Efficiency Improvement of Cooperative Spectrum Detection in Cognitive Radio System
[email protected]

In this paper we presents a study of hard combining and decision
fusion schemes for cooperative cognitive radio to solve the problem
of inefficient use of radio frequency spectrum to attack the upcoming
spectrum crunch issue the CR detection efficiency . Simulation
comparison between hard combining and decision fusion schemes(
AND ,OR, and K / N ) for cooperative cognitive radio are achieved.
The hard combination and decision fusion schemes provides a good
tradeoff between the detection performance . The OR – rule gives a
better spectrum detection efficiency at high values of SNR and small
values of Q m d with at a different values of Q f a and adaptive
threshold (?) level .
Key words: Energy Detection, cognitive radio (CR), Cooperative
spectrum sensing , combining data fusion , ROC, Q d , Q m d ,Q f a ,
and threshold (?)
I – Introduction The wireless communication systems and other applications uses the
frequencies in the range of 3 KHz to 300 GHz ( Radio Spectrum) .
The demand for wireless communication is increasing continuously
and the radio spectrum has afinite resource . Additionally, according
to the statics of the Federal Communicatins Commission ( FCC ) ,
temporal and geographical variations in the utilization of assigned
spectrum range from 15 to 85 percent . At the present time , it has
become necessary to use the available spectrum more efficiently to
upstay further growth of wireless communications..Consequently ,
cognitive radio is a revolutionary communication paradigm to solve
the problem of inefficient use of radio frequency spectrum to attack
the upcoming spectrum crunch issue .
The temporally unused spectrum is referred to as a white space or
spectrum hole . These frequency bands ( channels ) are assigned to
specific system users called licensed users or primary users (PU) ,
and the assigned frequency bands are called licensed bands .
Cognitive radio users also called unlicensed or secondary users (SU),
who can find unused authorized spectrum hole dynamically for its
own use without causing any interference to primary users .

Figure (1) : spectrum holes concept .
So, primary users (PU) can be defined as the users who have the
authorized license on the usage of a specific part of the spectrum.
Secondary users can be defined as the users who have the
conditional license and should not cause any interference to the
primary users (PU) when using the idle channel.
There are two major subsystems in a cognitive radio : – 1. a cognitive unit that makes decisions based on various Inputs .
2. a flexible SDR unit whose operating software provides a range of
possible operating modes . In ddition to a spectrum sensing
subsystem is also often included in the architectural of cognitive
radio to measure the signal environment to determine the

presence of other users . Cognitive radio technology will
intelligently determine whether a certain part of the frequency
spectrum is idle, or if it is being utilized. If the cognitive radio can
successfully determine with a high degree of certainty that a
specific part of the spectrum is being idle, it can then transmit on
these frequencies without interfering with the licensed owner of
the spectrum, thus achieving better spectral resource efficiency.

Figure ( 2) : Mechanism of cognitive radio for efficient use of the available radio frequency spectrum
The requirement of no interference is the key for developing of
cognitive radio to invent fast and highly robust ways of determining
whether a frequency band is available or being occupied. This is the
area of spectrum sensing for cognitive radio.
Spectrum sensing will be the backbone of any autonomous
cognitive radio. Therefore, more simple and reliable spectrum
sensing technique is needed . Energy detector retrofit simplicity
and serves as apractical spectrum sensing technique .
To improve the spectrum sensing for cognitive radio network
(CRN), cooperative spectrum sensing methodology is proposed to
withstand some spectrum sensing drawbacks such as fading,
shadowing, and receiver uncertainty problems .
The aim of cooperation spectrum sensing is to improve the
cognitive radio detection performance by taking advantage of the
spatial diversity, in order to protect the primary user( PU ) against
the interference , and reduce the probability of false alarm to get an
efficient utilization of the spectrum holes .

Figure (3):Cooperative spectrum Sensing showing drawbacks
(Multi-path and Shadowing , receiver uncertainty )

In a fading environment , the spectrum detection is braved by
uncertainty due to the fading of the channel i.e. the secondary user
need to differentiate between a white space, where there is no
primary signal, and a deep fade where it is detect the primary
signal. Similar difficulties arises in the case of shadowing.
To make a treatment for these issues, many different secondary
users can cooperate with each other to detect the presence of
primary user or signal . The advantage of diversity gain
accomplished through cognitive users cooperation helps to improve
fading and shadowing effects . Cooperative detection also helps in
improving the detection performance . The CR users ( Receivers )
can measure signal properties and can even estimate what the CR
system (Transmitter) meant to send , but it also should be able to
tell the transmitter about how to change its waveform in ways that
will suppress interference . In other words , the cognitive radio
users ( receivers ) needs to convert this information into a
transmitted message to send back to the CR system ( Transmitter )
.
Cognitive Radio Main Functions : –
1 – Spectrum Sensing : –
This is the main function in CR to enable cognitive radio
users (CRs) to detect the spectrum white space and occupy the
vacant spectrum band and improve overall spectrum efficiency .
2 – Spectrum Management : –
CR decide on the best spectrum band and the channels within
the available bands to meet the QoS requirements over all
available spectrum channels i.e It captures the best available
vacant spectrum holes from detected spectrum holes.
3 -Spectrum Mobility : –
Cognitive radio networks aim to use the spectrum dynamically
by allocating the radio terminals ( CR users ) to operate in the
greatest available frequency channels i.e cognitive radio users
are considered as guests on the spectrum. So , if a particular
spectrum band is desired to assign for the primary user, the
communication has to be continued in another idle spectrum band
4 -Spectrum Sharing : –
Providing an efficient and fair dynamic spectrum allocation
schemes to distribute the primary spectrum holes to the
competitive secondary users .
Cognitive Radio Cycle model : –
The cognitive radio cycle process can be achieved through three
steps : –
1 – Spectrum Sensing .
2 – Spectrum analysis .
3 – Spectrum Decision .

Figure (4) : Cognitive Radio Cycle Model .
Cognitive radio is one of the most promising solutions to spectrum-
scarcity problem. In cognitive radio networks, the most crucial
activity is the spectrum efficiency (SE) and energy efficiency (EE)
In cognitive radio , the SU senses the spectrum to check the
presence or absence of the PU signal depending on sensing
parameters like signal-to-noise ratio (SNR), bandwidth, bit error
probability, spectral efficiency, throughput, and spectral efficiency
are beer in mind and studied.
II – Spectrum Sensing Techniques
There are several popular and top performing spectrum sensing
algorithms for cognitive radio under low signal to noise ratio
scenario, namely : – 1 – Cyclo-stationary Feature Detector ( CFD ) .
2 – Energy Detector ( ED ) .
3 – Matched Filter Detector ( MFD ) .
The accuracy of sensing is call requirement to accurate sensing
which provides successful access operation of CRN. For known
condition, Energy Detection (symbolically expressed as ED)
technique is the simplest way of sensing. The strength of received
signal energy is based on threshold value of ED. For greater energy
of received signal compared to threshold value assumes the
presence of a user, otherwise the physical channel is free.
Spectrum sensing is to differentiate between two hypotheses ,
The primary user is present, hypothesis ( H1 ) .
The primary user is absent, hypothesis ( H0 ) .

= ,
=

Where is the signal received by secondary user , is
the primary user’s transmitted signal, is the additive white
Gaussian noise (AWGN) , and is the amplitude gain of the
Channel between the PU and the kth CR user.
III – System Model
Suppose a cognitive radio network, with number of cognitive users
(K ) indexed by k = 1, 2. . . K .
Suppose each CR performs local spectrum sensing autonoumously
by using N samples of the received signal and all cooperative CR
users send their sensing results (m1, m2… mK) via the control
channel. Consequently , the FC fuses the received local sensing
information to make a final decision about the presence or absence
of the PU .
Here we express the SU received signal process in the form of

and the ongoing SU received signal can be formulated as

spu (t) is the PU transmitted signal, ssu (t) represents the leakage
from the SU transmitted signal , hpu is the PU channel gain while hsu
represents the SU leakage signal gain, and t is the time.
1

2

An energy detector is employed in each SU to determine the
state of the PU. The ED output statistic in each SU is given as ,

Where S is the number of averaged samples .
The ED output for both the hypotheses can be expressed as ,

In this paper, we will study three different hard combining and
decision rules for CSS ( OR , AND , and K / N rules) , thus deducing
the effects on the Detection Efficiency under certain conditions .
A . AND _ Rule
decides that the PU signal is present if all CR users have detected
the PU signal. The cooperative test using the AND rule can be
formulated as follows : –

Where K : Number of CR users
The probability of detection and the probability of false alarm also
formulated as follows : –

Where is the final detection . The special case for the AND- –
Rule corresponds to M = K

= 1 –

B . OR_ Rule : –
The OR rule decides that the PU signal is present if any of the CR
users detect the PU signal . Hence, the cooperative test using the
OR rule can be formulated as follows : –

Where is the final decision ,The special case for the OR-
– Rule Corresponds to the case M = 1

= 1 – ( )

C . K / N _ Rule
The third rule is called majority or Voting rule that decides the
presence of the PU signal if at least M of K users have detected
with 1 ? M ? K , and is formulated as : –

A majority decision is a special case of voting rule for M = K/2
The probability of detection Qd and probability of false
alarm Qf are defined as : –

= 1 – ( )

D . The main parameters that controls the
performance of the CR Spectrum Sensing
• Signal – to – Noise Ratio ( SNR )
• Probability of Correct Detections
Qd { decision , Y = H1 | H1 }
Qd { decision , Y = H0 | H0 }
• Probability of False Alarm : –
Qfa { decision , Y = H1 | H0 }

• Probability of Miss-Detection : –
Q m d { decision , Y = H0 | H1 }

• Total Error Rate : –
Q e = Q f a + Q m d
• Threshold ( ? ) .
• Number of CR users .
IV – Simulation and Analysis of the Results
In this paper we was proposed a coop –
-erative hard combining and decision rules
to improve the CR detection efficiency .
There are three hard combining and decision based
cooperative sensing rules are used .
The OR rule gives a decision H1 when at least one of
CR users detects PU signal .
The AND rule gives a decision H1 if all CR users send
their status detection as bit – 1 as a local detection of
the PU .
The K / N rule gives a decision H1 if at least half of
CRs Local detection status is bit – 1
Each CR user makes its own decision with respect to the
presence or absence of the PU and sends the one bit
decision status (1 or 0) to the FC or cooperative groups to
make data fusion.

Our simulations was made for Cognitive Radio
Network with cooperative Seven Secondary Users
( K = 7 SU s ) . AWGN channel also proposed for
our simulations . A SNR ranges from (-18 dB to
– 6 dB , -12 dB to 0 dB , and -10 dB to 2 dB ) with
Qfa = 0.01 .
Also SNR ranges from (-18 dB to -6 dB , -16 dB to –
4 dB , -14 dB to -2 dB , -12 dB to 0 dB , and
-10 dB to 2 dB ) with
We are using QPSK modulation for test with
modulation index m = 6 , with a number of
simulations n = 2000 for each value of SNR ,and
Number of Samples / Signal N = 1500 .

At Figure ( 7 ) : SNR ( -10 dB to 2 dB) Vs Q d , at Q f a = 0.01
0.01 .
Figure ( 5 ) : SNR ( -18 dB to -6 dB) Vs Q d , at Q f a = 0.01
Figure ( 6 ) : SNR( -12 dB to 0 dB ) Vs Q d , at Q f a = 0.01

Figure ( 5 ) , Figure ( 6 ) , and Figure ( 7 ) shows the
receiver operating characteristic (ROCs) curves for
the hard decision fusion rules ( AND , OR , and K / N)
and non-cooperative through energy detector and
with a Qfa is 0.01 .
The last three(ROCs) curves shows that the
probability of detection increases as the SNR
increased . Also the (ROCs) curves shows that the
“OR” rule detection performance is the best one for
spectrum detection than the other hard decision
fusion rules. Also, the majority or “HV” hard decision
fusion rule is lower in the detection efficiency than
the ” OR – Rule ” , but is better than the ” AND –
Rule ” , and the ” AND – Rule ” is better in
performance than non-cooperative CR users .
Consequently , we will made the same simulation at
the same conditions , but with probability of false
alarm Qfa is 0.1 (Qfa increased ) .
Figure ( 8 ) : SNR ( -18 dB to – 6 dB) Vs Q d , at Q f a = 0.1
.

Figure ( 10 ) : SNR ( -10 dB to 2 dB) Vs Q d , at Q f a = 0.1

It is clear from (ROC s) curves in ( figures 8 , 9 ,
and 10 ) that the probability of detection Qd
increases by increasing the probability of False
Alarm Qfa in the same time of increasing the
SNR .
We can note that from (ROC) curve in
figure ( 5 ) with SNR range ( -18 dB to -6 dB) at
the point of SNR = – 10 dB with Q f a = 0.01 :-
the Q d – OR = 83 % , Q d – K/N = 44 % ,
Q d – AND = 20 % , Q d – Non-cooperative = 9 %

Also , We can note that from (ROC) curve in
figure ( 8 ) with the same SNR range ( -18 dB to
– 6 dB) as in fig. ( 5 ) at the point of SNR = – 10
dB , but with Q f a = 0.1 : –
the Q d – OR = 99 % , Q d – K/N = 88 % ,
Q d – AND = 77 % , Q d – Non-cooperative = 68%
So,it is obvouies that the Q d increases
by increasing the Q f a and SNR.
Consequently the ” OR – Rule ” gives
abetter efficiency of the spectrum
sensing detection .
Figure ( 9 ) : SNR ( -12dB to 0 dB ) Vs Q d , at Q f a = 0.1

From the last results we can confirm the relation
between the Q d , Q f a , and SNR and their
effets on the spectrum detection efficiency .
Also we know that there a relation
between the Q d and Q m d as follows : –
Q m d = 1 – Q d
Q d have a relation to Q f a , and SNR
hence there should arelation between
Q m d and Q f a as mentioned in eq(.)
So , we will simulate the effect of the
relation between Q f a , Q m d for the hard
decision fusion rules ( OR , AND , and K / N )
and Non-cooperative CR network , and their
effets on the spectrum detection efficiency .
To simulate the relation between Q m d
and Q f a .
we analyze the spectrum detection
efficiency under the target of the
probability of miss -detection and
probability of false alarm at ( K = 7 SU s )
SNR = – 10 dB , time bandwidth factor U = 100 ,
and probability of false alarm is used from 0.01 to
1 by increasing 0.01 and AWGN channel
considered .

Figure ( 11) : Q f a Vs Q m d , at SNR = -10dB
So, from Figure ( 9 ) , we can note that the
OR rule gives a minimum Q m d Versus
values of Q f a when compared to the
other cooperative spectrum sensing
techniques ( K / N or Majority rule , AND
rule ) . So the OR rule is best among hard
combination data fusion for cooperative
spectrum sensing in Cognitive Radio and
gives the better performance than ( K/N or
Majority rule , AND rule).
Since the aim of cognitive radio
cooperative spectrum sensing is to
improve the detection performance and
protect the primary user( PU ) against the
interference which results from the large
values of Q m d , and reduce the probability
of false alarm to get an efficient utilization
of the spectrum holes .
Consequently , we want to keep the
probability of missed detections Q m d very
low, so the probability of false alarms Q f a
increases and this would result in low
spectrum utilization. This Parallelize
a low probability of false alarms would
result in high missed detection probability
which increases the interference to the
primary users. This trade-off has to be
carefully considered.
Then the threshold is set in order to achieve a
constant level of false alarm to satisfy the
condition of minimum Q m d . Then the
threshold level (?) is raised and lowered
during detection in order to maintain
acceptable level of probability of false alarm
Q f a . Consequently we can formulate this
meaning as Adaptive Threshold – Constant
False alarm Rate (AT -CFAR) detection .

Figure ( 12 ) : Threshold (?) Vs Q f a

Number of Samples / Signal N = 1000 , Noise only
received i.e PU – absent , Q f a = 0.01 : 0.01 : 1
It is clear that from ROC for Threshold (?) Vs Q f a
The Q f a decreases as the the Threshold level (?)
Increased .
So , our reults satisfy the proper condition
for designing a cognitive radio network
with cooperative spectrum sensing with
high spectrum detection with minimum
interference and more efficient spectrum
utilization .

V – Conclusion
In this paper we was presented a study
and simulation of various effective
cognitive radio hard combining
cooperative spectrum sensing techniques
and signal detection base on hard decision
combining technique in data fusion centre
compared with non-cooperative one .
In cooperative technique, OR and AND ,
and majority or “HV” rules are employed
and evaluate the system performance by
using probability of detection (Q d) ,Q f a ,
Q m d , and SNR .
We was proved that by simulation the
probability of detection of hard combining
cooperative spectrum sensing techniques
better performance compared with non-
cooperative one , and increases as the SNR
increased . The (ROCs) curves shows that
the “OR” rule detection performance is
the best one for spectrum detection than
the other hard decision fusion rules. Also,
the majority or “HV” hard decision fusion
rule is lower in the detection efficiency
than the ” OR – Rule ” , but is better than
the ” AND – Rule ” , and the ” AND –
Rule ” is better in performance than non-
cooperative CR users . Also ,it is obvouies
that the Q d increases by increasing the Q f a
and SNR. Furthermore, a minimum of 7
cooperated users relatively in cognitive
radio system can achieve optimal value of
probability of detection. However, it
depends on the threshold (?) value used in
spectrum detection .