IRJET- Simulating Spectrum Sensing in Cognitive Radio Network using Cyclostationary Technique

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 08 | Aug 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 1380
Simulating Spectrum Sensing in Cognitive Radio Network using
Cyclostationary Technique
Fele Taiwo1, Ogunlola Okunola Olasunkanmi2, Fele Yetunde Olamide3
1Computer Science Department, The Federal Polytechnic, Ado-Ekiti, Nigeria
2Computer Science Department, The Federal Polytechnic, Ado-Ekiti, Nigeria
3Logistics Management and Coordination Unit, Ministry of Health, Ado-Ekiti, Nigeria
----------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The demand for wireless communication
applications are increasing and the available
electromagnetic spectrum band is getting crowded
geometrically. Spectrum sensing helps to detect the
spectrum holes (unutilized bands of the spectrum) in
providing high spectral resolution capability. Therefore for
efficient utilization of spectrum, we need to sniff the
spectrum to determine whether it is being used by licensed
owner or not. In an attempt to contribute to the possibility
of adopting dynamic spectrum access as an alternative
radio spectrum regulation system. The paper review
different spectrum sensing techniques used in finding
spectrum holes in available radio resource, model and
simulate cyclostationary based spectrum sensing technique
and classify primary user signals of different modulation
scheme. The results of this study show that accurate and
prompt modulation classification is possible beyond the
lower bound of 5 dB acclaimed in literature. The
performance of the detection technique is measured in
terms of the ROC curve. The proposed model is simulated on
a Laptop PC running on Windows 10 platform and requires
MATLAB R2015a/Simulink and LibSVM.
Key Words: Spectrum sensing, Cognitive radio,
Modulation classification, Spectrum holes, Cyclostationary
detection.
1. INTRODUCTION
The need for wireless communication applications are
increasing and the available Electromagnetic Spectrum
band is getting crowded day by day. According to many
researches it has been found that the allocated spectrum
(licensed spectrum) is not utilized properly because of
static allocation of spectrum. It has become most difficult
to find vacant bands either to set up a new service or to
enhance the existing one. In order to overcome these
problems we are going for “Dynamic Spectrum
Management” which aims at improving spectrum
utilization [1].
Wireless multimedia applications and other real-time
applications need high bandwidth, as static frequency
allocation techniques cannot resolve the problems of an
increasing number of high data rate services. This problem
can be resolved by improving spectrum resource
utilization. In this paper we investigate the performance of
Cyclostationary Spectrum Sensing technique. Specifically
we investigate a cyclostationary based sensing detector’s
ability to differentiate between a BPSK or a QPSK
modulated signal. The objectives of this study are to: study
and analyse existing spectrum sensing techniques; Design
optimized sensing technique based on cyclostationary
and; simulate the design above.
1.1 Cognitive Radio (CR)
Cognitive Radio (CR) is a form of wireless
communication in which a transmitter / receiver can
intelligently detect communication channels that are in use
and those which are not, and can move to unused channels.
This optimizes the use of available radio frequency
spectrum while minimizing interference with other users.
A primary feature of cognitive radios is the ability to adapt
the transmission parameters given a dynamic wireless
environment. Cognitive Radio works on dynamic Spectrum
Management principle which solves the issue of spectrum
underutilization in wireless communication in a better
way. This radio provides a highly reliable communication.
Fig - 1 shows the Dynamic Spectrum Access in Cognitive
Radio.
Fig -1: Dynamic Spectrum Access [2]
CR technologies utilize a radio frequency (RF) sensor to
detect unused spectrum that is available and capable of
communications. CR understands the properties inherent
to the user such as battery life, signal interface, and
attenuation, which are then used in a set of decision-
making algorithms to provide the best capabilities for each
user.

Cognitive radios can also change frequencies dynamically
to maintain reliable communications [3]. As a result, CR
helps improve the efficiency of spectrum usage. CR
technologies can also be used to ensure that new
unlicensed users do not interfere with TV signals when
databases of incumbent licensees are available for a given
location, devices can be instructed to avoid those
frequency bands [4].
Cognitive radio has four major functions. They are:
Spectrum Sensing, Spectrum management, Spectrum
Sharing and Spectrum Mobility [5]. Spectrum Sensing is to
identify the presence of licensed users and unused
frequency bands i.e., white spaces in those licensed bands.
Spectrum Management is to identify how long the
secondary users can use those white spaces [7]. Spectrum
Sharing is to share the white spaces fairly among the
secondary users. Spectrum Mobility is to maintain
unbroken communication during the transition to better
spectrum [8].
In terms of occupancy, sub bands of the radio spectrum
may be categorized as follows:
i. White spaces: These are free of RF interferers,
except for noise due to natural and/or artificial
sources.
ii. Gray spaces: These are partially occupied by
interferers as well as noise.
iii. Black spaces: The contents of which are
completely full due to the combined presence of
communication and (possibly) interfering signals
plus noise [10].
Fig - 2 shows the White Spaces and Used Frequencies in
Licensed Spectrum.
Fig - 2: White Spaces in Licensed Bands. [2]
When compared to all other functions, Spectrum Sensing
is the most crucial task for the establishment of cognitive
radio based communication networks.
1.2 Frequency Management Policy
Radio frequency spectrum is one of the key natural
resources of great economic value as a result of its direct
application in telecommunications, broadcasting, military
operations, and scientific research in addition to a range of
other socioeconomic activities such as social services, law
enforcement, education, healthcare, transportation, etc. As
a result, many industries depend heavily on the efficient
utilization of radio frequency spectrum.
These crucial factors therefore, make it mandatory for the
government to develop comprehensive and clear-cut
policies that will ensure that spectrum resource is
optimally utilised for the overall benefit of the nation [6].
2. SPECTRUM SENSING IN COGNITIVE RADIO
Spectrum sensing is the ability to measure, sense and be
aware of the parameters related to the radio channel
characteristics, availability of spectrum and transmit
power, interference and noise, radio’s operating
environment, user requirements and applications,
available networks(infrastructures) and nodes, local
policies and other operating restrictions [9]. It is done
across Frequency, Time, Geographical Space, Code and
Phase. A number of different methods are proposed for
identifying the presence of signal transmission all of which
are in early development stage. They are:
i. Energy-Detection Based
ii. Waveform Based
iii. Cyclostationary Based
iv. Radio Identification Based
v. Matched filtering Based
2.1 Cyclostationary Feature Detection
Cyclostationary feature detection based on introduction of
periodic redundancy into a signal by sampling and
modulation. The periodicity in the received primary signal
to identify the presence of Primary Users (PU) is exploited
by Cyclostationary feature detector [10] which measures
property of a signal namely Spectral Correlation Function
(SCF) given by
( ) ∫ ( )
Where ( ) is cyclic autocorrelation function (CAF).
Cyclostationary feature detector implementation can
differentiate the modulated signal from the additive noise,
distinguish Primary User signal from noise [11]. It is used
at very low SNR detection by using the information
embedded in the Primary User signal which does not exist
in the noise. This technique is robust to noise

discrimination and it performs better than energy
detector. It has disadvantage of more computational
complexity and longer time observation [12].
3. METHODOLOGY
The paper started with literature review of previous
research works relevant to the topic was highlighted. This
work assumed a cognitive radio network with N primary
users and W secondary users. For any one of the
secondary user, the presence of the primary user can be
summarized as a hypothesis test model of two elements:
H0 : x(t) = w(t) (1)
H1 : x(t) = s(t) + w(t) (2)
Based on received signal x(t), which is a function of
transmitted signal s(t) and white additive Gaussian noise
w(t) there are two hypothesis: in which when the primary
user is present, H1 and the other, in which the primary
user is absent, H0.
Considering a known signal s(t) corrupted by additive
white Gaussian noise w(t) as the received signal x(t). Then,
x(t) = s(t) + w(t) 0 t T (3)
A continuous time signal x(t) is said to be cyclostationary
(in wide sense), if it exhibits a periodic auto-correlation
function which is given by:
( ) , ( ) ( )- (4)
Where E[·] represents statistical expectation operator.
Since Rx (t, τ) is periodic, it has the Fourier series
representation.
( ) ∑ ( ) (5) (5)
( ) ∫ ( ) ( ) (6)
Where sum is taken over integer multiple of fundamental
cyclic frequency, α and ( ) is Cyclic Autocorrelation
Function (CAF). Considering a time series of length T, the
expectation in the definition of autocorrelation can be
replaced by time average. So that:
( ) ∫ ( ) (7)
‘n/T’ represent the cyclic frequencies and can be written
as ‘α’. A wide sense stationary process is a special case of a
wide sense cyclostationary process for ‘ n/T = α = 0’.
Therefore, a signal exhibit second-order cyclostationarity
in the wide sense when its cyclic autocorrelation function,
( ) is different from zero for some non-zero α. For zero
frequency shift, the spectral correlation density is
equivalent to standard power spectral density. The
amount of correlation between frequency shifted versions
of x(t) in the frequency domain is determined. The
spectral correlation density (SCD) function is defined as
Fourier transform of Cyclic Autocorrelation Function of
x(t). The SCD of a signal x(t) is given by:
( ) ∫ ( ) (8)
To analyze a signal in the frequency domain, the power
spectral density (PSD), Sx(f), is used to characterize the
signal, which is obtained by taking the Fourier Transform
of the autocorrelation Rx(τ ) of the signal x(t). The PSD and
the autocorrelation of a function, Rx(τ), are mathematically
related by the Einstein-Wiener-Khinchin (EWK) relations,
namely:
( ) ∫ ( )
( ) ∫ ( ) (9)
Using the EWK relations, we can derive some general
properties of the power spectral density of a stationary
process, such as:
( ) ∫ ( )
* ( )+ ∫ ( )
( )
( ) ( )
Power Spectral Density is a special case of SCF when α = 0.
The power spectral density, appropriately normalized, has
the properties usually associated with a probability
density function:
( )
( )
∫ ( )
(10)
Using H(f) to denote the frequency response of the system,
we can relate the power spectral density of input and
output random processes by the following equation:
( ) ( ) ( ) (11)
Where X(f) is the PSD of input random process and Y(f) is
the PSD of output random process.
We calculate SCF for each one of BPSK, QPSK, MPSK and
QAM and found that a sinusoidal signal with carrier
frequency fc have four peaks in CSD at (α = 0, f = ± fc ) and
(α = ± 2fc, f = 0). We compute the SCF of the received signal
y(t) taking into account that the primary user transmits a
cyclostationary signal, so, its SCF has nonzero component
at some nonzero cyclic frequency. Hence we can rewrite
the hypothesis in (7) and (8) as:

( ) {
( )
( ) ( )
(12)
Where ( ) ( )is SCF of AWGN and primary signal
respectively and λ is threshold. This states that
( ) that is, x(t) is cyclostationary signal, then we
can robustly detect the presence of the primary signal. We
focus only on frequencies (α = 0, f = ± fc ) and (α = ± 2fc, f =
0) and look for peaks. These peaks are compared to a pre-
determined threshold, so that if they are greater than the
threshold and the other values on the same frequencies (α
= 0 and f = 0), then the signal exists in the band under
sensing or the band is free otherwise[10].
Fig - 3: Block diagram of the proposed Cyclostationary
Feature Detection.
3.1 Signal Generation
The first step of this experiment was to generate signals
belonging to different families of modulation schemes. This
was done in order to highlight how certain modulation
schemes can vary spectrally while others possess similar
characteristics. This model features four very basic
modulation schemes that are pulse shaped for over the air
transmission BPSK, QPSK, M-PSK and QAM. Vectors of each
transmission were saved to the MATLAB workspace.
Fig -4a: Simulink model of cyclostationary without using
modulation
ig -4b: Simulink model of cyclostationary without using
modulation (output)
Fig - 5a: Simulink model of cyclostationary using BPSK
modulation
Fig – 5b: Simulink model of cyclostationary using BPSK
modulation (output)
Fig – 6a: Simulink model of cyclostationary spectrum
sensing using QPSK

Fig – 6b: Simulink model of cyclostationary spectrum
sensing using QPSK (output)
Fig – 7a: Simulink Model of Cyclostationary spectrum
sensing using MPSK
Fig – 7b: Simulink Model of Cyclostationary spectrum
sensing using MPSK (output)
Fig – 8a: Simulink model of cyclostationary spectrum
sensing using QAM
Fig – 8b: Simulink Model of Cyclostationary spectrum
sensing using QAM (output)
When the transmitted signal used is QPSK modulated and
the carrier frequency is 200 Hz and cyclostationary
spectrum sensing is performed, we get two peaks at
400Hz frequency. The two peaks signify that the
modulation scheme used by the primary user is QPSK and
the reason for getting peaks at double the carrier
frequency is autocorrelation of the received signal. When
the transmitted signal used is BPSK modulated and the
carrier frequency is 200 Hz and cyclostationary spectrum
sensing is performed, we get a single peak at 400Hz
frequency which signifies that the modulation scheme
used by the primary user is BPSK and the reason for
getting a peak at double the carrier frequency is
autocorrelation of the received signal.
4. SIMULATION RESULTS
The cyclic spectral density using cyclostationary detector
has been plotted for white Gaussian noise, QPSK
modulated signal for different SNR values. By observing
the cyclic spectral density of signal, the decision about the
signal presence and its modulation scheme can be made.
When the transmitted signal used is QPSK modulated and
the carrier frequency is 200 Hz and cyclostationary
spectrum sensing is performed, we get two peaks at
400Hz frequency. The two peaks signify that the

modulation scheme used by the primary user is QPSK and
the reason for getting peaks at double the carrier
frequency is autocorrelation of the received signal.
The performance of a detector is characterized by two
parameters, the probability of missed detection (PMD) and
the probability of false alarm (PFA), which are defined as:
ɛ = PFA = Prob {Decide H1|H0} and
δ = PMD = Prob {Decide H0|H1}.
A typical receiver operating characteristic (ROC), which is
a plot of 1−δ, the probability of detection (PD), versus the
probability of false alarm (PFA), is shown in Fig -9, 10 and
11.
Fig – 9: Performance at SNR of -10dB
The graph represents probability of false alarm x axis and
in y axis the probability of detection is represented. In Fig -
9 at low SNR -10dB, it begins with high probability of
detection ( 0.7210). In Fig - 10 When decrease low SNR to -
15dB, cyclostationary detection begin with probability of
detection(0.1840)and in Fig - 11 at low SNR to -20dB,
cyclostationary detection is begin with probability of
detection( 0.073). It is observed that cyclostationary
detection probability is decrease when decreased SNR.
3. CONCLUSION
In this paper we have implemented Simulink based
spectrum sensing using Cyclostationary Detection
technique. Cyclostationary feature detector
implementation can differentiate the modulated signal
from the additive noise, distinguish Primary User signal
from noise. It is used at very low SNR detection by using
the information embedded in the Primary User signal
which does not exist in the noise. The merits of the
Cyclostationary Detection technique is that it is robust in
low SNR and robust to interference, whereas the demerits
of this technique is that it requires partial information of
the primary user and that it has a high computational cost.
With Cyclostationary spectrum sensing, the primary user’s
modulation scheme can also be easily found out.
REFERENCES
[1] M. Ghozzi, M. Dohler Marx and J. Palicot.
Cognitive radio: methods for the detection of free
bands, ComptesRendus Physique, vol. 7, no. 7,
pp. 794-804, 2006.
[2] M. Bodepudi, R. C. Kolli, and R. K. Rayala,
Spectrum sensing technique and issues in
cognitive radio. International Journal of
Engineering Trends and Technology(IJETT)
Volume 4, 2013.
[3] D. Cabric, and R. W. Brodersen, Physical layer
design issues unique to cognitive radio
systems,” in Proc. IEEE PIMRC, 2005.
[4] M. António and B. C. Nuno, White Spaces
Communications in Europe ,Working Paper,
Universidade de Aveiro, Aveiro, Portugal, 2011.
[5] Hong L., Junfei, Fangmin, X. Shurong, L. and Zheng,
Z. (2008). Optimization of Collaborative Spectrum
Sensing for Cognitive Radio, in Proc. IEEE
International Conference on Networking, Sensing
and Control, ICNSC’ 08, Sanya, , pp. 1730–1733.
[6] G. Amir, and S. Elvino, Spectrum Sensing in Cognitive
Radio Networks: Requirements, Challenges and
Design Trade-offs, IEEE Communication Magazine.
Page (s): 32-39, 2008.

[7] S. Haykin, Cognitive Dynamic Systems,
Proceedings of the IEEE, vol. 94, no. 11, pp. 1910-
1911, 2006.
[8] W. Ejaz, N. Hasan, M. Asam and S. Kim,
Improved local spectrum sensing for cognitive
radio networks. EURASIP J. Adv. Signal Process,
2012.
[9] I. F. Akyildz, B. F. Lo, and R. Balakrishnan,
Cooperative spectrum sensing in cognitive radio
networks: A survey. Physical Communication, vol. 4,
no. 1, pp. 40-62, 2011.
[10] M. M. Buddhikot and K. Ryan, Spectrum
management in coordinated dynamic spectrum
access based cellular networks, in: Proc. IEEE
DySPAN 2005, pp. 299-307, 2005.
[11] Y. Helin, X. Xianzhong, and W. Ruyan, SOM- GA-SVM
Detection Based Spectrum Sensing in Cognitive
Radio. National Nature Science Foundation of
China, 2012.
[12] K. Hamdi and K. B. Letaief, Cooperative
communications for cognitive radio networks, in
8th Postgrad. Symp. Converg. Telecom., Net.
Broadcasting, Liverpool John Moores University,
2007.

IRJET- Simulating Spectrum Sensing in Cognitive Radio Network using Cyclostationary Technique

More Related Content

What's hot (19)

Similar to IRJET- Simulating Spectrum Sensing in Cognitive Radio Network using Cyclostationary Technique (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Simulating Spectrum Sensing in Cognitive Radio Network using Cyclostationary Technique