NON-COOPERATIVE CROSS-CHANNEL GAIN ESTIMATION USING FULL-DUPLEX
AMPLIFY-AND-FORWARD RELAYING IN COGNITIVE RADIO NETWORKS
Bijia Huang
†
, Guodong Zhao
†
, Liying Li
‡
, Xiangwei Zhou
[
, and Zhi Chen
†
†
National Key Laboratory of Science and Technology on Communications, UESTC, Chengdu, China
‡
Automation Engineering, UESTC, Chengdu, China
[
School of EECS, Louisiana State University, Baton Rouge, LA, USA
ABSTRACT
In this paper, we propose a new estimation method to obtain the
cross-channel gain, which avoids the severe interference to the pri-
mary receiver (PR) in existing relay-assisted estimation methods. In
our method, we let the cognitive transmitter add a time delay when it
conducts the full-duplex amplify-and-forward relaying. This forces
the time-difference-of-arrival (TDOA) between the direct and relay
signals to be large enough rather than randomly large or small. Then
we develop our estimation method only in the large TDOA case and
precisely control the interference to the PR. Simulation results indi-
cate that the proposed method can significantly reduce the interfer-
ence to the PR.
1. INTRODUCTION
In cognitive radio networks, it is very challenging to estimate the
cross-channel gain between the cognitive transmitter (CT) and the
primary receiver (PR) [1–3] in frequency-division duplexing (FDD)
systems. This is because the cross-channel gain can only be esti-
mated at the PR and sent back to the CT via the backhaul link be-
tween cognitive and primary systems. However, such an assumption
is not always valid.
Recently, a non-cooperative method, called proactive estima-
tion, has been introduced into cognitive radio networks [4–8], which
enables the CT to autonomously estimate the cross-channel gain in
FDD systems. However, this method may cause severe interference
to the PR since it requires the CT to send jamming signal to the PR.
Very recently, a new kind of proactive estimation methods, called
relay-assisted proactive estimation, has been proposed in [9–11] to
deal with the interference issue. Instead of sending jamming sig-
nal to the PR, the CT conducts the full-duplex amplify-and-forward
(AF) relaying to forward the primary signal to the PR. This can ef-
fectively reduce the interference to the PR. However, it requires the
time-difference-of-arrival (TDOA) between the direct and relay sig-
nals to be small enough. In fact, this may not be true since the TDOA
is actually a random variable and it is determined by the locations of
the three nodes [12–14]. For example, if the CT is located in the
segment between the PT and PR, the TDOA is small. However, if
the CT is far away from the PT and PR, the TDOA becomes large.
Therefore, the existing relay-assisted methods designed under the
small TDOA case may still cause severe interference to the PR since
the TDOA may be randomly small and large in practical systems.
This work was supported in part by the National Natural Science Foun-
dation of China under Grant 61201130, by the U.S. National Science Foun-
dation under Grants 1343380 and 1560437, and by the High-Tech Research
and Development (863) Program of China under Grant 2015AA01A707. The
corresponding author is Guodong Zhao (gdngzhao@gmail.com).
Fig. 1. System model.
In this paper, we investigate the relay-assisted cross-channel
gain estimation in FDD systems and design a new estimation method
to obtain the cross-channel gain, in which a preset interference prob-
ability to the PR can be satisfied. In our method, we let the CT
artificially add a time delay during the full-duplex AF relay. This
forces the small TDOA to be the large one no matter where the three
nodes are located. Then we can design our estimation method only
in the large TDOA case, which allows us to precisely control the
interference to the PR.
2. SYSTEM MODEL
Fig. 1 provides the system model of this paper. In the figure, the PT
serves the PR that is uniformly located inside the disk region with
the center PT and the radius R. At the same time, a CT intends to
estimate the cross-channel gain from the CT to the PR for spectrum
sharing, and the distance between the CT and the PT is r. Here, we
denote h
k
√
g
k
(k = 0, 1, and 2) as the channel coefficients among
the three nodes, where h
k
and g
k
represent the small-scale fading
and large-scale path loss coefficients, respectively. In the small-scale
fading, the coefficient h
k
follows Rayleigh distribution with unit
variance. In the large-scale
1
path loss [15], the coefficient g
k
can
be obtained by g
k
(dB) = −128.1−37.6 log
10
(l) for l ≥ 0.035 km,
where l denotes the distance between two nodes. Furthermore, we
consider the block fading channel, where the Rayleigh fading coef-
ficients are constant within each block and they are independent for
different blocks.
1
To facilitate the development of the proposed method, we do not con-
sider the shadowing in our model, which allows us to obtain some important
closed-form expressions. However, in the simulation, we consider the shad-
owing to evaluate the performance of our method.