SPECTRUM SENSING IN COGNITIVE RADIO

NETWORKS: A SURVEY

P.T.V.BHUVANESWARI

Contents
1 Introduction
1.1 Cognitive Radio Network
1.2 Features of cognitive radio Network
1.3 Fundamental requirements of CRN
2 Spectrum Sensing
2.1 Principle of Spectrum Sensing
2.2 Issues and Challenges in spectrum sensing
2.3 Spectrum Sensing Mechanism
2.3.1 Cooperative Detection
2.3.2 Non-Cooperative Detection
2.3.3 Interference Based Detection
2.4 Comparison of Spectrum Sensing approaches
2.5 Need for Wideband Spectrum sensing
3 Conclusions
Acknowledgements
References
Biographical sketch

1. INTRODUCTION:

The available electromagnetic radio spectrum is in scarcity as its usage is increasing day
to day. On the other hand, due to conventional approach of spectrum allocation and management
and also due to existing inflexibility in operating frequency bands, it is found that the allocated
spectrum is underutilized. As most of the useful radio spectrums are already allocated, it is
cumbersome to detect the vacant bands to either deploy new services or to enhance existing ones.
To overcome this situation, Dynamic Spectrum Access (DSA) scheme [1] has been introduced to
improve spectrum utilization. The outcome of this scheme is Cognitive Radio Network (CRN)
1
technology, that aims to address the issues related to spectrum underutilization in wireless
communication.

1.1 Cognitive Radio Network:

The idea for Cognitive Radio Network (CRN) has been developed by Joseph Mitola at
the Defense Advanced Research Projects Agency (DARPA) in the United States. Full cognitive
radio is therefore sometimes known as “Mitola radio” software. According to Federal
Communications Commission (FCC), CRN is defined as "A radio or system that senses its
operational electromagnetic environment and can dynamically and autonomously adjust its
radio operating parameters to modify system operation, such as maximize throughput, mitigate
interference, facilitate interoperability, access secondary markets” [2].

In CRN, the users are classified into primary user and secondary user.

 Primary User (PU): A user who has higher priority or legacy rights on the usage of a
specific part of the spectrum.
 Secondary User (SU): A user who has a lower priority and therefore exploits the spectrum
in such a way that it does not cause interference to primary users.

Thus, the primary goals of a cognitive radio network are to (1) Provide highly reliable
communication to all users in the network, wherever and whenever required and (2) Provide cost
effective mechanism for efficient utilization of the radio spectrum.

1.2. Features of cognitive radio Network:

Features that CRN incorporates to enable more efficient and flexible usage of the spectrum is
given below [3]:

 Frequency Agility: It is the ability of CRN to adapt to changing operating frequency

environment.
 Dynamic Frequency Selection: It is the ability of the CRN to dynamically select the
optimal frequency from the sensed spectrum.
 Adaptive Modulation: It is the reconfiguration of transmission characteristics and
waveforms to enable efficient spectrum usage.
 Transmit Power Control: The transmission power need to be adaptive in order to enhance
spectrum utilization.

1.3 Fundamental requirements of CRN:

a. Coexistence of SUs with the PUs:

The fundamental requirement of the CRN is to tackle the coexistence of SUs with the
PUs. It can be achieved through three different paradigms [4].

2
  Interweave paradigm: The SU is denied on the access of the band that is under
usage by the PU in order to maintain the level of interference caused by the SU to
the PUs.
 Underlay and Overlay paradigms: The SU is permitted to concurrently access the
band used by the PU, thus enabling coexistence of SU with PU. However, the
interference caused by the SU to PU is kept minimal. In order to achieve this
requirement, it is necessary that SU be aware of various Channel Side Information
(CSI) such as channel conditions, encoding mechanism, etc. about the PU
network.

b. Spectrum Awareness

The above mentioned access paradigms require reliable spectrum awareness to obtain
information about the spectrum status. Spectrum awareness can be classified into two
approaches:
 Passive-based awareness: In this approach, the SU acquires the information about
the spectrum availability through external entities such as beacons or geo-location
databases.
 Active-based awareness: In this approach, the SU performs local sensing to
determine the spectrum status, and this is commonly known as Spectrum Sensing.
In both the approaches, periodic monitoring of the spectrum sensed and occupied
by the SU is mandatory, because the sudden reappearances of the PUs may
degrade the Quality of Service (QoS) of the SU.

c. Sensing reliability

Due to the stochastic nature of the wireless channel, the task of spectrum sensing performed
by the SU can be incorrect. Thus sensing becomes unreliable which may lead to degradation in
the QoS of both primary and secondary [5]. Hence, to combat fading in wireless channels,
cooperative communication that provides spatial diversity gains by enabling multiple SUs to
cooperate together in spectrum sensing and access is devised. This mechanism is called Co-
operative Cognitive Radio (CCR).

Cognitive Radio Network includes spectrum sensing, spectrum management, and

spectrum sharing and spectrum mobility.

•Spectrum sensing: The task of obtaining awareness about the spectrum usage and existence of
primary users in a geographical area.
•Spectrum management: The task of capturing the best available spectrum to meet user
communication requirements.
•Spectrum mobility: The task of maintaining seamless communication requirements during the
transition to better spectrum.
•Spectrum sharing: The task of providing the fair spectrum scheduling method among coexisting
users.

3
 This chapter details the issues and challenges in spectrum sensing. It also presents the
state of art in the spectrum sensing algorithms.

2. SPECTRUM SENSING

Spectrum Sensing (SS) is the ability to measure, sense and be aware of various
parameters related to the radio channel characteristics such as availability of spectrum, transmit
power, interference, noise, operating frequency of the radio, user requirements and applications,
available networks and nodes, local policies and other operating restrictions. It is done across
frequency, time, geographical space, code and phase. To allow reliable operation of cognitive
radios, following task has to be carried out in Spectrum sensing [6].

Task of spectrum sensing:

1) Detection of spectrum holes

2) Spectral resolution of each spectrum hole
3) Estimation of the spatial directions of incoming interferences
4) Signal classification

In practice, the Secondary Users (SUs) also known as unlicensed users need to
continuously monitor the activities of the Primary Users (PUs) who are also called as licensed
users in order to identify the presence of the Spectrum Holes (SHs). This procedure is called
spectrum sensing. There are two types of SHs, namely temporal and spatial SHs, respectively.

 Temporal Spectrum Holes: Appears whenever PU transmission does not occur for certain
period of time, and SUs is allowed to use the spectrum for transmission.
 Spatial Spectrum Holes: Appears whenever the PU transmission occurs within certain
geographical area and the SUs is allowed to use the spectrum outside that area.

2.1 Principle of Spectrum Sensing

The principle of spectrum sensing is illustrated in Figure 1, from which it can be

observed that the PU transmitter is sending data to PU receiver in their allotted licensed spectrum
band. In the same instant, pair of SUs also intends to access the spectrum. In order to protect the
PU transmission, spectrum sensing needs to be performed by the SU transmitter which determine
the presence of PU receiver that are present within the coverage of the SU transmitter.

4
 Figure 1. Principle of Spectrum Sensing

2.2 Issues and Challenges in spectrum sensing:

In this section, various issues such as channel uncertainty, noise uncertainty, sensing
interference limit etc., that has to be addressed by spectrum sensing in cognitive radio networks
[7] are discussed.

 Channel Uncertainty: In wireless communication networks, uncertainties in received

signal strength can arise due to channel fading or shadowing. In spectrum sensing, the
deep fade experienced by the primary user signal may be wrongly interpreted by the
secondary user as white space. Then the secondary user transmission can become
interference to PU. Hence, cognitive radios have to be more sensitive to distinguish a
faded or shadowed primary signal from a white space. Cooperative Sensing is the
mechanism devised to handle this issue, in which group of cognitive radios share their
local measurements and a collective decision can be made on the occupancy status of the
licensed band.
 Noise Uncertainty: The detection sensitivity can be defined as the minimum SNR (signal
to noise ratio) at which the primary user signal can be accurately detected by the
cognitive radio and is given by,

(1)
Where N is the noise power, Pp is transmitted power of the primary user, D is the
interference range of the secondary user, and R is maximum distance between primary
transmitter and its corresponding receiver.

Limitation: Determining noise power is difficult in a practical scenario. Hence it need to be

estimated which may have calibration errors due to change in thermal noise. Hence to calculate
noise uncertainty, it is necessary to have more sensitive detector.

5
  Aggregate Interference: It future, the possibility of widespread deployment of secondary
users operating over the same licensed band increases; this may result in uncertainty
which is known as aggregate interference. Even though, the primary user may lie outside
the interference range of a secondary user, detection of white space may become faulty
due to aggregate interference. Hence, in order to enhance the accuracy of detection, usage
of more sensitive detector becomes mandatory.
 Sensing Interference: The Primary goal of spectrum sensing is to detect the spectrum
status which may be either idle or occupied, so that access can be made by the unlicensed
user. However the crucial challenge faced in practice is errors involved in interference
measurement. In order to compute the level of interference caused by the unlicensed user
(SUs) to the licensed user (PUs), SUs need to be aware of the location of the PUs.
However, when the PU device becomes passive, determination of location becomes
cumbersome. Hence, these factors need to be given more attention while calculating the
sensing interference limit.

2.3 Spectrum Sensing Mechanism:

Spectrum sensing mechanism can be classified into narrowband sensing and wideband
sensing. This section deals with some of the common narrowband spectrum sensing mechanism
and through some lights on wideband spectrum sensing mechanism.

The taxonomy of narrowband spectrum sensing mechanism is illustrated in Figure 2.

They can be broadly classified under two categories: Cooperative Detection Technique and Non-
cooperative Detection Technique [8].

6
 Figure 2. Taxonomy of narrowband spectrum sensing mechanism

2.3.1. Cooperative Detection:

In this technique, group of Cognitive Radios (SUs) share sensing information to each
other to enhance the sensing accuracy. In this process, group of SUs collect the channel
occupancy information and represent them as a spectrum map. This information is then
periodically transmitted to the Central Coordinator which determines the set of available UHF
(Ultra High Frequency) channels by performing a bitwise-OR operation on the spectrum maps.
Then the best available channel is selected and broadcasted back to SU. This technique exploits
the spatial diversity intrinsic to a multi-user network. It can be accomplished in a centralized or
distributed fashion [9]. Cooperative spectrum sensing adopts three approaches namely
Centralized, Distributed and Relay-assisted cooperative.

a. Centralized approach: In this approach, Fusion Center (FC) or central processor node collects
sensing information from all the secondary user radios that are within its coverage as shown in
Figure 3. Based on the analyses performed over them, the optimal frequency of operation for the
SU can be determined [10].

7
 Figure 3. Centralized Approach

b. Distributed approach: In this approach, the concept of central processing is not adopted.
Instead, each cognitive radio can share the sensing information gathered by them with other
radios that are within its coverage as shown in Figure 4. Thus, an ad-hoc approach is adopted
which requires higher level of autonomy for each radio [11].

Figure 4. Distributed Approach

c. Relay-assisted cooperative: In this approach, relay cognitive radios are used between
secondary user and fusion centre to mitigate the effect of fading in the wireless medium as
shown in Figure 5. This enhances accuracy involved in spectrum sensing.

8
 Figure 5. Relay-assisted cooperative approach

Procedure involved in the cooperative sensing:

Cooperative sensing consist of two phase such as detection of PU and reporting them to the
FC.
• Detection of PU: In this phase, secondary user attempt to detect primary unused channel.

• Reporting: In this phase, the detected information is reported to the Fusion Centre for
processing.

2.3.2. Non-Cooperative Detection:

In this technique, individual radios perform spectrum occupancy measurements and

analysis locally to detect primary unused channel. Hence each node is autonomous in nature.
Non - Cooperative spectrum sensing adopts two approaches namely blind sensing, and Signal
Specific sensing.

a. Blind Sensing: In this approach, no prior information related to primary users and noise
power is not required to perform spectrum sensing. For a given number of observation samples
of primary channels, statistical analysis is performed which reflects the channel condition. Based
on the obtained result, detection threshold value is computed. Through efficient decision making
algorithm, spectrum sensing is achieved. Some of the popular blind sensing mechanisms are
detailed in this section.

(i) Energy Detector based sensing: Energy detector based sensing also known as
radiometry or periodogram, is the most popular spectrum sensing method because of its low
computational and implementation complexities [12]. In this approach, the receiver does not
require any prior knowledge about the Primary Users (PU). The signal of PU is detected by
comparing the output of the energy detector with a threshold value which depends on the noise
floor. However few challenges exist in this approach which includes:

9
 Selection of adaptive threshold value to detect primary users
 Inability to differentiate interference from primary users and noise
 Poor performance under low signal-to-noise ratio (SNR) values and detection of spread
spectrum signals.

The block diagram of Energy detector based sensing approach is illustrated in figure 6.

Figure 6. Energy detector based Sensing

In this approach, the received signal obtained from the channel is passed through band
pass filter of the bandwidth W and is integrated over time interval. The output from the integrator
block is then compared to a predefined threshold λE to discover absence of the primary user. The
threshold value λE can be either fixed or variable based on the channel conditions.

Let s(n) be the signal transmitted by the primary user which is to be detected, w(n) be the
additive white Gaussian noise (AWGN) sample and n be the sample index, then the received
signal y(n) can be expressed as

(2)

Let N be the period over which the detection of spectrum occupancy is observed. Then
the decision metric M made in the energy detector can be expressed as:

(3)

The decision on the occupancy of a band can be obtained by comparing the decision
metric M against a fixed threshold λE.
Two hypothetical conditions namely Ho and H1 represent the status of detection algorithm.
H0 : y(n) = s(n) + w(n), indicate presence of primary user
H1 : y(n) = w(n), when s(n) = 0, indicates the absence of primary user. Hence the spectrum can be
assigned to the secondary user.

The performance of the energy detector based sensing algorithm can be analyzed using
metrics namely probabilities:
 Probability of detection PD
 Probability of false alarm PF

10
 PD is the probability of detecting the presence of PU signal when it is present. It can be
formulated as

(4)
while PF is the probability of detecting the presence of PU signal when it is not present.

(5)

In order to prevent the underutilization of spectrum, it is necessary that PF should be kept

as small as possible when compared to PD. The value of λE aids to achieve an optimum balance
between PD and PF.

(ii) Eigen value based Sensing: In this approach, Eigen value of the covariance matrix
of the received signal is used in detection of primary user. The ratio of the maximum Eigen value
to the minimum Eigen is computed using random matrix theory. Detection threshold value is
chosen from this ratio [13,14].

(iii) Covariance based Sensing: It is a known fact that statistical covariance matrices of
the received signal and that of noise are normally different. By differentiating this difference,
desired signal component can be obtained from the background noise [15].

b. Signal Specific: This sensing approach requires prior knowledge of Primary User (PU) signal.

(i)Waveform based Sensing: In this approach, the prior knowledge of the Primary User
signal pattern is known through various methods such as preamble, midamble, pilot signals, and
spreading sequence. Preamble is a known sequence transmitted before each burst and midamble
is a sequence transmitted in the middle of a burst. It is also known as waveform-based sensing
or coherent sensing. This approach outperforms energy detector based sensing in terms of
reliability and convergence time. Further, performance of this approach is enhanced when the
length of the known signal pattern is increased [16].
The decision matrix M can be computed from the given expression:

(6)

Where s(n)* is the conjugate of S(n).

In absence of PU, the metric value becomes

(7)
Similarly, in the presence of a PU, it becomes

11
 (8)

The decision on the detection of the presence of a PU signal is made by comparing the
decision metric M with the fixed threshold λE. Since this approach require short observation
period T, it is susceptible to synchronization errors.

(ii) Transmitter Based Sensing: In this approach, secondary user (SU) can detect the
presence of the primary users which are in its vicinity. When the received signal strength (RSS)
of the primary user with respect to secondary user is weak, it can be considered to be located far
away from the Secondary user. Then, the sensed spectrum can be made available for SU. The
basic hypothesis for transmitter detection is given below:

(9)
where, x(t) is the received signal, s(t) is the primary user transmitted signal, n(t) is
additive white Gaussian noise (AWGN) and h is the amplitude gain of the channel. As mentioned
earlier H0 indicate presence of primary user and H1 indicates the absence of primary user. This
approach suffers under fading channel condition, as distance between PU and SU alone cannot
be considered for RSS degradation.
The three main hypothesis based transmitter detection approach are described below.

 Energy Sensing: Whenever information gathered about the signal of the primary user
become insufficient, and power of the random Gaussian noise is alone known, then
spectrum sensing can be done with optimal energy detector. The operation of the energy
sensing model is shown in figure 7. This model is simple and can be implemented
efficiently using Fast Fourier Transform (FFT) algorithm [17].

Figure 7. Energy Sensing model

 Matched Filter Sensing: A matched filter (MF) is a linear filter designed to maximize
the output signal to noise ratio (SNR) for a given input signal. When secondary user has a
priori knowledge of primary user signal, matched filter detection can be applied. The
operation of the matched filter based sensing which is equivalent to correlation is
illustrated in figure 8.

12
 Figure 8. Matched filter based sensing

The primary user signal x is convolved with h, the impulse response of the matched filter
which the time shifted version of a reference signal. The operation of matched filter detection is
expressed as:

(10)

Where x is the primary user signal, h is the impulse response of matched filter.H0 indicate
presence of primary user and H1 the absence of primary user.

This approach is computationally efficient as it involves few samples and provides promising
result when primary user signal information is known to the secondary user [18]. However, it
suffers from few drawbacks such as accurate prior knowledge of every primary signal and
dedicated receiver to acquire signal from different type of primary user

 Cyclostationary Based Sensing: In this approach, the inherent cyclostationary features

such as periodicity and spectral correlation of the received primary signal is exploited to
sense the availability of the spectrum as shown in figure 9. This approach uses cyclic
spectral correlation function parameter to detect presence of the primary user signals.
Due to noise rejection capability, this approach outperforms energy detection method
even in low SNR regions. It is more suitable for practical scenario, when prior knowledge
about primary user signal remains unknown. Further the type of modulation scheme used
by the primary user signal can also be determined. However, this approach suffer
drawbacks such as spectral leakage of high amplitude signals, non-linearity, high
computational complexity [19].

Figure 9. Cyclostationary Based Sensing

13
c. Radio Identification Based Sensing: From the above discussed sensing approaches, it can be
clearly understood that the impact of interference between transmitter (primary user) and
receiver (secondary user) play a crucial factor affecting sensing accuracy. Hence, in this
approach a study on interference with respect to transmitter is analyzed. It is observed that by
regulating the output transmission power and out-of-band emissions based on its location with
respect to other users, the inference caused due to transmitter can be minimized [20].

2.3.3. Interference Based Detection

In section 1.3, the Coexistence of SUs with the PUs in CRN has been discussed. The
above mentioned spectrum sensing approaches falls under the category of Interweave paradigm.
However in practical scenario Underlay and Overlay paradigms also exist. In this section, two
spectrum sensing approaches under this category are presented.

a. Primary Receiver Detection: In real time wireless communication environment, the local
oscillator (LO) present in the receiver emit leakage power from its front end while receiving data
from the transmitter. Such leakage power may cause interference to the adjacent transceiver. In
CRN, this scenario can be experienced in an underlay or overlay paradigm. When LO of the
primary receiver emits leakage power while receiving the data from primary transmitter. It
causes interference to the other cognitive users that are within its vicinity. Therefore, to alleviate
this problem, a low cost sensor node is mounted close to the receiver of the primary user to
detect the leakage power emission. The local sensor node then report the sensed information to
the CR users from which spectrum occupancy status can be identified [21].

b. Interference Temperature Management: Similar to the previous approach, the basic idea
behind the interference temperature management is to fix an upper interference limit for the
given frequency band used by the CR user according to the geographic location, so that the
interference caused by them when using the specific band in specific area can be minimized.
This is achieved by regulating the transmission power of the CR user transmitters based on their
locations with respect to primary users. The interference caused by the CR user is measured at
the receiver (primary users) [21]. Thus, coexistence of SUs with PUs is achieved by restricting
the interference temperature level.

2.4. Comparison of Spectrum Sensing Approaches:

Figure 10 shows the comparison of various spectrum sensing approaches. It can be

inferred that Waveform-based sensing approach is more robust when compared to both energy
detector and cyclostationary based sensing as it involves coherent processing using deterministic
signal component. However, it requires priori information about the characteristic of the primary
user in the form of pilot or pattern which adds complexity on the sensing module. The
performance of energy detector based sensing is not promising when noise is not stationary and
its variance is not known. Further the effects of baseband filter and spurious tones also influence
the performance degradation. From the literature [19], it is observed that cyclostationary based
approach performance may be worse when compared to energy detector based sensing, whenever
noise happen to be stationary.

14
 Figure 10. Comparison of various spectrum sensing approaches

However, noise becomes non-stationary in the presence of co-channel or adjacent channel

interferers. Under this circumstance energy detector based sensing approach fails while
cyclostationary-based approach does not get affected. On the other hand, cyclostationary features
may be completely lost due to channel fading [22]. Hence certain tradeoffs need to be
considered while selecting a spectrum sensing approach. The characteristics of primary users are
found to be major design factor in selecting a spectrum sensing approach. Other factors include
required accuracy, sensing duration requirements, computational complexity, and network
requirements.

2.5 Need for Wideband Spectrum sensing:

The cooperative sensing adopted in narrowband spectrum sensing exploits the spatial
diversity of cooperating CR (SUs) focusing sensing of one frequency band in each round of
cooperation. In order to determine the availability of the spectrum in multiple bands, each time
the CR users has to switch to each band separately which may leads to synchronization problem.
This process can incur significant switching delay and synchronization overhead. Wideband
spectrum sensing is an alternative approach, in which CR users can cooperatively sense multiple
bands simultaneously involving lesser sensing time [23]. Wideband cooperative sensing has
recently gained much attention in the literature. However, it raises many open research issues
such as narrowband–wideband tradeoff and signal classification. Although wideband sensing
when compared to narrowband sensing, reduces sensing time and channel switching overhead,
but it requires additional hardware to facilitate simultaneous detection of multiple bands. Hence,
a tradeoff between both sensing approaches has to be investigated in terms of various detection
performance matrices such as sensing delay, complexity and throughput. Several approaches
exist in the literature [23]. However it is beyond the scope of this chapter.

15
3. CONCLUSIONS

In this chapter, fundamental concept and requirements of the cognitive radio network
along with four primary functions of a cognitive radio such as spectrum sensing, spectrum
management, spectrum sharing, and spectrum mobility has been described. Followed by that,
principle and various narrowband spectrum sensing approaches are explained and a comparison
has been made. Finally, the need for wideband spectrum sensing along with advantage and
disadvantage are discussed. This chapter can be used as base references who wish to pursue
research in this field.

ACKNOWLEDGEMENTS
I would like to thank my research scholars Saraswathi Priyadharshini . A, Bino.J, and
Ezhilarasi.S for their support in making this chapter.

REFERENCES:

1. Yan Zhang ; Simula Res. Lab., Lysaker, “Dynamic Spectrum Access in Cognitive
Radio Wireless Networks”, Proceeding of IEEE International Conference on
Communications pp.no. 4927 – 4932, 19-23 May 2008.
2. Anant Sahai, Niels Hoven, Shridhar Mubaraq Mishra, Rahul Tandra, “Fundamental
tradeoffs in robust spectrum sensing for opportunistic frequency reuse”, IEEE Journal
Selected Areas of Communications, 2006.
3. Tulika mehta, Naresh kumar and Surender s saini, “Comparison of Spectrum Sensing
Techniques in Cognitive Radio Networks” International Journal on Electronics &
Communication Technology, vol. 4, special issue - 3, pp.no. 33-37, April - June 2013.
4. Ghaith Hattab and Mohamed Ibnkahla, “Multiband Spectrum Access: Great Promises
for Future Cognitive Radio Networks”, Proceedings of the IEEE, Vol. 102, No.3,
pp.no.282 -306, March 2014.
5. E. Biglieri, A. J. Goldsmith, L. J. Greenstein, and N. B. Mandayam, “Principles of
Cognitive Radio”, Cambridge University Press, 2012.
6. Weifang Wang, “Spectrum Sensing for Cognitive Radio” Proceedings of the IEEE
Computer society, pp. no.110-112, 2009.
7. Mansi Subhedar and Gajanan Birajdar, “Spectrum Sensing Techniques in Cognitive
Radio Networks: A Survey”, International Journal of Next-Generation Networks
(IJNGN) Vol.3, No.2, June 2011.
8. Manish Jaiswal, Anuj Kumar Sharma and Vikash Singh, “A Survey on Spectrum
Sensing Techniques for Cognitive Radio”, Conference on Advances in
Communication and Control Systems, pp. no. 647 to 660, 2013.
9. Vatsal Sharan and Prenit Wankhede, “ Spectrum Sensing for Cognitive Radio”. (white
paper).
10. A.D Cabric, S. Mishra, R. Brodersen, “Implementation issues in spectrum sensing for
cognitive radios”, Processing of Asilomar conference on signals, systems, and
computers, vol. 1, pp. no. 772–776, 2004.

16
11. S. Visotsky, Kuffner, R. Peterson, “On collaborative detection of TV transmissions in
support of dynamic spectrum sharing,” Proceedings of IEEE DySPAN, pp. 338–345,
2005.
12. F.F. Digham, M.S. Alouini, M.K. Simon, “On the energy detection of unknown
signals over fading channels”, IEEE Transactions on Communications, Vol 55, issue
1, pp. no.21-24, 2007.
13. Narushan Pillay, HongJun Xu, “Eigenvalue-based spectrum sensing using the exact
distribution of the maximum eigenvalue of a Wishart matrix”, IET Signal Processing
Vol. 7, Issue. 9, pp. 833–842, 2013.
14. Shu Wang, Junjie Bao, Bin Shen, Qiong Huang and Qianbin Chen, “Eigenvector
Based Cooperative Wideband Spectrum Sensing for Cognitive Radios” Proceedings
of IEEE International Conference on Ubiquitous and Future Networks, pp. 346-351,
2014.
15. Y. Zeng and Y.-C. Liang, “Eigenvalue-based Spectrum Sensing Algorithms for
Cognitive Radio,” IEEE Transaction on Communications, vol. 57, no. 6, pp. 1784–93,
2009.
16. Y. Zeng and Y. C. Liang, “Spectrum-sensing algorithms for cognitive radio based on
statistical covariances”, IEEE Transaction on Vehicular Technology, Vol. 58, no. 4,
pp. 1804–1815, 2009.
17. Xin Liu, Min Jia , Xuemai Gu and Xuezhi Tan, “Optimal Periodic Cooperative
Spectrum Sensing Based on Weight Fusion in Cognitive Radio Networks”, Sensors
Vol 13, 5251-5272, 2013.
18. Ian F. Akyildiz, Brandon F. Lo, Ravikumar (2011), “Cooperative spectrum sensing in
cognitive radio networks: A survey, Elsevier Journal on Physical Communication”,
pp: 40-62, 2011.
19. P. D. Sutton, J. Lotze, K. E. Nolan, and L. E. Doyle, “Cyclostationary signature
detection in multipath Rayleigh fading environments,” Proceedings of IEEE
International conference on Cognitive Radio Oriented Wireless Networks and
Communincation, 2007.
20. Tevfik Yucek and Huseyin Arslan, “Spectrum characterization for opportunistic
cognitive radio systems”, Proceedings of IEEE conference of Military
Communication, pp. no. 1-6, 2006.
21. L.Thanayankizil, A.Kailas, “Spectrum Sensing Techniques (II): Receiver Detection
and Interference Management”, 2008 (white paper).
22. Tevfik Yucek and Huseyin Arslan, “A Survey of Spectrum Sensing Algorithms for
Cognitive Radio Applications” Proceedings of the IEEE communications surveys &
tutorials,Vol. 11,No.1, First quarter, pp. no.116-130, 2009.
23. Mario Bkassiny and Sudharman K. Jayaweera, “Robust, Non-Gaussian Wideband
Spectrum Sensing in Cognitive Radios”, IEEE Transactions on Wireless
Communications, Vol. 13, No. 11, pp. no.6410-6421, November 2014.

17