Chandrasekhar Korumilli, Chakrapani Gadde, I.

Hemalatha

/ International Journal of
Engineering Research and Applications (IJERA) ISSN: 2248-9622
www.ijera.com Vol. 2, Issue 4, July-August 2012, pp.1004-1009

1004 | P a g e
Performance Analysis of Energy Detection Algorithm in
Cognitive Radio

Chandrasekhar Korumilli
1
, Chakrapani Gadde
2
, I.Hemalatha
3
(
1
P.G.Student Department of Electronics and Communications, Sir CRR College of engineering, Eluru)
(
2
P.G.StudentDepartment of Electronics and Communications, Sir CRR College of engineering, Eluru)
(
3
Associate prof Department of Electronics and Communications, Sir CRR College of engineering, Eluru)

ABSTRACT
In cognitive radio systems, secondary
users should determine correctly whether the
primary user is absent or not in a certain
spectrum within a short detection period.
Spectrum detection schemes based on fixed
threshold are sensitive to noise uncertainty; the
energy detection based on dynamic threshold
can improve the antagonism of noise
uncertainty; get a good performance of
detection while without increasing the computer
complexity uncertainty and improves detection
performance for schemes are sensitive to noise
uncertainty in lower signal-to-noise and large
noise uncertainty environments. In this paper
we analyze the performance of energy detector
spectrum sensing algorithm in cognitive radio.
By increasing the some parameters, the
performance can be improved as shown in the
simulation results.
Keywords - cognitive radio, detection threshold,
dynamic threshold detection, noise uncertainty

1. INTRODUCTION
With the development of a host of new
and ever expanding wireless applications and
services, spectrum resources are facing huge
demands. Currently, spectrum allotment is done by
providing each new service with its own fixed
frequency block. As more and more technologies
are moving towards fully wireless, demand for
spectrum is enhancing. In particular, if we were to
scan the radio spectrum, including the revenue-rich
urban areas, we find that some frequency bands in
the spectrum are unoccupied for some of the time,
and many frequency bands are only partially
occupied, whereas the remaining frequency bands
are heavily used [1]. It indicates that the actual
licensed spectrum is largely under-utilized in vast
temporal and geographic dimensions [2].A remedy
to spectrum scarcity is to improve spectrum
utilization by allowing secondary users to access
under-utilized licensed bands dynamically
when/where licensed users are absent.
Cognitive radio is a novel technology
which improves the spectrum utilization by
allowing secondary users to borrow unused radio
spectrum from primary licensed users or to share
the spectrum with the primary users. A cognitive

radio is able to able to fill in the spectrum holes and
serve its users without causing harmful interference
to the licensed user. To do so, the cognitive radio
must continuously sense the spectrum it is using in
order to detect the re-appearance of the primary
user [3]. Once the primary user is detected, the
cognitive radio should withdraw from the spectrum
instantly so as to minimize the interference. This is
very difficult task as the various primary users will
be employing different modulation schemes, data
rates and transmission powers in the presence of
variable propagation environments and interference
generated by other secondary users [1].
The paper is organized as follows. Section 2
discuss the spectrum sensing problem, overview of
spectrum sensing methods and the performance of
energy detector spectrum sensing algorithm in
cognitive radio. Section 3 discusses the
performance of dynamic threshold based spectrum
detection in cognitive radio systems. Section 4
discusses the conclusion and future in this field of
study.

2. SPECTRUM SENSING PROBLEM

2.1 ENERGY DETECTION
Spectrum sensing is a key element in
cognitive radio communications as it must be
performed before allowing unlicensed users to
access a vacant licensed band. The essence of
spectrum sensing is a binary hypothesis-testing
problem

H
0
: X (N) =W (N)
H
1
: X (N) =S (N) +W (N) (1)

Where N is the number of samples,
N=2TW, T is duration interval ,W is bandwidth, S
(N) is the primary users signal, W (N) is the noise
and X (N) is the received signal. H
0
and H
1
denote
that the licensed user is present or not, respectively.
The noise is assumed to be additive white Gaussian
noise (AWGN) with zero mean and is a random
process. The signal to noise ratio is defined as the
ratio of signal power to noise power
Chandrasekhar Korumilli, Chakrapani Gadde, I.Hemalatha

/ International Journal of
Engineering Research and Applications (IJERA) ISSN: 2248-9622
www.ijera.com Vol. 2, Issue 4, July-August 2012, pp.1004-1009

1005 | P a g e
Under the assumption of absolutely no
deterministic knowledge about the signal X (n),
i.e., we assume that we know only the average
power in the signal. In this case the optimal
detector is energy detector or radiometer can be
represented as [23]

( ) ( ) ( ) > =


=
n X n Y
N
Y D
N
n
1
0
1

1
H

<
0
H

(2)

Where D(Y) is the decision variable and is
the decision threshold, N is the number of samples.
If the noise variance is completely known, then
from the central limit theorem the following
approximations can be made [24]


( ) ( ) N N H Y D
n n
4 2
0
2 , ~ o o


( ) ( ) N P P N H Y D
n n
2 2 2
1
) ( 2 , ~ o o + +
(3)
Where P is the average signal power and
2
n
o is the
noise variance. Using these approximations
The probability expressions are
( )
|
|
|
.
|

\
|

= > =
N
Q H Y D P P
n
n
r FA
4
2
0
2
) (
o
o

(4)
( )
|
|
|
.
|

\
|
+
+
= > =
N P
P
Q H Y D P P
n
n
r D
2 2
2
1
) ( 2
) (
) (
o
o

(5)


|
|
|
.
|

\
|
+
+
= =
N P
P
Q P P
n
n
D Md
2 2
2
) ( 2
) (
1 1
o
o
(6)
Where Q ( ) is the standard Gaussian
complementary cumulative distribution function
(CDF). P
D
, PF
A
and P
Md
represent detection
probability, false alarm probability and missing
probability respectively.

From (4) and (5) eliminating threshold




( ) ( ) | | ( )
2
2
1 1
2

= SNR P Q P Q N
D
FA
(7)
Where
2
2
,
n
n
P
SNR o
o
= is the normalized noise
power.
Figure 1 shows the numerical results of
(7) for given P
FA
(0.0.9), sample number N=500,
with different SNR values with that the
performance is improved by increasing SNR value

Figure 1 ROC curves of energy detection
scheme with different SNR

Figure 2 is the numerical results of (7) for
given P
FA
(0.0.9), SNR=-15dB. It shows that the
performance is improved by increasing N, and
probability of detection can be improved by
increasing N value even if the SNR is much lower,
as long as N is large enough without noise
uncertainty.


Figure 2 ROC curves of energy detection
scheme with different N







Chandrasekhar Korumilli, Chakrapani Gadde, I.Hemalatha

/ International Journal of
Engineering Research and Applications (IJERA) ISSN: 2248-9622
www.ijera.com Vol. 2, Issue 4, July-August 2012, pp.1004-1009

1006 | P a g e
2.2 NOISE UNCERTAINTY

Now, considering the case with uncertainty in the
noise model [20], the distributional uncertainty of
noise can be represented as

] , / [
2 2 2
n n
o o o e
is the noise uncertainty coefficient and >1

Now

|
|
.
|

\
|

=
N
Q P
n
n
FA
/ 2
2
2
o
o
(8)

|
|
.
|

\
|
+
+
=
N P
N P
Q P
n
n
D
/ 2 ) / (
) / (
2
2
o
o
(9)


eliminating threshold

and equating both
equations we have

( ) ( ) | | ( )
2
2
1 1
) / 1 ( ) / 1 ( 2

+ = SNR P Q SNR P Q N
D
FA
(10)

Comparing (10) with (7), there is almost
no contribution to the whole expression results if
there is a tiny change of ; however, SNR
2
and
(SNR ( 1/))
2
should be mainly discussed
and compared. When 1, then SNR2 (SNR
( 1/))
2
, the numerical value of (10) and (7) are
almost the same; When is larger and suppose =
1.05, then ( 1/) = 0.0976 0.1, if SNR = 0.1,
well then (SNR ( 1/))
2
0, substituting into
equation (10) to be N . In other words, only
infinite detection duration can complete detection,
which is impracticable. A tiny fluctuation of
average noise power causes performance drop
seriously, especially with a lower SNR.
Figure 3 shows the numerical result of
(10) probability of false alarm on X-axis and
probability of detection on Y-axis for an SNR=-
15dB, P
FA =
(0, 0.9), N=500 and varying the noise
uncertainty value.
From the Figure it is seen that the
performance gradually drops as the noise factor
increasing. This indicates that Energy detector is
very sensitive to noise uncertainty. It means that
cognitive users predict the spectrum to be idle no
matter whether there are primary users present or
absent. Consequently, cognitive users are harmful
to licensed users when primary users are present.
This situation often occurs in cognitive radio
systems, particularly in lower signal-to-noise ratio
environments. In order to guarantee a good
performance, choosing a suitable threshold is very
important. Traditional energy detection algorithms
are based on fixed threshold and we have verified
that performance decreased under noise uncertainty
environments. This indicates that the choice of a
fixed threshold is no longer valid under noise
uncertainty and threshold should be chosen flexible
based on the necessities.


Figure 3 ROC curves of energy detection
scheme with different
3. DYNAMIC THRESHOLD
Performance of cognitive radio declined
sharply due to noise uncertainty and cognitive
users accessing will be serious interference to
licensed users. This should be avoided in dynamic
spectrum access technology. For this reason, a new
algorithm combating the noise uncertainty is
presented [21][20].
Assume that the dynamic threshold factor
and

> 1 the distributional of dynamic threshold

\
|

=
N
P
n
n
FA
/ 2
'
2
2
o
o
(11)


( )
( )
|
|
|
|
|
.
|

\
|
+
+
=
N
P
P
P
n
n
D
2
2
2
o
o
(12)


After simplifying (11) and (12) we get the
relationship for SNR, N,
FA
P ,
D
P and '


( ) ( ) ( ) | |
( ) ( )
2
2 ' 2 '
2
1 2 1
1
1 ' 2

+
+ =


SNR
P Q SNR P Q N
D FA
(13)


Chandrasekhar Korumilli, Chakrapani Gadde, I.Hemalatha

/ International Journal of
Engineering Research and Applications (IJERA) ISSN: 2248-9622
www.ijera.com Vol. 2, Issue 4, July-August 2012, pp.1004-1009

1007 | P a g e
Figure 4 shows the performance of energy
detection scheme, probability of false alarm on X-
axis and probability of detection on Y-axis.

Figure 4 ROC curves of energy detection
scheme with no noise uncertainty, with noise
uncertainty, and with dynamic threshold

Here we have taken a SNR=-15dB, ( ) 09 . 0 , 0 e
FA
P ,
N=1500, noise uncertainty1.02 and dynamic
threshold1.001.It is observed that the performance
is improved by using a dynamic threshold

3.1 NOISE UNCERTAINTY AND DYNAMIC
THRESHOLD

We have discussed two cases that existing
noise uncertainty and dynamic threshold
respectively, this section will give the expressions
that considering noise uncertainty and dynamic
threshold together, we got expressions of false
alarm probability and detection probability

The noise variance in the interval
(

e
2
2
,
n
n
o

o
o
Now the probability relations are represented as

|
|
|
.
|

\
|

=
N
Q P
n
n
FA
2
'
2
2
o
o
(14)

|
|
|
|
|
|
.
|

\
|
|
|
.
|

\
|
+
|
|
.
|

\
|
+
=
N
P
P
Q P
n
n
D
2
'
2
2

o

o

(15)
Eliminating threshold
we get the inter relationship for SNR, N,
FA
P ,
D
P
and '

( ) ( ) ( ) ( )
2
2
1 1
'
'
'
1
' ' 2

|
|
.
|

\
|
+
(

|
|
.
|

\
|
+ =

SNR
P Q SNR P Q N
D FA
(16)

In (16), when ' and ' / ' 1,
( )
2
' ' '

+ SNR (SNR)
2
and
' (1/+SNR) (1+SNR). We substitute (16) with
the above approximate unequal expressions, and
we can get that the numerical value of (16) is
almost the same to (7). Therefore, dynamic
threshold detection algorithm can overcome the
noise uncertainty as long as a suitable dynamic
threshold factor is chosen. Comparing (16) with
(13), supposing SNR = 0.1 and ' and both
closer to 1, it is clear that
( ) ( ) ( )
2 2
1 ' ' '

>> + SNR SNR .
Consequently, detection duration N has been
shortened largely to N= 500 with the same
probability parameters P
D
and P
FA
as shown in
Figure 5 It can be concluded that as long as the
dynamic threshold factor is suitable, even if there is
noise uncertainty, we can get a better spectrum
performance. To attaining the same performance,
the detection time of dynamic threshold energy
detection Algorithm is less than the traditional
version.
Figure 5 is the numerical results of (7),
(13) and (16). With the same parameters as before.


Figure 5 ROC curves of energy detection
scheme with N=500, different and '
Chandrasekhar Korumilli, Chakrapani Gadde, I.Hemalatha

/ International Journal of
Engineering Research and Applications (IJERA) ISSN: 2248-9622
www.ijera.com Vol. 2, Issue 4, July-August 2012, pp.1004-1009

1008 | P a g e

Where = 1.00 denotes that the average
noise power keeps constant (without noise
uncertainty); ' = 1.00 denotes that the algorithm
did not use dynamic threshold (the threshold is
fixed); otherwise, it represents cases with noise
uncertainty and dynamic threshold. From Figure 5
it indicates that a tiny fluctuation of average noise
power causes a sharp decline in detection
performance. The dynamic threshold makes the
performance improve significantly as the dynamic
threshold factor increasing. If a suitable dynamic
threshold factor is selected, the falling proportion
of performance caused by noise uncertainty can be
omitted and the performance may be more
accurate.

4. CONCLUSION
Energy detection based on fixed threshold
are sensitive to noise uncertainty, a fractional
change of average noise power causes decreasing
the performance quickly. According to the
drawback in Matched filter which not sensitive to
noise uncertainty, by using dynamic threshold the
performance can be improved as compared with the
fixed threshold. The computer simulations of the
dynamic threshold based energy detection
algorithm in cognitive radio improve the detection
performance but in practical how acquire the
detection threshold and how to improve the
detection performance by other sensing methods.

REFERENCES
[1] S.Haykin,Cognitive radio:brain-empowered
wireless communications, IEEE
J,Select.Areas in Commun, vol. 23, no. 2,
pp. 201220, february 2005.
[2] V. K.Bhargava and E. Hossain, Cognitive
Wireless Communication Networks, 1st ed.
Springer, 2007.
[3] J.Mitola, Cognitive radio an integrated
agent architecture for software defined,
Ph.D. dissertation, KTH Royal Institute of
Technology, Stockholm,Sweden, 2000.
[4] R. RUBENSTEIN, Radios get smart,
IEEE Spectrum,consumer electronics, pp.
4650, February 2007.
[5] J.Mitola, Cognitive radio:making software
radio more personal, IEEE Personal
Communications, vol. 6, no. 4, pp. 1318,
1999.
[6] B. A.Fette, Cognitive Radio, 1st ed. Newnes,
2006.
[7] FCC, Notice of proposed rulemaking and
order, ET, Docket 03-222, December 2003.
[8] J. O. Neel, Analysis and design of cognitive
radio networks and distributed radio
resource management algorithms, Doctoral
Dissertation, Virginia Polytechnic Institute
and State University, September 2006.
[9] FCC, Spectrum policy task force,
Technology Advisory Council
(TAC),Briefing, Tech. Rep., 2002.
[10] Z. L. W. H. S. J. Carl R.Stevenson, Gerald
Chouinard and W. Caldwell, Ieee 802.11
the first cognitive radio wireless regional
area network standard, IEEE Standards in
Communications, pp. 130138, January
2009
[11] T. Yucek and H. Arsla, A survey of
spectrum sensing algorithms for cognitive
radio applications, IEEE Communications
Surveys & Tutorials, vol. 11, no. 1, pp. 116
130, March 2009.
[12] X. C.X.Wang, H.H.Chen and M.GuiZani,
Cognitive radio network management,
IEEE Vehicular Technology Magazine, vol.
3, no. 1, pp. 2835, 2008.
[13] W. Gardner and C. Spooner, Signal
interception: performance advantages of
cyclic-feature detectors, IEEE Transactions
on Communications, vol. 40, no. 1, pp. 149
159, January 1992.
[14] H. P. Zhi Quan, Shuguang Cui and A.
Sayed, Collaborative wideband sensing for
cognitive radios, IEEE Signal Processing
Magazine, vol. 25, no. 6, pp. 6073,
November 2008.
[15] Y. Zeng and Y.-C. Liang, Covariance based
signal detections for cognitive radio, 2nd
IEEE International Symposium on New
Frontiers in Dynamic Spectrum Access
Networks, pp. 202207, April 2007.
[16] K.B.Letaief and W. Zhang, Cooperative
communications for cognitive radio
networks, Proceedings of the IEEE, vol. 97,
no. 5, pp. 878893, May 2009.
[17] H.Urkowitz, Energy detection of unknown
deterministic signals, Proceedings of the
IEEE, vol. 55, no. 4, pp. 523531, April
1967.
[18] M. F.F.Digham and M.K.Simon, On energy
detection of unknown signal over fading
channel, Proceedings of IEEE International
Conference on Communications (ICC03),
pp. 35753579, May 2003.
[19] A.H.Nuttal, Some integrals involving qm
function, IEEE Transaction on Information
Theory, vol. 21, no. 1, pp. 9596, January
1975.
[20] T. L. Jinbo Wu, Guicai and G. Yue, The
performance merit of dynamic threshold
energy detection algorithm in cognitive
radio systems, The 1st International
Conference on Information Science and
Engineering (ICISE2009), IEEE Computer
Society, 2009.
Chandrasekhar Korumilli, Chakrapani Gadde, I.Hemalatha

/ International Journal of
Engineering Research and Applications (IJERA) ISSN: 2248-9622
www.ijera.com Vol. 2, Issue 4, July-August 2012, pp.1004-1009

1009 | P a g e
[21] H. L. Gui-cai Yu, Yu-bin Shao and G. xin
Yue, Dynamic threshold based spectrum
detection in cognitive radio systems,
Wicom09 5th International Conference on
Wireless Communs,Networking and Mobile
Computing, 2009.
[22] R. Tandra and A. Sahai, SNR walls for
signal detection, EEE Journal of Selected
Topics in Signal Processing, vol. 2, no. 1,
pp. 417, February 2008.
[23] R. Tandra, Fundamental limits on detection
in low SNR, Masters thesis, University of
California, Berkeley, 2003.
[24] S.M.Kay, Fundamentals of Statistical Signal
Processing: Detection Theory. Englewood
Cliffs: Prentice-Hall, 1998, vol. 2.



Authors
1
Chandrasekhar Korumilli
received his B.Tech from
Pragati Engineering College
in 2008. Currently he is
pursuing his M.Tech from SIR C.R.Reddy
College of engineering college, Eluru west
Godavari district, Adhrapradesh. His areas of
interest are wireless communications, computer
networking and Cognitive Radio. He has published
a paper in international journals relating to
Throughput Improvement in Wireless Mesh
Networks by Integrating with Optical Network
and presented papers in many international
conferences.



2
Chakrapani Gadde
received his B.Tech from
Lakireddy Balireddy College
of engineering in 2008.
Currently he is pursuing his
M.Tech from SIR C R
Reddy College of
engineering college, Eluru west Godavaridistrict,
Andhrapradesh. His areas of interest are wireless
communications and computer networking. He has
published papers in international journals relating
to Mitigation of near far effect in GPS receivers.


3
I.Hemalatha received B.Tech from Sir CRR
College of Engineering, M.Tech from JNTU
Kakinada. Currently she is working as an Associate
Prof at Sir C R Reddy College of Engineering,
Eluru west Godavari district, Andhrapradesh. She
is having Twelve years of teaching experience and
guided many students during this period. Her areas
of interest are Speech Processing and wireless
communications. Published papers in many
international journals and Conferences.