.