Deep Reinforcement Learning Based Resource Allocation with Radio Remote Head Grouping and Vehicle Clustering in 5G Vehicular Networks

Hyebin Park,Yujin Lim

doi:10.3390/electronics10233015

Hyebin Park, Yujin Lim

Open Access

https://doi.org/10.3390/electronics10233015

Copy DOI

Journal: Electronics	Publication Date: Dec 2, 2021
Citations: 9	License type: CC BY 4.0

Affiliation: Sookmyung Women's University

Abstract

With increasing data traffic requirements in vehicular networks, vehicle-to-everything (V2X) communication has become imperative in improving road safety to guarantee reliable and low latency services. However, V2X communication is highly affected by interference when changing channel states in a high mobility environment in vehicular networks. For optimal interference management in high mobility environments, it is necessary to apply deep reinforcement learning (DRL) to allocate communication resources. In addition, to improve system capacity and reduce system energy consumption from the traffic overheads of periodic messages, a vehicle clustering technique is required. In this paper, a DRL based resource allocation method is proposed with remote radio head grouping and vehicle clustering to maximize system energy efficiency while considering quality of service and reliability. The proposed algorithm is compared with three existing algorithms in terms of performance through simulations, in each case outperforming the existing algorithms in terms of average signal to interference noise ratio, achievable data rate, and system energy efficiency.

Highlights

With increasing data traffic requirements in vehicular networks, vehicle-to-everything (V2X) communication has become imperative in improving road safety to guarantee reliable and low latency services
The baseband unit (BBU) that process signals are centralized in a BBU pool and remote radio head (RRH) that serve as radio-frequency transceivers are distributed in the cells
We present a deep reinforcement learning (DRL) based resource allocation algorithm with RRH grouping and vehicle clustering for cellular V2X communication while guaranteeing quality of service (QoS) and reliability requirements

Summary

Introduction

With increasing data traffic requirements in vehicular networks, vehicle-to-everything (V2X) communication has become imperative in improving road safety to guarantee reliable and low latency services. A DRL based resource allocation method is proposed with remote radio head grouping and vehicle clustering to maximize system energy efficiency while considering quality of service and reliability. With the increasing data traffic in wireless and vehicular networks, it has become difficult to meet data traffic requirements To this end, cellular network architectures have been researched as a potential solution. The BBUs that process signals are centralized in a BBU pool and RRHs that serve as radio-frequency transceivers are distributed in the cells. It causes serious interference because of the dense deployment.

Objectives

Results

Conclusion