QoE-Aware Resource Allocation for D2D
Communications in Unlicensed Spectrum
Yu Jin and Shaoyi Xu
School of Electronics and Information Engineering, Beijing Jiaotong University, Beijing 100044, China
E-mail: {15125014, shyxu}@bjtu.edu.cn
Abstract—This paper studies the QoE-aware resource alloca-
tion for device-to-device communications in unlicensed spectrum
(D2D-U), which can meet the explosively growing data traffic and
provide higher user’s satisfaction in 5G networks. Specifically,
channel access and power allocation are jointly considered for
the D2D-U system and an optimization problem is formulated
with the objective of balancing the overall system throughput
and fairness of D2D-U and WiFi for coexistence. In order to
reduce the complexity, this problem is decomposed into two
subproblems, accordingly, two algorithms are proposed which are
QoE-aware power allocation algorithm and rate-based fair duty
cycle algorithm (F-DCA). In the QoE-aware power allocation
algorithm, we employ the difference convex (D.C.) method to
transfer the non-convex problem into convex optimization, which
can greatly reduce the complexity. In the F-DCA, we adopt an
iteration method to allocate the appropriate time proportion for
the D2D-U system adaptively according to the load of the WiFi
network. Simulation results show that our proposed algorithms
achieve better performance compared with the traditional meth-
ods.
I. INTRODUCTION
The fifth generation (5G) wireless network raises two
most obvious requirements which are better user experience
and higher system capacity [1]. With the development of
network technology, the differences among services tend to
be assimilated. In the face of homogenous services, users’
demands for service quality are more and more harsh, which
requires operators to pay more attention to user experience.
Currently, the most widely used standard for evaluating service
quality is called quality of service (QoS), which only reflects
the performance from the physical layer including network
throughput, delay, packet loss rate, jitter and so on. QoS can
not directly reflect the degree of user acceptance of the service,
if the service can not guarantee the user experience, the user
will be lost and the revenue of the operator will be affected.
Therefore, to better achieve the requirement of user-centric
in 5G system, the quality of experience (QoE) is proposed
instead of QoS to measure system performance [2]. QoE is
defined as the user’s satisfaction with the service, it is a
combination of subjective and objective factors. At the same
time, there are three mainstream technologies can meet the
demand for higher system capacity: Ultra-dense network [3],
expanding the network in unlicensed bands [4] and innovation
of techniques such as device-to-device(D2D) [5].
D2D communication allows two users in proximity com-
municate directly rather than go through the base station and
can obtain proximity gain, reuse gain, and hop gain without
having to invest in network-side hardware upgrades or new cell
deployments. However, the lack of licensed spectrum hinders
the deployment of D2D communication. Hence, the device-to-
device communication in the unlicensed spectrum (D2D-U) is
becoming a promising technology. In D2D-U networks, D2D-
U users share the unlicensed spectrum with WiFi users and
how to ensure the harmonious coexistence between D2D-U
users and existing WiFi users is one of the critical challenges.
In [6], the authors show that by allowing D2D to operate
in the unlicensed bands with protective fairness measures
for WiFi transmissions, D2D is able to operate across the
whole coverage area and efficiently scale the overall network
capacity. [7] presented a joint mode selection and resource
allocation algorithm to minimize the overall interference that
cellular and WiFi users suffer from D2D communications.
[8] designed a feasible sensing-based D2D-U protocol for the
D2D-U users to utilize the unlicensed spectrum.
Nevertheless, existing works still have following problems.
Firstly, these works mainly concentrate on improving the
QoS of system rather than QoE, thus the user experience is
ignored. Secondly, these algorithms are not necessarily global
optimal leading to the optimal solution can not be obtained.
Finally, they can not provide higher fairness between D2D-
U and WiFi while maintaining high total system throughput.
Motivated by the above problems, in our paper, we formulate
an optimization problem with the objective of balancing the
overall system throughput and fairness of D2D-U and WiFi for
coexistence, then we propose a QoE-aware power allocation
algorithm for the multiple D2D-U pairs sharing the same
unlicensed channel and devise a rate-based fair duty cycle
algorithm (F-DCA) to allocate time proportion for D2D-U
system dynamically.
Our major contributions in this paper are summarized as
follows.
1) A QoE-aware power allocation algorithm is proposed for
multiple D2D-U pairs sharing same unlicensed bands. Instead
of using traditional QoS to constrain the utility function, in
our work, QoE is applied to be the constraints.
2) Aim to balance the system throughput and fairness, the
original problem is formulated as non-convex optimization
problem. To solve such a problem, the D.C. method is pro-
posed to obtain the global optimal solution.
3) A rate-based fair duty cycle algorithm by allocating time
proportion for D2D-U system dynamically is devised, which
can provide higher fairness between D2D-U and WiFi while
978-1-5386-4328-0/18/$31.00 ©2018 IEEE