Smart Transportation Networks Leveraging AI and Edge Computing for Real-Time Route Optimization

Rapid urbanization and the subsequent rise in vehicular traffic volume pose significant challenges to Intelligent Transportation Systems (ITS), primarily related to congestion, latency, and resource efficiency. The integration of Artificial Intelligence (AI) and Edge Computing offers a paradigm shift for real-time traffic management and route optimization. This paper proposes a distributed hybrid architecture leveraging AI and Edge Computing for real-time route optimization in ITS. The system utilizes Edge Nodes (at Road-Side Units and in vehicular units) for local data preprocessing, feature extraction, and immediate Reinforcement Learning (RL)-based decision-making. This decentralized approach minimizes the latency inherent in cloud-centric systems, enabling timely adaptive traffic signal control and dynamic routing adjustments. The cloud layer is reserved for large-scale predictive modeling and regional model synchronization via Federated Learning, which ensures data privacy and system scalability. Edge AI provides low-latency computation and maintains dependable performance, particularly in areas with unreliable or limited cloud connectivity (cloud-sparse regions). Tested via SUMO and VISSIM simulations, the proposed framework demonstrates significant improvements, including a reduction in average travel time (up to 30%), fuel consumption, and CO2 emissions. This work validates the efficacy of hybrid AI and edge computing for creating resilient, privacy-preserving, and high-efficiency smart transportation networks.

Ҳали таржима қилинмаган