Forecasting and Improving Latency in 5G V2X Networks for Autonomous Driving Scenarios

The advent of Artificial Intelligence (AI) has transformed telecommunications, particularly within 5G networks, amplifying their significance, especially in autonomous driving. Within this context, our study delves into the critical challenge of minimizing Round-Trip Time (RTT) in 5G Vehicle-to-Everything (V2X) networks. This endeavor is essential for ensuring efficient and real-time communication, a factor pivotal to the success of autonomous driving, and aligns seamlessly with the rigorous standards set by the 3rd Generation Partnership Project (3GPP). Our approach utilizes Machine Learning (ML) on real 5G mobility data to predict and manage RTT values exceeding defined thresholds for advanced driving scenarios. Additionally, we conduct tests involving numerology adjustments in 5G testbeds, exploring how this proactive mechanism can effectively lower the latency for the predicted high RTT cases obtained through the prediction approach, optimizing the communication in 5G V2X networks.