Content uploaded by Imran Rashid
Author content
All content in this area was uploaded by Imran Rashid on Jun 25, 2020
Content may be subject to copyright.
Primary User Emulation Attack and
their Mitigation Strategies: A Survey
Bilal Naqvi‡, Imran Rashid‡, Faisal Riaz†, Baber Aslam‡,
‡National University of Sciences and Technology, Islamabad, Pakistan
†Mirpur University of Science & Technology, Mirpur (A.K), Pakistan
‡{bilal.is-10, irashid, ababer}@mcs.edu.pk , †faisal.f.r@ieee.org
Abstract— Cognitive Radio emerged as a solution to solve
the spectrum scarcity by allowing unlicensed user to exist with
the licensed users of the network whenever the licensed users
are in idle state. Cognitive radio since its inception has received
a lot of importance as far as research is concerned, however
security threats also emerged with this promising technology
among which Primary User Emulation (PUE) attack is one. We
cover an understanding of the PUE attack in this paper and
then discuss existing solutions to mitigate it along with the
limitations of the proposed solutions. In end we identify the
need of solution to address a case that has not been addressed
yet.
Keywords— Cognitive Radio Network, Primary User
Emulation Attack, Sensing, Opportunistic access
I. INTRODUCTION
The ever increasing demand of spectrum utilization due
to rapidly growing number of wireless users has led to
spectrum shortage problem, which is expected to further
increase in the coming years. To counter this situation
Federal Communication Commission (FCC), in September
2010, approved new rules to better manage and address the
spectrum shortage problem [1]. The technology that
addresses this spectrum shortage is known as Cognitive
Radio. It solves the spectrum shortage problem by letting
unlicensed user to consume the spectral resources when the
licensed user (also known as Primary User, PU) is not active.
The users that use the resources when spectrum allocated to
PU is vacant are called the Secondary Users (SU) [2]. The
SU can only use the resources till the time the PU is not
using the allocated resources; as soon as the PU becomes
active, the SU has to leave the channel for use by the PU.
The cognitive radio, therefore, provides opportunistic access
to the SU whenever the channel is vacant [18].
The SU has to perform sensing to detect presence of primary
user (PU) and only if the channel is vacant (i.e. the primary
user is not active), then the SU will be able to use the
channel [3]. To ensure proper advantages of Cognitive
Radio, i.e., to provide spectral resources to unlicensed users;
two issues are of prime importance: (1) to find available slots
(called whitespaces) in the bandwidth for use by the SU, (2)
to have a mechanism that ensures no interference with the
PU. To solve the first issue, out-band sensing is performed to
find out white spaces in the network bandwidth. Based on
the results of out-band sensing, the SU selects whitespace
according to its Quality of Service (QoS) requirements. After
selecting a white space the SU starts using it to communicate
with desired SU. Once the transmission starts the SU enters
into periodic transmission and sensing cycle. Each sensing
interval is followed by a transmission interval the SU
transmits data to communicate with other SU. During
sensing interval referred to as in-band sensing the SU
ensures that the PU, whose whitespace is underutilization, is
not active. This also ensures noninterference with the PU,
whose whitespace is underutilization. The SU has to vacate
the channel within a certain amount of time after the re-
emergence of the PU. The upcoming IEEE 802.22 standard,
governing use of Cognitive Radio over the unused bandwidth
of TV channels, states that the SU has to vacate the channel
within two seconds of detection of PU via in-band sensing
[3]. After performing in-band sensing the SU transmits for a
short time then does in-band sensing again, this process
continues till the time PU is active or when the transmission
of SU is over and does not require resources any more.
Due to the role of cognitive radio in solving spectrum
shortage problem and its increasing use in many
applications, it is susceptible to many attacks including the
primary user emulation (PUE) attack, hello flood attack,
sinkhole attack, jamming attack, lion attack etc. [1]. Of these
attacks, the attack under consideration is the PUE attack that
if successful forms a basis of other attacks like a Lion attack
or in extreme form it can result into a Denial of Service
(DoS) attack. PUE attack is done by a malicious SU during
the sensing time masquerading as a PU to obtain the network
resources, thus denying other SUs to use the resources. There
are two types of this attack: first is the Selfish PUE attack in
which the focus of the attacker is to increase its own share in
the spectral resources. The second is the Malicious PUE
attack in which the attacker prevents legitimate SUs from
using the resources [1], [20].
The rest of paper is organized as follows. In section II we
cover similar/related work on the same topic. In section III
we discuss PUE attack in detail. In section IV we cover
literature survey. In section V we cover critical analysis of
existing solutions. In section VI we conclude the paper and
present our future works.
II. RELATED WORK
Many surveys conducted on the PUE attacks, are a part of
today’s literature content. Some of the surveys specifically
2013 2nd National Conference on Information Assurance (NCIA)
978-1-4799-1288-9/13/$31.00 ©2013 IEEE
95
Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on June 25,2020 at 07:17:42 UTC from IEEE Xplore. Restrictions apply.
[1] and [20] have been cited by us in our paper. However [1]
is basically a survey on security threats to Cognitive Radio
and is not focused specifically towards PUE attacks. The
authors discussed many solutions to counter PUE attacks but
focus was not on identifying limitations of those solutions.
The authors in [20] focused primarily on PUE attack but they
did not discuss limitations of the proposed solutions.
This paper covers an understanding of PUE attack in
depth than most of available content. Simple scenarios have
been considered as in Section II to give reader a detailed
understanding of the PUE attack. Also we have tried to
group the solutions according to their types, that is another
unique feature of this paper. In the end we summarize the
paper by identifying the limitations of each proposed
solution, which is not found in most of the surveys available.
III. PRIMARY USER EMULATION (PUE) ATTACK
In this section we discuss the PUE attack in detail by
considering a simple scenario. We assume a cell consisting
of one primary user (PU) and three secondary users (SU). As
discussed above once the SU starts utilizing a bandwidth it
enters into cycle of sensing and transmitting till the time the
SU needs the bandwidth. This series of sensing and
transmission cycles one after the other continues with time
and transmission can only be done once the SU determines
the availability of network resource as a result of sensing.
In figure 1 the activity of SU is represented showing
sensing and transmission slots. Each SU first senses for a slot
in the bandwidth and then based on sensing results transmits
during the transmission slots and then senses again to check
for reappearance of the SU.
Fig.1. Representation of usual activity of SU depicting sensing and
transmission slots.
The figure 1 showed the sensing and transmission activity
of SU in our simple scenario. In the figure 2, it can be seen
that there is one PU who is in inactive state initially and
during most of time in our consideration. When the PU is
inactive, SU1 appears and performs sensing to check for
availability of bandwidth. Upon sensing SU1 finds network
to be available and transmits. Once SU1 is transmitting SU2
appears and senses for availability of networks and finds
network to be unavailable therefore enters a timed wait
interval. In the meantime SU1 finishes its transmission and
leaves the network intentionally. SU2 senses again and finds
the network available and then enter transmission phase.
Upon finishing its transmission SU2 leaves. Then it can be
seen that SU3 appears and senses for network availability and
it finds the network available it enter into transmission phase.
Upon completion of first transmission slot SU3 has some
more data to send therefore it senses again to check for
reappearance of the PU and based on sensing SU3 finds PU
to be active therefore SU3 has to leave the network as the PU
has high priority and SU can only use the spectral resources
till the time PU is inactive.
Fig.2. Representation of ideal case consisting of a PU and three SUs
Now we consider an attack scenario where malicious SU
represented as SUm emulates itself as a PU and makes the
network unavailable for others SUs. In the figure 3 below it
can be that SU1 appears and performs sensing. Based on the
sensing results it transmits and then leaves the network.
While SU1 was transmitting, SU2 appeared and performed
sensing and entered into wait state. Once SU1 leaves the
network, a malicious SU emulates the characteristics of the
PU and pretends to be PU. SU2 senses again which is at point
in time when SU1 is no more using the network but a
malicious SU is pretending to be PU thus making SU2 to
enter in other waiting interval. Ultimately SU2 leaves the
network without getting a chance to transmit because the
network was made unavailable by phenomena that is called
the Primary User Emulation Attack.
Similarly SU3 appears and senses while emulation by
malicious SU is under way. SU3 senses for network
availability but finds it to be unavailable because of
emulation attack. The same is depicted by the figure 3.
Fig.3. Representation of attack scenario by malicious SU making network
resources unavailable.
96
Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on June 25,2020 at 07:17:42 UTC from IEEE Xplore. Restrictions apply.
If the PUE attack is successful then cognitive radio
technology is unable to deliver the purpose it has been
devised for i.e. providing access to unlicensed users
whenever the spectrum is vacant. However situation can get
more worst if malicious SU schedules PUE during each
sensing interval ultimately causing a denial of service (DoS)
attack.
To elaborate the attack in more detail we use a flow
diagram as in figure 4. It can be seen in figure 4 that a SU is
sensing whether the PU is active or not. Initially the SU
finds PU to be inactive and therefore starts transmission.
After transmission cycle the SU checks whether or not the
PU is active again, so the SU performs sensing and while
sensing it finds PU active because of emulation by the
malicious SU. The result is spectrum mobility i.e the SU has
to vacate the spectrum resource currently underutilization.
Fig.4. Flow Diagram depicting scenario of successful Primary User
Emulation Attack
IV. LITERATURE SURVEY
Since the inception of cognitive radio and its increasing
use, a lot of research has been carried out towards better
sensing techniques, coexistence mechanism etc. However
security issues have received a little less attention. Based on
our review of available literature the proposed solutions to
PUE attack fall in three categories as depicted by the figure
below.
Fig.5. A taxonomy representing kinds of solutions to PUE
Generally describing each type, distributed/ individual
node based schemes emphasize on applying the proposed
technique on individual user and guarding a particular user
from the attack. Whereas in centralized/ cooperation based
schemes the emphasis is on cooperation and communicating
with central authority upon discovery of unusual activity.
The main difference between the two is that of decision
authority. In distributed the decision making authority rests
with the individual user while in centralized scheme the
decision making rest with central authority. Literature survey
helped in establishing that Centralized/Cooperation based
schemes are particularly used in MANET and VANET
applications. Now we consider the proposed solutions one by
one categorizing it according to taxonomy in figure 5.
A. Distributed/ Individual node based solutions
In this section we will discuss the schemes that do not
require cooperation among the nodes in the Cognitive radio
network.
1) Practical Solutions: Chen et al., in [4] proposed the
transmitter verification scheme called LocDef [4]. This
scheme detects PUE attack based on RSS (received signal
strength) value. This scheme is particularly applicable in TV
networks where signals are generated by powerful base
station and such power is difficult to emulate by malicious
SU. In addition to the Distance Ratio Test (DRT), this
scheme takes advantage of the geographical location of a
base station using distance difference test (DDT). Since the
location of the base station is fixed, so whenever SU detects
primary user as active, it estimates the location from where
the received signal was generated. If the location from where
the received signal was generated matches that of the base
station it is assumed to be a legitimate primary user
otherwise an attack is detected [19].
Khare et al., in [5] studied the threats to cognitive radio
networks and suggested that the mitigation strategies in [4]
are applicable in detecting PUE attacks. The authors referred
to the proposed techniques, distance difference test (DDT)
and the distance ratio test (DRT). Both these techniques are
used for location verification of the received signal.
Employing the above techniques it can also be ascertained
whether the received signal strength is equal to the strength
generated by legitimate PU [4] or not. In addition the authors
in [5] proposed a Network User Management Center
(NUMC) based cognitive radio architecture where the role of
NUMC is the authorization of all the secondary users SUs.
Huang et al., in [6] proposed a strategy similar to [4],
featuring localization, but due to some shortcoming in DDT
and DRT, proposed an efficient positioning scheme based on
Time Difference of arrival (TDOA) and Frequency
Difference of arrival (FDOA).
Zhao et al., in [7] proposed a scheme to counter PUE
attacks that is based on transmitter fingerprinting. In this
technique the phase noise of the noisy career is extracted.
Categories of
solution to PUE
attack
Distributed/
Individual node
based
Centralized/
Cooperation based
Practical
Solutions
Analytical
Model
Intrusion
Detection System
Based
97
Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on June 25,2020 at 07:17:42 UTC from IEEE Xplore. Restrictions apply.
The extracted sample is then applied to identify the
transmitter.
2) Analytical Approaches: Jin et al., in [8] proposed an
analytical approach to detect PUE attack based on Neyman-
Pearson composite hypothesis test and Wald’s sequential
probability ratio test. In their approach firstly probability
density function is generated for the received signal. Based
on the probability density function result, Neyman-Pearson
composite hypothesis test is conducted which under certain
scenarios can have high probability of a false alarm.
Therefore Wald’s sequential probability ratio test is applied
which gives improvement in results by allowing user to
specify thresholds for false alarms and probability of a miss.
In another paper Jin et al., [9] proposed an analytical
scheme independent of the location information and does not
require any dedicated sensors. They present their analysis
using Fenton’s approximation and Wald’s sequential
probability ratio test. As discussed above using Wald’s
sequential probability ratio test allows user to set threshold
values on probability of successful attack and probability of
a miss thus giving improved results.
Anand et al., in [10] proposed an analytical approach
based on Fenton’s approximation and Markov inequality.
The proposed scheme uses received power at a SU and treats
it as log-normally distributed random variable. Fenton’s
approximation is then applied to determine the mean and
variance of power received at SU. The lower bound on
probability of the case of successful attack is then
determined using Markov inequality.
Clancy et al., in [11], in their study of threats to cognitive
radio networks suggest that development of better sensing
algorithms with low false positive rates can be helpful in
mitigating PUE attack.
B. Centralized/Cooperation based solutions
Ramzi Saifan, in [12] proposed a new scheme for in-band
sensing that can be used for countering PUE attacks. The
author proposed cooperative sensing scheme for his
framework where nodes have two distinct tasks that is
sensing and transmission. Nodes in sensing mode only
perform in-band sensing and send warning messages (about
re-appearance of the PU) to the nodes in transmission mode.
Nodes in transmission mode can only do transmission and
hear to the warning messages generated by the other set of
nodes that are in sensing mode. For PUE attack, he proposed
feature detection technique that is capable of identifying
modulation type of a primary signal [6]. Using this technique
the system will be able to distinguish whether the signal is
generated by primary user or by malicious secondary user.
Kaligineedi et al., in [13] propose their scheme to detect a
malicious user in cognitive radio network. The proposed
scheme requires cooperation among the nodes in the
network. All the nodes send their sensed data to a central
entity which makes the decision about presence or absence of
primary user. In the proposed scheme all the nodes use
energy detectors and each user is assigned outlier factors.
The assigned factors are used to identify and nullify the
effect of a malicious user. Once the nodes send their data to a
central entity (an access point), it makes a decision using
data fusion and detection schemes.
C. IDS Based Solutions
In addition to above the authors in [14], [15] propose an
Intrusion Detection System to identify malicious user in the
network. Identifying malicious user is broad term and one
activity of malicious user can be PUE attack. So IDS based
solutions proposed for identification of malicious user can be
useful for detecting PUE attack. The solutions proposed
employ both distributed and centralized schemes. However
the IDS based schemes have been proposed mostly for
wireless ad hoc networks and MANET.
V. CRITICAL ANALYSIS OF PROPOSED SOLUTIONS
All of the solutions discussed above either follow
physical layer level detection mechanism or analytical
solutions or are IDS based. All the solutions discussed have
one thing in common, i.e. they only provide detection
mechanisms to PUE attacks rather than its countermeasures.
Specifically in [4] and [5] detection is done on the basis of
static location and received signal strength. However this
solution is infeasible in case of ad-hoc networks where the
location of PU is not static, thus reducing the effectiveness of
DDT. These solutions are limited to the case where the
location of the PU is static and hence upon receiving a signal
the SU can determine whether the signal is generated by
legitimate PU or not. However the DDT can be compromised
by transmitting from vicinity/neighborhood of a legitimate
PU [1]. The technique proposed in [6] also focuses on
detection and not on mitigation. In [7] finger printing is done
to verify the transmitter but if signal is degraded by more
than expected noise which discards the signature portion of
the transmitter, and then the signal generated by legitimate
transmitter would be considered as false. In [8], cooperative
scheme was proposed but if node in sensing mode leaves the
network then there would not be any one to transmit warning
to the transmitting nodes and hence creating a chance of
collision with the PU which is against the FCC
recommendation which states that “no modification to the
incumbent system should be required to accommodate
opportunistic use of the spectrum by secondary users” [1].
All the solutions, discussed above, are deployed at the
physical layer such as transmitter finger printing and
received signal strength. Further, these focus on detection
rather than providing countermeasures. Also the proposed
solutions do not incorporate the case where the location of
the PU or SU is mobile. Also all IDS based solutions require
some central authority to manage them and configure the
rules to guard against specific attacks. Configuring such
rules on the run time might not be practical and feasible in
some situations.
98
Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on June 25,2020 at 07:17:42 UTC from IEEE Xplore. Restrictions apply.
VI. CONCLUSION & FUTURE WORK
To protect Cognitive Radio networks many schemes have
been proposed but none of the solutions is complete i.e.,
applicable in all scenarios of PUE attack. All schemes
employ almost same level of detection, like they employ
physical layer detection of the attack and do not incorporate
the case where location of PU is not static and changing at
rapid pace as in VANETs. Also the proposed schemes like
transmitter finger printing can be affected by noise and since
noise varies in different networks therefore we need to define
different threshold values for different networks. Hence the
proposed schemes are limited to the network they have been
proposed for. For Cognitive Radio applications deployed in
ad-hoc networks especially vehicular networks [16],[17], the
countermeasure to PUE attack should be flexible to
incorporate changing network environment and should
neither be too much dependent on physical layer attributes
(due to changing location) nor be an IDS based solution (due
to time scarcity) to deploy such a solution. In a network
environment where the location of PU/SU is not static, there
is not much time to deploy techniques as transmitter
fingerprinting and result in high number of false
positives/negatives. Therefore a solution against PUE attack
is required that incorporates the dynamic nature of the ad-hoc
networks, and gives correct detection results and provides
efficient protection (in case of a successful PUE attack). In
addition all solutions discussed in literature review section
have been summarized in table 1 below.
TABLE I. TABLE REPRESENTING SUMMARY OF ALL DISCUSSED
SOLUTIONS
Solution Protection Mechanism
Suggested
Evaluation
Loc Def
Detects PUE attack
based on RSS
(received signal
strength) value
Applicable mostly in
802.22 networks not in
ad-hoc networks
Network User
Management Center
(NUMC) based
cognitive radio
architecture
Employs distance
difference test (DDT)
and the distance ratio
test (DRT).
Applicable in the case
where location of PU/
SU is static
Time Difference of
arrival (TDOA) and
Frequency Difference
of arrival (FDOA).
Employs Time
Difference of arrival
(TDOA) and
Frequency Difference
of arrival (FDOA) for
location verification
Applicable in the case
where location of PU/
SU is static
Neyman-Pearson
composite hypothesis
test and Wald’s
sequential probability
ratio test
An analytical approach
employing Neyman-
Pearson composite
hypothesis test &
Wald’s sequential
probability ratio test is
applied which gives
improvement in results
by allowing user to
specify thresholds for
Just an Analytical
approach. Neyman-
Pearson composite
hypothesis test has
high probability of
false alarm and adding
Wald’s sequential
probability ratio test
adds complexity to the
solution.
false alarms and
probability of a miss
Fenton’s
approximation and
Wald’s sequential
probability ratio test.
An approach
independent of sensor
information and
employs Fenton’s
approximation and
Wald’s sequential
probability ratio test
for detection of the
attack
Just an Analytical
approach. Complexity
high to make it
practically applicable
in especially ad-hoc
networks.
Fenton’s
approximation and
Markov inequality.
The proposed scheme
uses received power at
a SU and applies
Fenton’s
approximation to
determine the mean
value. Both these
values are used to
determine lower bound
on probability of
successful PUE attack
using Markov
inequality
Just an Analytical
approach and based on
received power at the
SU. Received power
alone is not a perfect
metric to make a
decision regarding
occurrence of an
attack.
Feature detection
technique
Feature detection
technique that is
capable of identifying
modulation type of a
primary signal .Using
this technique the
system will be able to
distinguish whether the
signal is generated by
primary user or by
malicious secondary
user
Identifies the PUE
attack only does not
provide any counter
measure.
Centralized/ access
point based scheme
All the nodes send
their sensing data to an
access point which
makes the decision
about presence or
absence of primary
user.
Not applicable in
rapidly changing
networks like VANET.
Intrusion Detection
System to identify
malicious user in the
network
Detects an usual
happening and
generates alarm to
admin.
Time consuming and
need continuous
monitoring. Applicable
mostly in MANET.
Transmitter
fingerprinting
The phase noise of the
noisy career is
extracted and directly
applied to identify the
transmitter
Noise attenuation can
seriously degrade its
performance.
REFERENCES
[1] Wassim El-Hajj, Haidar Safa , Mohsen Guizani, “Survey
of Security Issues in Cognitive Radio Networks” in Journal of
Internet Technology Volume 12, No 2 ,2011.
[2] IEEE Std 802.22TM-2011, IEEE Standard for Wireless Regional
Area Networks
[3] Yonghong Zeng, Ying-Chang Liang, Anh Tuan Hoang, Rui Zhang,
“A review of spectrum sensing for cognitive radio: challenges and
solutions”, EURASIP Journal on Advances in Signal Processing.
[4] Ruiliang Chen, Jung-Min Park, Jeffrey H. Reed, “Defense against
Primary User Emulation Attacks in Cognitive Radio Networks” in
99
Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on June 25,2020 at 07:17:42 UTC from IEEE Xplore. Restrictions apply.
IEEE Journal on Selected areas in communication Volume 26
No1,2008
[5] Dr. Anubhuti Khare, Mahesh Saxena, Roshan Singh Thakur, Khyati
Chourasia, “Attacks and Preventions of Cognitive Radio Network-
A Survey” in International Journal of Advanced Research
in Computer Engineering and Technology (IJARCET) Volume
2, Issue 3, March 2013.
[6] Lianfen Huang, Liang Xie, Han Yu, Wumei Wang and Yan Yao,
“Anti-PUE Attack Based on Joint Position Verification in Cognitive
Radio Networks”, International Conference on Communications
and Mobile Computing (CMC), Vol.2, Shenzhen, China, April,
2010, pp.169-173
[7] Caidan Zhao,Wumei Wang, Lianfen Huang and Yan Yao, “Anti-
PUE Attack Based on the Transmitter Fingerprint Identification in
Cognitive Radio”, in 5th International Conference on Wireless
Communications, Networking and Mobile Computing (WiCom
‘09), Beijing, China, September, 2009, pp.1-5.
[8] Z. Jin, S.Anand, K.P Subbalakshmi, “Mitigating Primary User
Emulation Attacks in Dynamic Spectrum Access Networks using
Hypothesis Testing”, in Mobile Computing and Communication
review, Volume 13, Number 2.
[9] Z. Jin, S.Anand, K.P Subbalakshmi, “Detecting Primary User
Emulation Attacks in Dynamic Spectrum Access Networks”.
[10] S.Anand, Z. Jin, K.P Subbalakshmi, “An Analytical Model for
Primary User Emulation Attacks in Cognitive Radio Networks”.
[11] T. Charles Clancy, Nathan Goergen, “Security in Cogintive Radio
Networks Threats and Mitigation”.
[12] Ramzi Saifan, “Security issues in cooperative in-band sensing for
cooperative radio”, April 2010.
[13] Praveen Kaligineedi, Majid Khabbazian, Vijay K. Bhargava,
“Malicious User Detection in Cognitive Radio Cooperative
Sensing System” in IEEE transations on wireless communication
Vol 9,No 8, Auguest 2010.
[14] Bo Sun, Lawrence Osborne, Yang Xiao, Sghaier Guizani, Intrusion
detection techniques in mobile adhoc and wireless sensor
networks”.
[15] Amitabh Mishra, Ketan Nadkarni, Animesh Patcha, “Intrusion
Detection in Wireless Ad Hoc Networks”.
[16] Faisal Riaz, Imran Shafi,Syed Faraz Hassan,Waseem Akhtar "
Vehicle-to-Vehicle Communication enhanced by cognitive
approach and multi-radio technologies TECHNOLOGIES" in
the Proceedings of IEEE, ICET 2012 Muhammad Ali Jinnah
University,8-9 oct,pp.131-136
[17] Faisal Riaz, Saeed Ahmed, Imran Shafi, Faisal Iqbal, Muhammad
Ibrahim, “White Space Optimization Using Memory Enabled
Genetic Algorithm in Vehicular Cognitive Radio” in 11th IEEE
International Conference on Cybernetic Intelligent Systems August
23-24, 2012, Limerick, Ireland pp.133-138.
[18] Mitola, J.; Maguire, G.Q., Jr., "Cognitive radio: making software
radios more personal," Personal Communications, IEEE , vol.6,
no.4, pp.13,18, Aug 1999.
[19] Ruiliang Chen; Jung-Min Park, "Ensuring Trustworthy Spectrum
Sensing in Cognitive Radio Networks," Networking Technologies
for Software Defined Radio Networks, 2006. SDR '06.1st IEEE
Workshop in , vol., no., pp.110,119, 25-25 Sept. 2006.
[20] Deepa Das, Susmita Das “Primary User Emulation Attack in
Cognitive Radio Networks: A Survey” , IRACST – International
Jounral of Computer Networks and Wireless Communications
ISSN: 2250-3501 Vol 3, No 3 June 2013 pp 312-318.
100
Authorized licensed use limited to: NUST School of Electrical Engineering and Computer Science (SEECS). Downloaded on June 25,2020 at 07:17:42 UTC from IEEE Xplore. Restrictions apply.
