Comparison of Energy Detection Based Spectrum Sensing Methods over Fading Channels in Cognitive Radio

Komal Arora, Ankush Kansal & Kulbir Singh
Signal Processing: An International Journal (SPIJ), Volume (5) : Issue (2) : 2011 44
pComparison of Energy Detection Based Spectrum Sensing
Methods Over Fading Channels in Cognitive Radio
Komal Arora erkomalarora@yahoo.com
Department of Electronics and Communications,
Thapar University,
Patiala, 147004, India.
Ankush Kansal akansal@thapar.edu
Thapar University,
Kulbir Singh ksingh@thapar.edu
Thapar University,
Abstract
With the advance of wireless communications, the problem of bandwidth scarcity has become more
prominent. Cognitive radio technology has come out as a way to solve this problem by allowing the
unlicensed users to use the licensed bands opportunistically. To sense the existence of licensed
users, many spectrum sensing techniques have been devised. This paper presents the energy
detection based spectrum sensing technique. In the present work, the comparison of ROC curves has
been done for various wireless fading channels using squaring and cubing operation. The cubing
operation shows an improvement of up to 0.6 times as compared to the squaring operation for AWGN
channel. For Rayleigh channel, the improvement achieved is up to 0.4 times as we move from
squaring to cubing operation in an energy detector.
Keywords: Cognitive Radio, Spectrum Sensing, Probability of Detection.
1. INTRODUCTION
The progressive growth of the wireless communications, has led to under-utilization of the spectrum.
It has also been observed that major portion of the spectrum is rarely used as it is reserved only for
the licensed users while other is heavily used. Federal Communications Commission (FCC) proposed
the solution to this problem by allowing the unlicensed users to use the licensed bands
opportunistically and named it as Cognitive Radio (CR) [1]. Cognitive Radio is a sophisticated
wireless system that gathers the information about the surrounding environment and adapts its
transmission parameters accordingly [2]. There are two main characteristics of Cognitive Radio:
Cognitive Capability (i.e. gathering the information about the environment) and reconfigurability (i.e.
adapting its transmission parameters according to the gathered information) [3]. Since the unlicensed
secondary (users) are allowed to utilize a licensed band only when they do not cause interference to
the licensed (primary) users, spectrum sensing is considered as one of the most important elements
of the Cognitive Radio. Spectrum sensing aims at monitoring the usage and the characteristics of the
covered spectral band(s) and is thus required by the secondary users both before and during the use
of licensed spectral bands [1].
Energy detector based approach, also known as radiometry or periodogram, is one of the popular
methods for spectrum sensing as it is of non-coherent type and has low implementation complexity. In
addition, it is more generic as receivers do not require any prior knowledge about the primary user’s
signal [4]. In this method, the received signal’s energy is measured and compared against a pre-
defined threshold to determine the presence or absence of primary user’s signal. Moreover, energy
detector is widely used in ultra wideband (UWB) communications to borrow an idle channel from
licensed user. Detection probability ( ), False alarm probability ( ) and missed detection probability
( ) are the key measurement metrics that are used to analyze the performance of an

energy detector. The performance of an energy detector is illustrated by the receiver operating
characteristics (ROC) curve which is a plot of versus or versus [5].
This paper is organized as follows: Section 2 describes the performance analysis of energy detector.
Section 2.1 illustrates test statistic for energy detection using squaring operation. Section 2.2 and 2.3
describe the expressions for probability of detection using squaring operation for AWGN and Rayleigh
channel respectively. Simulation Results for squaring operation and cubing operation are presented in
Section 2.4 followed by conclusions in section 2.5.
2. PERFORMANCE ANALYSIS OF ENERGY DETECTOR
Energy detector is composed of four main blocks [6]:
1) Pre-filter.
2) A/D Converter (Analog to Digital Converter).
3) Squaring Device.
4) Integrator.
FIGURE 1: Block Diagram of Energy Detector [6].
The output that comes out of the integrator is energy of the filtered received signal over the time
interval and this output is considered as the test statistic to test the two hypotheses and [7].
: corresponds to the absence of the signal and presence of only noise.
: corresponds to the presence of both signal and noise.
2.1 Test Statistic Using Squaring Operation
Considering the following notations:
is the transmitted signal waveform, is the received signal waveform, is in-phase noise
component, is quadrature phase component, is noise bandwidth, is power-spectral
density (two-sided), is power spectral density (one-sided), is the sampling interval, is the signal
energy, is decision threshold.
The received signal is filtered by a pre-filter which is a band-pass filter. The filtered signal is then
passed through A/D converter i.e. converted to samples. Now, if noise is a band-pass random
process, its sample function can be written as [8, Eq. (5.4)]:
where is angular frequency. If is confined to a bandwidth of and has a power-spectral
density , then and can be considered as two low-pass processes with bandwidth less
than and power spectral density of each equal to . Now, if a sample function has Bandwidth
and duration , then it can be described approximately by a set of sample values 2 or degrees of
freedom will be 2 . Thus, and each will have degrees of freedom, d equal to [7].
Also, using approximation as in [9, Eq. (2.1-21)]:
As and are considered as low-pass processes, therefore according to sampling
theorem, can be written as [10]:
Pre-Filter Squaring
Device
IntegratorA/D

where and is a Gaussian random variable with zero mean and variance
, .
Now, using the fact as in [7],
Therefore using (3) and (4), we obtain:
As has degrees of freedom over the interval [7], therefore
And the integral over the interval can be written as
Similarly,
Substituting and in (7) and (8), and using (2), we arrive at [7]:
Similarly, considering transmitted signal as a band-pass process [8], we have
or,
where , and is the signal energy.
The output of the integrator is .Test statistic can be or any quantity monotonic
with . Taking as the test statistic [7]:
Now, Under Hypothesis , the received signal is only noise i.e. , therefore using (9) test
statistic can be written as:

Thus, Test statistic under is chi-square distributed [9] with degrees of freedom or
[11] .
Under Hypothesis , received signal is the sum of signal and noise i.e. . Again
considering as a band-limited process [8], using equations (2-10), we arrive at [7]:
Then, using (12) and (14), test statistic can be written as:
Thus, test statistic under has a non-central chi-square distribution [9] with degrees of
freedom and a non-centrality parameter given by [7]. Now, Defining Signal to Noise Ratio, in
terms of non-centrality parameter as in [11]:
Thus, test statistic under : [11]. Also, probability density function of can be
expressed as [ 9, Eq. (2.3-21) & Eq. (2.3-29) ]:
2.2 Probability Of Detection And False Alarm For AWGN Channel
Probability of detection and false alarm can be evaluated respectively by [11]:
where is the decision threshold. Also, can be written in terms of probability density function
as:[12, Eq. (4-16) & Eq. (4-22)]
Using (19),
Dividing and multiplying the R.H.S. of above equation by , we get
Substituting , and changing the limits of integration to , we get
or,

where is the incomplete gamma function [13]. Now, Probability of detection can be written by
making use of the cumulative distribution function [12, Eq. (4.22)].
The cumulative distribution function (CDF) of can be obtained (for an even number of degrees of
freedom which is in our case) as [9, Eq. (2.1-124)]:
Therefore, using (25) and (26), probability of detection for AWGN channel is [11]:
Using (16),
2.3 Probability Of Detection And False Alarm For Rayleigh Channel
Probability density function for Rayleigh channel is [12, Eq. (4-44)]:
The Probability of detection for Rayleigh Channels is obtained by averaging their probability density
function over probability of detection for AWGN Channel [11]:
where is the probability of detection for Rayleigh channel.
With (28) and (29), (30) becomes
Now, substituting in (31), we get
Considering the result [14]
Comparing (32) and (33), , , ,
Thus, using (33), Probability of detection for Rayleigh channel can be expressed as [11]:

2.4 Simulation Results
The performance of energy detector is analysed using ROC (Receiver operating characteristics)
curves for fading channels. Monte-Carlo method is used for simulation. It can be seen in the following
figures that with increase in SNR (Signal to Noise Ratio), the performance of energy detection
improves. FIGURE 2 and FIGURE 4 illustrates the ROC curves using squaring operation for AWGN
and Rayleigh channel respectively. FIGURE 3 and FIGURE 5 depicts improvement in the
performance of energy detector using cubing operation over AWGN and Rayleigh channel
respectively. We assume time-bandwidth product=5.
10
-4
10
-3
10
-2
10
-1
10
0
10
-4
10
-3
10
-2
10
-1
10
0
Probability of False alarm
ProbabilityofMissedDetection
SNR=-5dB
SNR=0dB
SNR=5dB
10
-4
10
-3
10
-2
10
-1
10
0
10
-4
10
-3
10
-2
10
-1
10
0
ProbabilityofMissedDetection SNR=-5dB
SNR=0dB
SNR=5dB
FIGURE 2: Complementary ROC Curves for FIGURE 3: Complementary ROC Curves for
AWGN using Squaring operation. AWGN using Cubing operation.
10
-4
10
-3
10
-2
10
-1
10
0
10
-4
10
-3
10
-2
10
-1
10
0
SNR=5dB
SNR=10dB
SNR=15dB
10
-4
10
-3
10
-2
10
-1
10
0
10
-4
10
-3
10
-2
10
-1
10
0
SNR=5dB
SNR=10dB
SNR=15dB
FIGURE 4: Complementary ROC for Rayleigh FIGURE 5: Complementary ROC for Rayleigh
using Squaring operation. using Cubing operation.
The results obtained using cubing operation show an improvement of roughly one order of magnitude
as compared to the energy detection method illustrated in [11]. The results obtained are quantified as
shown in TABLE 1 and TABLE 2. These results illustrate improvement in probability of detection using
cubing operation. This improvement has gone up to 0.4 times for Rayleigh Channel and 0.6 times for
AWGN Channel. We assume time-bandwidth product=5 and Average SNR=5dB.

Probability of
False Alarm
Probability of detection
for AWGN Channel
(Squaring Device)
for AWGN Channel
(Cubing Device)
Improvement
(times)
0.0001 0.5372 0.8656 0.6113
0.0441 0.8792 0.9594 0.0912
0.1681 0.9390 0.9758 0.0392
0.3721 0.9748 0.9862 0.0117
0.6561 0.9938 0.9952 0.0014
TABLE 1: Improvement using cubing operation for AWGN channel.
Probability of
False Alarm
for Rayleigh Channel
(Squaring Device)
for Rayleigh Channel
(Cubing Device)
Improvement
(times)
0.0001 0.6516 0.9746 0.4957
0.0441 0.4096 0.9982 0.0146
0.1681 0.9976 0.9996 0.0020
0.3721 0.9998 1.000 0.0002
0.6561 1.000 1.000 0
TABLE 2: Improvement using cubing operation for Rayleigh channel.
2.5 Conclusions
In the present work, the performance of energy detector is analysed. Closed form expressions for
Probability of detection and false alarm over AWGN and Rayleigh channels are described. Using
ROC Curves, it is shown that the Probability of detection is improved if cubing operation is used
instead of squaring operation. Energy detection has the advantage of low implementation and
computational complexities.
3. REFERENCES
[1] J. Ma, G. Y. Li, B.H. Juang. “Signal Processing in Cognitive Radio.” Proceedings of the IEEE,
vol. 97, pp. 805-823, May 2009.
[2] H. A. Mahmoud, T. Yucek and H. Arslan. “OFDM For Cognitive Radio- Merits and Challenges.”
IEEE wireless communications, vol. 16, pp. 6-15, April 2009.
[3] I. F. Akyildiz, W.Y.Lee, M.C. Vuran, S. Mohanty. “Next Generation/Dynamic Spectrum
Access/Cognitive Radio Wireless Networks: A Survey.” Comp. Net. J., vol. 50, pp. 2127–59,
Sept. 2006.
[4] T. Yucek and H. Arslan. “A Survey of Spectrum Sensing Algorithms for Cognitive Radio
Applications.” IEEE Communications Surveys & Tutorials, vol. 11, pp.116-130, Jan. 2009.
[5] S. Atapattu, C. Tellambura, and H. Jiang. “Energy detection of primary signals over η-µ fading
channels.” in Proc. Fourth International Conference on Industrial and Information Systems,
ICIIS, 2009, pp. 118-122.
[6] S. Ciftci and M. Torlak. “A Comparison of Energy Detectability Models for Spectrum Sensing.”
in IEEE GLOBECOM, 2008, pp.1-5.
[7] H. Urkowitz. “Energy detection of unknown deterministic signals.” Proc. IEEE, vol. 55, pp. 523–
531, April 1967.
[8] W. C. Van Etten. Introduction to Random signals and Noise. England: John Willey & Sons Inc.,
2005, pp. 101-102.
[9] J. G. Proakis. Digital Communications. New York: McGraw-Hill, 2001, pp. 25-47.
[10] C.E. Shannon. “Communications in the presence of noise.” Proc. IRE, vol. 37, pp. 10-21, Jan.
1949.

[11] F. F. Digham, M. S. Alouini, and M. K. Simon. “On the energy detection of unknown signals
over fading channels.” in Proc. IEEE Int. Conf. Commun., Anchorage, AK, vol. 5, pp. 3575–
3579, May 2003.
[12] A. Papoulis and S.U Pillai. Probability, Random Variables, and Stochastic Processes. New
York: McGraw-Hill, 2002, pp. 82-90.
[13] I. S. Gradshteyn and I. M. Ryzhik. Table of Integrals, Series, and Products. New York:
Academic Press, 7th edition, 2007, pp. 657-658.
[14] A. H. Nuttall. “Some integrals involving Q-M functions (Corresp.).” IEEE Transactions On
Information theory, vol. 21, pp. 95-96, 1975.

Comparison of Energy Detection Based Spectrum Sensing Methods over Fading Channels in Cognitive Radio

More Related Content

What's hot (19)

Viewers also liked (18)

Similar to Comparison of Energy Detection Based Spectrum Sensing Methods over Fading Channels in Cognitive Radio (20)

Recently uploaded (20)

Comparison of Energy Detection Based Spectrum Sensing Methods over Fading Channels in Cognitive Radio