V2I-RL: An Efficient RSU Selection Approach in VANET

Vehicular data traffic continues to increase thus the LTE networks face significant challenges in handling the high volume of vehicular communication demand, particularly in congested areas. Road Side Units (RSUs), which are strategically deployed along roadways, can alleviate network congestion by offloading traffic from the cellular network to dedicated vehicular communication channels. This offloading reduces the strain on base stations and enhances overall system capacity, reducing latency and improving throughput. In this paper, we propose a Vehicle-to-Infrastructure Reinforcement Learning (V2I-RL) approach aimed at intelligently selecting RSUs in congested areas to optimize the overall system performance in vehicular networks. Traditional communication protocols struggle to efficiently manage traffic and data transmission in high-density environments, leading to congestion, delays, and inefficient resource utilization. By leveraging reinforcement learning, our V2I-RL approach dynamically selects the most appropriate RSU based on historical traffic conditions, real-time vehicle mobility, and network load. This intelligent selection improves communication reliability, reduces latency, and enhances the overall throughput of the network, particularly in areas with high vehicular density. Our simulation results demonstrate that the proposed V2I-RL approach significantly outperforms conventional RSU selection methods, offering substantial improvements in system performance and network efficiency in urban environments.