QoE-enabled unlicensed spectrum sharing in 5G : a game-theoretic approach

Spectrum sharing is an important aspect of 5G new radio, as it plays a complementary role for fulfilling diversified service requirements. This paper studies unlicensed spectrum sharing, namely, local thermal equilibrium (LTE) over unlicensed bands (LTE-U), for providing a better quality of experien...

Full description

Saved in:
Bibliographic Details
Main Authors: Bairagi, Anupam Kumar, Abedin, Sarder Fakhrul, Tran, Nguyen H., Niyato, Dusit, Hong, Choong Seon
Other Authors: School of Computer Science and Engineering
Format: Article
Language:English
Published: 2018
Subjects:
Online Access:https://hdl.handle.net/10356/89794
http://hdl.handle.net/10220/46394
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Spectrum sharing is an important aspect of 5G new radio, as it plays a complementary role for fulfilling diversified service requirements. This paper studies unlicensed spectrum sharing, namely, local thermal equilibrium (LTE) over unlicensed bands (LTE-U), for providing a better quality of experience (QoE) in 5G networks. Specifically, unlicensed band selection and resource allocation (time, licensed, and unlicensed) are jointly designed, and an optimization problem is formulated with the objective of maximizing LTE users' QoE [measured in mean opinion score (MOS)] while protecting incumbent wireless systems such as Wi-Fi in the unlicensed spectrum. To solve the multi-player interaction in this spectrum space fairly, we employ a game-theoretic approach. A virtual coalition formation game (VCFG) is used to solve the unlicensed band selection problem. The outcome of the VCFG defines the optimization problem within each coalition. This optimization problem is then decomposed into two sub-problems: 1) time-sharing problem between the LTE-U and Wi-Fi systems and 2) a resource allocation problem for the LTE-U system. The cooperative Kalai-Smorodinsky bargaining solution is used for solving the first sub-problem, whereas the Q-learning algorithm is used for solving the second. VCFG and Q-learning-based resource allocation algorithms are proposed in this paper. In addition, the stability of VCFG and optimal sharing time are also proved in this paper. Simulation results show the advantages of the proposed approach over other baseline methods in terms of the MOS, percentage of unsatisfied users, and fairness. The results also show that the proposed approach can better protect the performance of Wi-Fi users compared to the conventional listen-before-talk scheme.