Hindawi Publishing Corporation

Journal of Electrical and Computer Engineering

Volume 2015, Article ID 591627, 6 pages
http://dx.doi.org/10.1155/2015/591627

Research Article
Fast Two-Step Energy Detection for Spectrum Sensing

Meiling Lai,1 Shengliang Peng,1 Xi Yang,2 and Lin Zhou1

1
Institute of Communications Technology, Huaqiao University, Xiamen 361021, China
2
School of Information Science and Technology, Jishou University, Jishou 416000, China

Correspondence should be addressed to Shengliang Peng; peng.shengliang@gmail.com

Received 22 January 2015; Accepted 20 March 2015

Academic Editor: John N. Sahalos

Copyright © 2015 Meiling Lai et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Spectrum sensing is one of the key tasks in cognitive radio. This paper proposes a fast two-step energy detection (FED) algorithm
for spectrum sensing via improving the sampling process of conventional energy detection (CED). The algorithm adaptively selects
𝑁-point or 2𝑁-point sampling by comparing its observed energy with prefixed double thresholds, and thereby is superior in
sampling time and detection speed. Moreover, under the constraint of constant false alarm, this paper optimizes the thresholds
from maximizing detection probability point of view. Theoretical analyses and simulation results show that, compared with CED,
the proposed FED can achieve significant gain in detection speed at the expense of slight accuracy loss. Specifically, within high
signal-to-noise ratio regions, as much as 25% of samples can be reduced.

1. Introduction target PU is not detected by ED. These schemes can achieve

better performance under awful noise conditions, while their
With the rapid growth of wireless communications, the issue complexity is high because of CFD. To solve this problem,
of spectrum resource scarcity has been causing deep concern. [10–12] consider performing ED twice to avoid CFD. In [10],
Cognitive radio (CR) is considered as an efficient approach to two thresholds are preset and decisions are made directly
solve this problem via opportunistic spectrum access [1, 2]. In in case that the observed energy is either large enough or
CR systems, the secondary user (SU) is not allowed to access small enough. Otherwise, the second ED is implemented
the licensed spectrum that is being occupied by a primary additionally. This method can improve detection accuracy
user (PU). Otherwise, if no active PU is found, the vacant to some extent but is sensitive to noise uncertainty. An
spectrum can be utilized [3]. Therefore, the task to detect adaptive double-threshold energy detection (AED) method
whether the target PU is present or not, namely, spectrum is proposed in [11], which adjusts its thresholds according to
sensing, is of great significance for CR. SNR to combat noise uncertainty. It should be pointed out
Three classic spectrum sensing technologies are matched that these methods are all committed to improve detection
filter detection (MFD), cyclostationary feature detection accuracy, while ignoring the importance of detection time.
(CFD), and energy detection (ED). MFD is viewed as the Reference [12] investigates the detection accuracy of AED
most accurate method but requires detailed prior knowledge and analyzes its detection time, indicating that the former is
of PU signal [4]. CFD performs reliably even with very low improved at the cost of the latter. Since long detection time
signal-to-noise ratio (SNR), while its computational com- impairs system throughput and agility [13], it is necessary to
plexity is a bottleneck [5]. Compared with them, ED does not dig into the issue of fast ED.
require any prior information and consumes little calculation As is known, the whole detection comprises sampling
[6] and thereby is widely applied to detect PU. and calculation processes. Due to the reasons that calculation
ED suffers from server performance degradation at low in ED is rather simple and time required by this process
SNR [6]. To improve its performance, [7–9] present some can be neglected, given certain sampling frequency, detection
ED-CFD hybrid schemes that firstly perform ED to search time of ED is mainly determined by its sampling number
for PU and then reuse the observations to conduct CFD if [14]. In this paper, a fast two-step ED (FED) scheme is
2 Journal of Electrical and Computer Engineering

proposed via improving the sampling process of conventional 2N

energy detection (CED) and reducing its sampling number. ···
This new scheme can achieve great gain in detection speed (a) CED
at the expense of slight accuracy loss. The rest of our N (the first step) N (the second step)
paper is organized as follows. CED is briefly introduced in ··· ···
Section 2. In Section 3, FED is discussed and its performance
is analyzed. A method to determine the optimal thresholds is (b) FED
also provided. Simulation results are presented in Section 4 Figure 1: Sampling process of CED and FED.
and conclusions are drawn in Section 5.

2. Conventional Energy Detection respectively. According to (3), we can rewrite the probabilities
as
In CED, whether a target PU is present is determined by
comparing the received energy with the preset thresholds, 𝜆−𝑛
𝑃𝑓 = 𝑄 ( ),
which can be modeled as a binary hypothesis described as √2𝑛
(6)
{𝑛 (𝑘) 𝐻0 𝜆 − 𝑛 (1 + 𝛾)
𝑦 (𝑘) = { (1) 𝑃𝑑 = 𝑄 ( ),
𝑠 + 𝑛 (𝑘) 𝐻1 ,
{ (𝑘) √2𝑛 (1 + 2𝛾)
where 𝑦(𝑘) represents the received signal at SU’s receiver; ∞ 2

𝑠(𝑘) and 𝑛(𝑘) denote the PU signal and the additive noise, where 𝑄(𝑥) = ∫𝑥 (1/√2𝜋)𝑒−𝑡 /2 𝑑𝑡 is the 𝑄-function.
respectively; 𝐻1 and 𝐻0 stand for the hypotheses of PU’s As for the detection time, since the time required for cal-
presence and absence, respectively. In this case, the test culation is negligible, it is mainly determined by the sampling
statistic of CED is given by number:
𝑛
𝑛
󵄨 󵄨2 𝑇𝑑 = , (7)
𝑉 = ∑ 󵄨󵄨󵄨𝑦(𝑘)󵄨󵄨󵄨 , (2) 𝑓𝑆
𝑘=1
where 𝑓𝑆 is the sampling frequency. In the rest of this
where 𝑉 denotes the received energy and 𝑛 means the sam- paper, we will discuss the sampling number 𝑛 instead of the
pling number. detection time 𝑇𝑑 for simplicity.
Assume the noise is AWGN with zero mean and unit var-
iance and PU transmits constant-power signal. Since 𝑛 is 3. Fast Two-Step Energy Detection
usually very large, 𝑉 approximately follows Gaussian distri-
butions [15]: 3.1. Algorithm Description. FED is proposed based on CED.
Assume 2𝑁 samples are collected to calculate energy in CED
{𝑁 (𝑛, 2𝑛) 𝐻0 (𝑛 = 2𝑁). We equally divide the samples into two parts, as
𝑉∼{ (3) illustrated in Figure 1(b). By comparing its observed energy
𝑁 (𝑛 (1 + 𝛾) , 2𝑛 (1 + 2𝛾)) 𝐻1 ,
{ with the preset thresholds, FED adaptively chooses 𝑁-point
or 2𝑁-point sampling.
where 𝛾 is the received SNR measured by SU.
The algorithm of FED is detailed as follows: two thresh-
In CED, SU can make decisions on whether PU is present
olds 𝜆 1 and 𝜆 2 (𝜆 1 > 𝜆 2 ) are preset initially. At the first
or not (𝐻1 or 𝐻0 ) as follows:
step, 𝑁 samples are gathered to measure the received energy
if 𝑉 ≥ 𝜆 𝐻1 𝑉1 . If 𝑉1 is high enough and greater than 𝜆 1 , hypothesis
(4) 𝐻1 will be accepted directly. Otherwise, the second step will
if 𝑉 < 𝜆 𝐻0 , be performed:

where 𝜆 is the preset decision threshold. if 𝑉1 ≥ 𝜆 1 accept 𝐻1

The detection accuracy of CED is usually evaluated in (8)
terms of two metrics, namely, the false alarm probability (𝑃𝑓 ) if 𝑉1 < 𝜆 1 do the second step.
and the detection probability (𝑃𝑑 ): At the second step, other 𝑁 samples will be collected to
∞ calculate the received energy 𝑉2 . If the average value of 𝑉1 and
𝑃𝑓 = 𝑃 (𝑉 ≥ 𝜆 | 𝐻0 ) = ∫ 𝑓 (𝑉 | 𝐻0 ) 𝑑V, 𝑉2 is greater than 𝜆 2 , 𝐻1 will be accepted. Otherwise, 𝐻0 will
𝜆
(5) be accepted:
∞
𝑃𝑑 = 𝑃 (𝑉 ≥ 𝜆 | 𝐻1 ) = ∫ 𝑓 (𝑉 | 𝐻1 ) 𝑑V, 𝑉1 + 𝑉2
𝜆 {
{if ≥ 𝜆2 accept 𝐻1
the second step { 2 (9)
where 𝑓(𝑉 | 𝐻0 ) and 𝑓(𝑉 | 𝐻1 ) are the conditional prob- { 𝑉1 + 𝑉2
if < 𝜆2 accept 𝐻0 .
ability density functions of 𝑉 under hypotheses 𝐻0 and 𝐻1 , { 2
Journal of Electrical and Computer Engineering 3

3.2. Detection Accuracy Analysis. Let 𝑃𝑓1 and 𝑃𝑓2 denote the maximizing 𝑃𝑑 for a given 𝑃𝑓 . The optimization problem can
false alarm probabilities at the first and the second steps of be described as
FED and 𝑃𝑑1 and 𝑃𝑑2 represent the corresponding detection
probabilities at two steps, respectively. Then we have 𝜆 1 − 𝑁 (1 + 𝛾)
Max: 𝑄 ( )
√2𝑁 (1 + 2𝛾)
𝜆1 − 𝑁
𝑃𝑓1 = 𝑃 (𝑉1 ≥ 𝜆 1 | 𝐻0 ) = 𝑄 ( ),
√2𝑁 𝜆 1 − 𝑁 (1 + 𝛾)
+ (1 − 𝑄 ( )) (13)
𝑉 + 𝑉2 𝜆 −𝑁 √2𝑁 (1 + 2𝛾)
𝑃𝑓2 = 𝑃( 1 ≥ 𝜆 2 | 𝐻0 ) = 𝑄 ( 2 ),
2 √𝑁
𝜆 2 − 𝑁 (1 + 𝛾)
𝜆 1 − 𝑁 (1 + 𝛾) ⋅ 𝑄( )
𝑃𝑑1 = 𝑃 (𝑉1 ≥ 𝜆 1 | 𝐻1 ) = 𝑄 ( ), √𝑁 (1 + 2𝛾)
√2𝑁 (1 + 2𝛾)
𝜆1 − 𝑁 𝜆 −𝑁
Subject to: 𝑄 ( ) + (1 − 𝑄 ( 1 ))
𝑉1 + 𝑉2 𝜆 − 𝑁 (1 + 𝛾) √2𝑁 √2𝑁
𝑃𝑑2 = 𝑃 ( ≥ 𝜆 2 | 𝐻1 ) = 𝑄 ( 2 ). (14)
2 √𝑁 (1 + 2𝛾) 𝜆 −𝑁
⋅ 𝑄( 2 ) = 𝛽,
(10) √𝑁

where 𝛽 is the predetermined false alarm probability.

This conditional extreme value model can be solved by
Note that 𝑁 is the sampling number of each step. Lagrange multiplier method. Define the following:
According to the algorithm, final false alarm probability
and final detection probability of FED, namely, 𝑃𝑓 and 𝑃𝑑 , can 𝜆1 − 𝑁 𝜆 −𝑁
be formulated as 𝑔 (𝜆 1 , 𝜆 2 ) = 𝛽 − 𝑄 ( ) − (1 − 𝑄 ( 1 ))
√2𝑁 √2𝑁
(15)
𝜆 −𝑁
⋅ 𝑄( 2 ).
𝑃𝑓 = 𝑃𝑓1 + (1 − 𝑃𝑓1 ) ⋅ 𝑃𝑓2 , √𝑁
(11)
𝑃𝑑 = 𝑃𝑑1 + (1 − 𝑃𝑑1 ) ⋅ 𝑃𝑑2 . Then the Lagrange function is

𝑓 (𝜆 1 , 𝜆 2 , 𝑐) = 𝑃𝑑 (𝜆 1 , 𝜆 2 ) + 𝑐 ⋅ 𝑔 (𝜆 1 , 𝜆 2 ) , (16)
Substituting (10) into (11), we get where 𝑐 is a constant. Rewriting 𝑃𝑑 (𝜆 1 , 𝜆 2 ) and 𝑔(𝜆 1 , 𝜆 2 ) with
the error function erf(𝑥) = 1 − 2𝑄(√2𝑥), we can get

𝜆1 − 𝑁 𝜆 −𝑁 𝜆 −𝑁 𝑓 (𝜆 1 , 𝜆 2 , 𝑐)
𝑃𝑓 = 𝑄 ( ) + (1 − 𝑄 ( 1 )) ⋅ 𝑄 ( 2 ),
√2𝑁 √2𝑁 √𝑁
3 1 𝜆 − 𝑁 (1 + 𝛾)
𝜆 1 − 𝑁 (1 + 𝛾) = − erf ( 1 )
𝑃𝑑 = 𝑄 ( ) 4 4 2√𝑁 (1 + 2𝛾)
√2𝑁 (1 + 2𝛾)

1 𝜆 − 𝑁 (1 + 𝛾)
𝜆 1 − 𝑁 (1 + 𝛾) − erf ( 2 )
+ (1 − 𝑄 ( )) 4 √2𝑁 (1 + 2𝛾)
√2𝑁 (1 + 2𝛾)

1 𝜆 − 𝑁 (1 + 𝛾) 𝜆 − 𝑁 (1 + 𝛾)
𝜆 2 − 𝑁 (1 + 𝛾) − erf ( 1 ) erf ( 2 )
⋅ 𝑄( ). 4 2√𝑁 (1 + 2𝛾) √2𝑁 (1 + 2𝛾)
√𝑁 (1 + 2𝛾)
1 𝜆 −𝑁 1 𝜆 −𝑁
(12) + 𝑐 ( erf ( 1 ) + erf ( 2 )
4 2√𝑁 4 √2𝑁
1 𝜆 −𝑁 𝜆 −𝑁 3
3.3. Optimization of Double Thresholds. As shown in (12), + erf ( 1 ) erf ( 2 ) − + 𝛽) .
detection accuracy is closely related to 𝜆 1 and 𝜆 2 . In this 4 2√𝑁 √2𝑁 4
subsection, we deduce the optimal double thresholds via (17)
4 Journal of Electrical and Computer Engineering

Taking the first-order partial derivative of (17) with 0.1

N = 100, 𝛾 = 5 dB, P0 = 0.5
respect to 𝜆 1 and 𝜆 2 , respectively,
0.09
𝜕𝑓 (𝜆 1 , 𝜆 2 , 𝑐)
= 0, (18) 0.08
𝜕𝜆 1 0.07
𝜕𝑓 (𝜆 1 , 𝜆 2 , 𝑐) 0.06
= 0, (19)
𝜕𝜆 2

Pf
0.05

and combining (18) and (19) to eliminate 𝑐, we can get the 0.04
relationship between 𝜆 1 and 𝜆 2 : 0.03
0.02
1 + erf ((𝜆 1 − 𝑁 (1 + 𝛾)) /2√𝑁 (1 + 2𝛾))
0.01
1 + erf ((𝜆 1 − 𝑁) /2√𝑁) 0
110 115 120 125 130 135
2 2 2
(−2𝛾𝜆1 +2𝑁𝛾𝜆 1 +𝑁 𝛾 )/4𝑁(1+2𝛾)
⋅𝑒 𝜆2
(20) 𝛽 = 0.1
1 + erf ((𝜆 2 − 𝑁 (1 + 𝛾)) /√2𝑁 (1 + 2𝛾)) 𝛽 = 0.05
= 𝛽 = 0.01
1 + erf ((𝜆 2 − 𝑁) /√2𝑁)
Figure 2: Final false alarm probability versus 𝜆 2 under different 𝛽.
(−2𝛾𝜆2 2 +2𝑁𝛾𝜆 2 +𝑁2 𝛾2 )/2𝑁(1+2𝛾)
⋅𝑒 .
N = 100, 𝛾 = 5 dB, P0 = 0.5
Moreover, based on (14), another equation can be 0.95
obtained as 0.9
𝛽 − 𝑄 ((𝜆 2 − 𝑁) /√𝑁) 0.85
𝜆 1 = 𝑁 + √2𝑁𝑄−1 ( ). (21)
1 − 𝑄 ((𝜆 2 − 𝑁) /√𝑁) 0.8

Consequently, given the parameters 𝑁, 𝛾, and 𝛽, the 0.75

Pd

optimal 𝜆 1 and 𝜆 2 can be deduced according to (20) and (21). 0.7

Based on these two thresholds, we can maximize detection
0.65
probability subject to certain false alarm probability.
0.6
3.4. Average Sampling Number Analysis. The proposed FED 0.55
scheme adaptively chooses 𝑁-point or 2𝑁-point sampling.
Let 𝑃0 and 𝑃1 represent the prior probabilities of hypotheses 0.5
110 115 120 125 130 135
𝐻0 and 𝐻1 , respectively. Then the average sampling number
𝜆2
of FED can be expressed as
𝛽 = 0.1 Optimal 𝜆2 when 𝛽 = 0.1
̃ = 𝑁 (𝑃0 ⋅ 𝑃𝑓1 + 𝑃1 ⋅ 𝑃𝑑1 )
𝑁 𝛽 = 0.05 Optimal 𝜆2 when 𝛽 = 0.05
(22) 𝛽 = 0.01 Optimal 𝜆2 when 𝛽 = 0.01
+ 2𝑁 (𝑃0 ⋅ (1 − 𝑃𝑓1 ) + 𝑃1 ⋅ (1 − 𝑃𝑑1 )) .
Figure 3: Final detection probability versus 𝜆 2 under different 𝛽.
Without loss of generality, assume 𝑃0 = 𝑃1 = 0.5 in this
paper. Then we have
4. Simulation Results
̃ = 𝑁 (4 − 𝑃𝑓1 − 𝑃𝑑1 ) .
𝑁 (23)
2 In this section, simulation results are provided to verify the
Since 0 < 𝑃𝑓1 + 𝑃𝑑1 < 1 [16], the bounds of average sampling FED algorithm and its theoretical conclusions.
number in FED are given by Figures 2 and 3 show final false alarm probability and
detection probability of FED algorithm versus 𝜆 2 , respec-
̃ < 2𝑁.
1.5𝑁 < 𝑁 (24) tively. Simulation settings are 𝑃0 = 0.5, 𝛾 = −5 dB, 𝑁 = 100
[17], and 𝛽 = 0.1, 0.05, 0.01. For any given 𝜆 2 , corresponding
Note that CED always performs 2𝑁-point sampling. So 𝜆 1 is obtained according to (21). As shown in Figure 2, final
FED certainly collects fewer samples and is faster than CED. false alarm probability is always equal to 𝛽. In other words, a
In addition, as much as 25% of samples and detection time pair of thresholds satisfying (21) ensures that the constraint
can be reduced at most compared with CED. on false alarm probability is met. In Figure 3, the optimal
Journal of Electrical and Computer Engineering 5

N = 100, P0 = 0.5 𝛽 = 0.01, P0 = 0.5

1 600
0.8
550
0.6
Pd

0.4 500

Average sampling number

0.2 450
0
−12 −10 −8 −6 −4 −2 0 2 400
𝛾
350
Average sampling

200 300
190
number

180 250
170
160 200
150 150
−12 −10 −8 −6 −4 −2 0 2
𝛾 100
−12 −10 −8 −6 −4 −2 0 2
𝛾
𝛽 = 0.1, CED 𝛽 = 0.1, FED
𝛽 = 0.05, CED 𝛽 = 0.05, FED N = 300, CED N = 300, FED
𝛽 = 0.01, CED 𝛽 = 0.01, FED N = 200, CED N = 200, FED
N = 100, CED N = 100, FED
Figure 4: Performance of CED and FED versus 𝛾 under different 𝛽.
Figure 5: Comparison of average sampling number under different
𝑁.
𝜆 2 (115.0, 118.3, 124.6), which is derived from (20) and (21),
is marked on the 𝑥-axis. Since final detection probability
achieves maximum values with the optimal 𝜆 2 as expected, N = 100, 𝛽 = 0.05
1
the effectiveness of (20) and (21) is certainly verified.
In order to evaluate the performances of CED and 0.8
FED, Figure 4 plots their detection probabilities and average 0.6
Pd

sampling numbers versus SNR under different false alarm 0.4

constraints 𝛽 = 0.1, 0.05, 0.01 with 𝑃0 = 0.5 and 𝑁 = 100. 0.2
According to the top subfigure of Figure 4, in spite of the
0
fact that detection probability of FED is slightly lower, two −12 −10 −8 −6 −4 −2 0 2
detection probability curves almost coincide, indicating that 𝛾
detection accuracy of CED and FED is approximately the
200
same. As shown in the bottom subfigure of Figure 4, CED
Average sampling

consistently takes 200-point sampling, while the sampling 180

number

number of FED is between 150 and 200. Since its sampling 160
number is reduced, FED requires less sampling time and 140
completes the detection faster. 120
In Figure 5, the average sampling number of CED and 100
FED is detailedly compared under different 𝑁 to highlight −12 −10 −8 −6 −4 −2 0 2
the superiority of the latter. It can be seen from this figure 𝛾
that, no matter how 𝑁 changes, the sample number of CED
P0 = 0.1, CED P0 = 0.1, FED
is always 2𝑁 and that of FED varies from 1.5𝑁 to 2𝑁, which
P0 = 0.5, CED P0 = 0.5, FED
agrees well with (24). Specifically, in the scenarios with high P0 = 0.9, CED P0 = 0.9, FED
SNR (SU is close to the PU transmitter), 0.5𝑁 samples are
reduced and as much as 25% of detection time can be saved. Figure 6: Performance of CED and FED versus 𝛾 under different 𝑃0 .
To demonstrate the effect of 𝑃0 on detection performance,
Figure 6 plots detection probabilities as well as average
sampling numbers of CED and FED versus SNR under
different prior probabilities 𝑃0 = 0.1, 0.5, 0.9. According to 5. Conclusions
the top subfigure of Figure 6, detection probabilities of these
two algorithms are little affected by 𝑃0 , while in the bottom A FED scheme for spectrum sensing is proposed via improv-
subfigure of Figure 6, we can see that the sample number of ing the sampling process of CED. Detection accuracy of
CED is always 2𝑁 and that of FED increases as 𝑃0 increases. this new scheme is analyzed. Moreover, we deal with the
The reason is that higher 𝑃0 means larger possibility of PU’s optimization problem and provide a method to deduce its
absence, and FED prefers taking 2𝑁-point sampling when optimal double thresholds. The average sampling number of
PU is absent. FED is also discussed. We prove that, compared with CED,
6 Journal of Electrical and Computer Engineering

the proposed FED certainly uses fewer samples and 25% of of the IEEE 12th International Conference on Computer and
samples can be reduced at most with high SNRs. Information Technology (CIT ’12), pp. 824–827, Chengdu, China,
October 2012.
[12] Q. Su, T. Song, J. Hu, and L. Shen, “Adaptive double-threshold
Conflict of Interests energy detection algorithm for cognitive radio,” Journal of
The authors declare that there is no conflict of interests Southeast University, vol. 27, no. 4, pp. 351–356, 2011.
regarding the publication of this paper. [13] Y.-C. Liang, Y. Zeng, E. C. Y. Peh, and A. T. Hoang, “Sensing-
throughput tradeoff for cognitive radio networks,” IEEE Trans-
actions on Wireless Communications, vol. 7, no. 4, pp. 1326–1337,
Acknowledgments 2008.
[14] S. Peng, X. Yang, S. Shu, P. Zhu, and X. Cao, “Adaptive sequential
This research is supported by the National Natural Sci- cooperative energy detection scheme for primary user detection
ence Foundation of China (Grant nos. 61102089, 61201264, in cognitive radio,” IEICE Transactions on Communications, vol.
61302095, and 61362018) and Huaqiao University (Grant nos. 94, no. 10, pp. 2896–2899, 2011.
12BS219 and 13BS101). [15] J. Ma, G. Zhao, and Y. Li, “Soft combination and detection
for cooperative spectrum sensing in cognitive radio networks,”
References IEEE Transactions on Wireless Communications, vol. 7, no. 11, pp.
4502–4507, 2008.
[1] S. Haykin, “Cognitive radio: brain-empowered wireless com- [16] S.-L. Peng, X. Yang, S. Shu, and X. Cao, “Exploitation of
munications,” IEEE Journal on Selected Areas in Communica- temporal persistence for accuracy improvement in primary user
tions, vol. 23, no. 2, pp. 201–220, 2005. detection,” IET Communications, vol. 4, no. 15, pp. 1855–1864,
[2] C.-S. Sum, G. P. Villardi, M. A. Rahman et al., “Cognitive 2010.
communication in TV white spaces: an overview of regulations, [17] J.-Y. Wu, C.-H. Wang, and T.-Y. Wang, “Performance analysis
standards, and technology,” IEEE Communications Magazine, of energy detection based spectrum sensing with unknown
vol. 51, no. 7, pp. 138–145, 2013. primary signal arrival time,” IEEE Transactions on Communi-
[3] Y. Zeng, Y.-C. Liang, A. T. Hoang, and R. Zhang, “A review on cations, vol. 59, no. 7, pp. 1779–1784, 2011.
spectrumsensing for cognitive radio: challenges and solutions,”
Eurasip Journal on Advances in Signal Processing, vol. 2010,
Article ID 381465, 15 pages, 2010.
[4] I. F. Akyildiz, W.-Y. Lee, M. C. Vuran, and S. Mohanty, “NeXt
generation/dynamic spectrum access/cognitive radio wireless
networks: a survey,” Computer Networks, vol. 50, no. 13, pp.
2127–2159, 2006.
[5] N. Sai Shankar, C. Cordeiro, and K. Challapali, “Spectrum agile
radios: utilization and sensing architectures,” in Proceedings
of the 1st IEEE International Symposium on New Frontiers in
Dynamic Spectrum Access Networks (DySPAN ’05), pp. 160–169,
November 2005.
[6] F. F. Digham, M.-S. Alouini, and M. K. Simon, “On the energy
detection of unknown signals over fading channels,” IEEE
Transactions on Communications, vol. 55, no. 1, pp. 21–24, 2007.
[7] S. Maleki, A. Pandharipande, and G. Leus, “Two-stage spectrum
sensing for cognitive radios,” in Proceedings of the IEEE Inter-
national Conference on Acoustics, Speech, and Signal Processing
(ICASSP ’10), pp. 2946–2949, Dallas, Tex, USA, March 2010.
[8] Z. Khalaf, A. Nafkha, J. Palicot, and M. Ghozzi, “Hybrid
spectrum sensing architecture for cognitive radio equipment,”
in PRoceedings of the 6th Advanced International Conference
on Telecommunications (AICT ’10), pp. 46–51, Istanbul, Turkey,
August 2010.
[9] Y. Zhang, L. Zhang, and C. Tang, “Joint detection of cyclo-
stationary and energy in cognitive radio,” in Proceedings of
the IEEE International Conference on Intelligent Systems and
Knowledge Engineering (ISKE ’10), pp. 182–186, Hangzhou,
China, November 2010.
[10] J.-B. Wu, T. Luo, and G. Yue, “An energy detection algorithm
based on double-threshold in cognitive radio systems,” in
Proceedings of the 1st International Conference on Information
Science and Engineering (ICISE ’09), pp. 493–496, IEEE, Nan-
jing, China, December 2009.
[11] J.-Q. Xie and J. Chen, “An adaptive double-threshold spectrum
sensing algorithm under noise uncertainty,” in Proceedings
 International Journal of

Rotating
Machinery

International Journal of
The Scientific
Engineering Distributed
Journal of
Journal of

Hindawi Publishing Corporation

World Journal
Hindawi Publishing Corporation Hindawi Publishing Corporation
Sensors
Hindawi Publishing Corporation
Sensor Networks
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Journal of

Control Science
and Engineering

Advances in
Civil Engineering
Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Submit your manuscripts at

http://www.hindawi.com

Journal of
Journal of Electrical and Computer
Robotics
Hindawi Publishing Corporation
Engineering
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

VLSI Design
Advances in
OptoElectronics
International Journal of

International Journal of
Modelling &
Simulation
Aerospace
Hindawi Publishing Corporation Volume 2014
Navigation and
Observation
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
in Engineering
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Engineering
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com
http://www.hindawi.com Volume 2014

International Journal of
International Journal of Antennas and Active and Passive Advances in
Chemical Engineering Propagation Electronic Components Shock and Vibration Acoustics and Vibration
Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014