IEEE COMMUNICATIONS LETTERS, VOL. 17, NO. 3, MARCH 2013 487
Outage Probability of TDBC Protocol in
Multiuser Two-Way Relay Systems with Nakagami-
m Fading
Xianfu Lei, Lisheng Fan, Diomidis S. Michalopoulos, Pingzhi Fan, and Rose Qingyang Hu
Abstract—This paper considers a multiuser two-way relay sys-
tem with time division broadcast (TDBC) protocol, where multiple
end-sources compete to exchange information with another end-
source through the help of a single half-duplex amplify-and-
forward (AF) relay. We derive a tight closed-form lower bound for
the system outage probability over Nakagami-m fading channels
with integer fading parameters. An asymptotic expression for the
outage probability in high signal-to-noise ratio (SNR) regime is also
obtained. Numerical results corroborate the theoretical analysis.
Index Terms—Outage probability, two-way relay system, time
division broadcast (TDBC), Nakagami-m fading.
I. INTRODUCTION
T
WO-WAY relaying has received many attentions because
of its relatively high spectrum efficiency. There are two
major spectrally efficient protocols proposed for two-way relay
systems, i.e., analogy network coding (ANC) and time division
broadcast (TDBC). The two-way relay systems employing ANC
and TDBC protocols have been studied in the literature [1]–
[7]. Sp ecifically, the outage probability of ANC and TDBC
protocols was studied for a three-node system in [1] and
[2], respectively. The outage probability of ANC and TDBC
protocols with optimal relay selection was investigated in [3] for
multi-relay systems. The outage performance of ANC protocol
over Nakagami-m fading channels has been studied for multi-
relay system in [4]. The authors in [5] analyzed the outage
proba bility of ANC proto col in a multiuser two-way relay
system with optimal source scheduling over mixed Rayleigh
and Rician fading channels. For the same scheme, the average
sum rate was derived in [6], when operating over Rayleigh
fading. The authors in [7] studied the outage probability of
TDBC protocol in a multiuser two-way relay system in Rayleigh
fading, where the source scheduling is only based on the direct
links. However, the performance of TDBC protocol has not been
investigated in multiuser two-way relay systems with optimal
source scheduling, even in a Rayleigh fading environmen t.
Manuscript received October 20, 2012. The associate editor coordinating the
review of this letter and approving it for publication was M. Xiao.
This work was supported by the National Basic Research
Program of China (973 Program No. 2012CB316100), NSFC
(No. 60971116/61032002/61002015), the Open Research Fund of State
Key Laboratory of ISN (No. ISN13-04), the NSF of Guangdong Province (No.
S2012010010062), and the 111 project (No. 111-2-14).
X. Lei is with Utah State Uni versity, Logan, Utah, USA. He was with South-
west Jiaotong Uni versity, Chengdu, China (e-mail: xflei81@yahoo.com.cn).
P. Fan is with Southwest Jiaotong University , China (e-mail: p.fan@ieee.org).
L. Fan is with Shantou University, Shantou, China (e-mail: lsfan@stu.edu.cn).
X. Lei and L. Fan are also with the State Key Laboratory of Integrated
Services Networks, Xidian University, Xi’an, China.
D. S. Michalopoulos is with the Unive rsity of British Columbia, Vancouver,
Canada (e-mail: dio@ece.ubc.ca).
R. Q. Hu is with Utah State University, Logan, Utah, USA (e-mail:
rosehu@ieee.org).
Digital Object Identifier 10.1109/LCOMM.2013.020413.122338
A
B
1
B
k
B
K
R
3rd phase
2nd phase
1st phase
Fig. 1. System model.
In this paper, we study the outage probability of the TDBC
protocol in a multiuser two-way relay system with optimal
source scheduling, operating over Nakagami-m fading channels
with integer fading parameters. A tight closed-form lower bound
for the outage probability of the scheme under consideration is
derived. The asy mptotic outage behavior is then investigated
in high signal-to-noise ratio (SNR) regime, providing useful
insights into the system’s diversity order.
Notation: We use CN(0,σ
2
) to denote a circularly symmetric
complex-valued Gaussian random variable with mean 0 and
variance σ
2
. The notation X ∼ Naka(m, α) denotes that the
random variable X follows the Nakagami-m distribution so
that the probability density function (PDF) of Y = |X|
2
is
f
Y
(y)=
m
m
y
m−1
α
m
Γ(m)
e
−
my
α
. Pr[·] denotes p robability. Moreover,
Γ(x) and Γ(α, x) are the Gamma function [8, eq. (6.1.1)] and
incomplete Gamma function [8, eq. (6.5.3)], respectively.
II. S
YSTEM MODEL
The scheme under investigation is depicted in Fig. 1. We
consider a m ultiuser two-way relay system with on e amplify-
and-forward (AF) relay R, in which end-source A intends to
exchange information with one out of other K end-sources B
k
(k =1, 2, ···,K). We use h
B
k
,R
∼ Naka(m
1
,α), h
A,B
k
∼
Naka(m
2
,β) and h
A,R
∼ Nak a(m
3
,ε) to denote the channels
of B
k
-R, A-B
k
and A-R, respectively. All the channels are
reciprocal. Moreover, all terminals have a single antenna and
operate in the half-duplex mode. For each transmission, one
out of K end-sources B
k
is selected to exchange information
with end-source A. Before presenting the selection criteria in
(9), we first give the details on how the TDBC protocol works,
as follows. Assuming B
k
is selected, terminal A will exchange
information with B
k
during three phases. Specifically, in the first
phase, A sends signal s
A
with unit energy, hence the signals
received at R and B
k
can be expressed as
y
(1)
R
= h
A,R
√
Ps
A
+ n
(1)
R
, (1)
y
(1)
B
k
= h
A,B
k
√
Ps
A
+ n
(1)
B
k
, (2)
where P is the transmit power, n
(1)
R
∼CN(0, 1) and n
(1)
B
k
∼
CN(0, 1) are the additive white gaussian noise (AWGN) at the
relay and B
k
, respectively, in the first phase. In the second
1089-7798/13$31.00
c
2013 IEEE