Design and Analysis of an AI-Driven Adaptive Multi-Intersection Traffic Signal Coordination System for Smart Urban Mobility: A Review

Rapid urbanization and the exponential growth of motorized I. INTRODUCTION private vehicles have overwhelmed existing municipal transportation networks, creating chronic traffic congestion, elevated tailpipe emissions, severe travel delays, and increased emergency response vulnerabilities. Urban traffic congestion represents one of the most persistent and costly Traditional traffic control systems predominantly operate using isolated operational challenges faced by modern municipal authorities. The fixed-time schedules or rudimentary actuated timers that lack real-time exponential surge in vehicular volume, combined with static urban road visibility and inter-intersection coordination. Consequently, green phases infrastructure, has pushed municipal arterial networks past their designed are routinely misallocated to vacant approaches while perpendicular saturation thresholds. Conventional traffic control systems—which corridors suffer massive queue build-ups. This paper presents a continue to govern the vast majority of municipal road networks—rely comprehensive review of intelligent transportation systems, real-time primarily on pre-programmed, fixed-time signal cycles derived from edge computer vision, and multi-agent reinforcement learning (MARL) historical averages [1]. Because these legacy controllers function without for multi-intersection traffic signal management. Based on the continuous situational awareness, they cannot adapt to unpredictable synthesized literature, an AI-Driven Adaptive Multi-Intersection Traffic fluctuations, localized bottlenecks, or directional peak-hour surges [2]. Signal Coordination System is reviewed and formulated. The proposed A fundamental structural flaw in existing urban mobility infrastructure is the framework employs YOLOv8 edge perception on localized nodes (such the operational isolation of adjacent intersections. In conventional setups, as NVIDIA Jetson) for spatial queue length and velocity profiling, a individual traffic signal controllers function as disconnected entities. multi-agent Deep Q-Network (MADQN) collaborative framework for Even when isolated junctions incorporate basic sensing hardware, such dynamic green phase optimization and green-wave corridor progression, as inductive loops or static infrared sensors, they process traffic volume FastAPI and WebSocket microservices for central telemetry and phase independently [3]. Flushing a dense vehicle queue from an upstream synchronization, Redis and PostgreSQL for low-latency caching and junction without communicating with downstream controllers persistent traffic analytics, and SUMO/CityFlow microscopic simulation systematically dumps an unmanageable wave of vehicles onto adjacent for reproducible validation. Component boundaries, data flow pipelines, lights, triggering cascading shockwaves and corridor-wide gridlocks [4], closed-loop MARL formulations, emergency vehicle preemption protocols, and evaluation metrics are systematically detailed without [5]. reporting uncollected empirical field data. The economic, environmental, and public safety consequences of this uncoordinated operation are severe. Prolonged vehicle idling and Index Terms—Adaptive traffic signal control, multi-intersection repetitive stop-and-go acceleration patterns consume millions of barrels coordination, multi-agent reinforcement learning (MARL), YOLOv8 of fuel unnecessarily and elevate urban concentrations of hazardous object detection, edge computing, SUMO simulation, intelligent greenhouse gases, including carbon dioxide (CO2), carbon monoxide transportation systems (ITS), green wave progression, emergency (CO), and nitrogen oxides (NOx) [6], [7]. Furthermore, legacy controllers preemption, smart urban mobility.