Page 157 - Proceedings of the 2017 ITU Kaleidoscope

SECRECY ENERGY EFFICIENCY OPTIMIZATION FOR ARTIFICIAL NOISE AIDED
PHYSICAL-LAYER SECURITY IN COGNITIVE RADIO NETWORKS

Yuhan Jiang, Jian Ouyang, and Yulong Zou*

School of Telecomm. and Inform. Eng., Nanjing Univ. Posts & Telecomm., Nanjing, P. R. China
*Corresponding author (email: yulong.zou@njupt.edu.cn)

ABSTRACT of total power and sum rate in amplify-and-forward-based
energy harvesting two-way relaying systems.
In this paper, the artificial noise (AN) is used to improve the
Generally, the aforementioned work is either the case only
secrecy energy efficiency (SEE) for underlay cognitive ra-
concerned about secrecy rate (SR) or the case focused on
dio networks (CRNs). A joint zero-forcing (ZF) beamform-
EE. In order to achieve a better trade-off between the SR and
ing and power allocation problem is formulated to maximize
EE, the secrecy energy efficiency (SEE), which is defined as
the SEE under the constraints of the total transmit power, the
the ratio of the SR to the total power consumption, was in-
secrecy rate (SR) of cognitive user (CU) and the quality-of-
troduced in [12]. Motivated by this, this paper presents an
service (QoS) requirement of primary user (PU). As a conse-
SEE maximization (SEEM) scheme with the constraints of a
quence, we firstly transform the formulated non-convex opti-
total transmit power, the SR of cognitive user (CU) and the
mization problem in fractional form into an equivalent one in
quality-of-service (QoS) requirement of primary user (PU).
subtractive form, and use the difference of two-convex func-
A two-tier iterative power allocation algorithm is adopted to
tions (D.C.) approximation method to obtain an equivalent
solve our formulated SEEM problem, where the outer iter-
convex problem. Then, a power allocation algorithm is p-
ative mainly focuses on converting the formulated fractional
resented to obtain the optimal solution. Simulation results
problem into a subtractive one which is then transformed into
show the advantage of the proposed scheme.
a convex problem with the help of difference of two-convex
Keywords— Power allocation, artificial noise, energy effi- functions (D.C.) approximation method in the inner iterative.
ciency, secure communication, cognitive radio networks Finally, the advantage of the proposed SEEM scheme is illus-
trated by numerical results.
The reminder of this paper is organized as follows. In Section
1. INTRODUCTION II, we present the system model of secure communications in
CRNs and formulate an SEEM problem to maximize the SEE
Recently, physical-layer security (PLS) [1] has attracted in- of cognitive transmissions. Next, Section III presents an opti-
creasing attention, especially in cellular cognitive radio net- mal solution to the formulated SEEM problem as well as the
works [2]. Under this situation, extensive research efforts corresponding iterative algorithm, followed by Section IV,
have been devoted to improving the PLS of wireless com- where simulation results are given to show the advantage of
munication by employing beamforming [3], artificial noise proposed SEEM scheme. Finally, some concluding remarks
(AN) [4], cooperative jammers [5], and so on. For example, are provided in Section V.
a scheme by jointly utilizing multiple antennas and coopera-
tive jammers was investigated in [6] to maximize the secrecy
rate. In [7], the authors injected the AN at the transmitter 2. SYSTEM MODEL AND PROBLEM
to degrade eavesdropping channel and exploited spatial de- FORMULATION
grees of freedom in massive multiple-input multiple-output
(MIMO) systems to enhance the PLS. As shown in Fig. 1, we consider an underly cognitive radio
Also, there is an increasing interest in energy efficiency (EE), network (CRN) where a cognitive base station (CBS) with N
which is regarded as an effective metric to balance the spec- antennas transmits confidential message to a single-antenna
tral efficiency (SE) and power consumption especially in the CU in the presence of an eavesdropper (ED) with a single
condition of the tremendous demand of high SE and low antenna. The CBS uses the licensed spectrum that is unoccu-
power consumption [8]. In [9], a resource allocation algorith- pied by a PU to transmit confidential message. To improve
m for EE communication was studied in an orthogonal fre- the PLS of cognitive transmissions, AN signals are used to
quency division multiple access (OFDMA) system. An up- confuse the ED.
link energy minimization problem was investigated with non- Thus, the transmit signal of CBS can be expressed as x =
orthogonal multiple access for machine-to-machine commu- v c x c + v z z, where x c is the confidential signal, satisfying
2 2
nications in [10]. Besides, the authors of [11] studied a power E(|x c | ) = P c , z is the AN signal with E(|z| ) = P z , v c
allocation problem to maximize the EE under the constraints and v z are the beamforming (BF) weight vectors of the con-

978-92-61-24291-6/CFP1768P-ART © 2017 ITU – 141 – Kaleidoscope