Deep Learning-Based Predictive Traffic Signal Optimization for Smart Cities: A Comparative Analysis of Centralized and Edge-AI Urban Transportation Architectures

Abstract

Rapid urbanization, increasing vehicle ownership, and growing mobility demand have intensified congestion, energy inefficiency, and environmental degradation within urban transportation systems. Conventional traffic signal control infrastructures frequently rely on static timing mechanisms or centralized adaptive control architectures that struggle to respond effectively to real-time traffic variability and heterogeneous mobility conditions. This article comparatively evaluates two intelligent transportation architectures for predictive traffic signal optimization: centralized cloud-based traffic intelligence and decentralized edge-AI traffic management systems. Using computational transportation simulation, deep reinforcement learning, urban traffic-flow analytics, and Internet of Things (IoT)-enabled sensing environments, the study investigates how distributed transportation intelligence influences congestion mitigation, travel-time efficiency, energy consumption, and environmental sustainability. Comparative analysis demonstrates that edge-AI traffic optimization significantly improves adaptive signal responsiveness, reduces average vehicle delay, enhances intersection throughput, and lowers transportation-related emissions under highly dynamic urban traffic conditions. The findings further reveal that decentralized transportation intelligence strengthens scalability and resilience within smart-city infrastructures by minimizing communication latency and enabling localized decision autonomy. This article contributes to transportation engineering scholarship by integrating intelligent transportation systems theory, deep reinforcement learning, edge computing, and sustainable urban mobility analysis into a unified systems-engineering framework. The study also provides operational implications for future smart-city governance, intelligent infrastructure deployment, and sustainable urban transportation transformation.