Enhancing Cognitive Radio Network Design with New Energy
Detection versus Pilot and Radio Based Techniques
Rizwana N. A. and Nagaraju V.
Research Scholar, Department of Computer Science and Engineering, Saveetha School of Engineering,
Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, Tamil Nadu, 602105, India
Keywords: Network Dataset, Energy Efficiency, Cognitive Radio Networks, Novel Energy Detection Technique, Pilot
and Radio Based Detection Technique, Spectrum Sensing, Technology.
Abstract: This study aimed to enhance the energy efficiency (EE) and accuracy of the Cognitive Radio Network (CRN)
system design by using a unique energy detection approach, contrasting it with the conventional Pilot and
Radio Based Detection Technique. A model was developed and processed in Python, using a network dataset
for initial exploration, sourced from the UCI Machine Learning Repository. Statistically, with a confidence
interval of 95% and sample size of 140, the energy detection's precision was assessed. In evaluating spectrum
allocation, the conventional technique had a slightly higher accuracy. However, our proposed energy detection
method achieved an impressive 95.2713% accuracy. Surprisingly, it processed in just 4 seconds, half the time
taken by the conventional method. The results confirm the new method's superiority in energy efficiency.
1 INTRODUCTION
Utilising CR technology can more effectively address
the issue of frequency underutilisation in wireless
transmission (Hamdan et al.). The emerging trend of
addressing the EE aspect of WSN is motivated by
rapidly escalating energy costs and stringent
environmental regulations (Li and Kara 2017). From
a green perspective, given that spectrum is a natural
resource meant to be shared rather than wasted,
cognitive users can significantly enhance the EE of
wireless links (Sun et al. 2013). While the CR
community has been effective in promoting the
concept of CR and developing prototypes,
programmes, and fundamental components, it has
faced unexpected challenges in clearly defining the
boundaries of what constitutes a CR (Neel 2006).
Over the last five years, more than 150 research
articles covering varied CRN concepts have been
published in Science Direct journals, and nearly 300
research articles are available on GS (Google
Scholar). Previous research indicates that the EE of
the system can be conceptualised as a joint
optimisation process; that is, trying to determine the
optimal parameters of criteria and sleep percentages
that minimise power consumption across the entire
CSS system (Maleki et al. 2014, 2015), set against
global predictive performance and false alarm
probability constraints (Maleki et al. 2015). Due to
the complexity of the joint optimisation problem,
specific assumptions are essential for finding the
optimal solution numerically, such as a flat-fading
environment with consistent SNR across all detectors
(Wu, Ng, and Lam 2022). Collaboration can notably
enhance bandwidth efficiency within CRN networks.
Spectrum gap allocation has been influenced by
network access, a recognised research gap (Lacunae).
Both distributed and centralised systems could be
accessed (Mukherjee and Nath 2015). The approach
adopted in this research is predicated on datasets from
users who have previously accessed it, focusing on
one or multiple functionalities. This approach creates
spectral gaps for SUs. However, the distribution of
spectrum gaps might deter smooth network use due to
the significantly faster and superior bandwidth
sharing being replicated. The process of relaying the
results of local channel estimation requires additional
energy. It's vital to efficiently use the energy of SUs,
especially when their resources are finite (Hu et al.
2015). The proposed energy detection method has
proven to be more accurate than existing pilot and
radio techniques.
520
A., R. and V., N.
Enhancing Cognitive Radio Network Design with New Energy Detection versus Pilot and Radio-Based Techniques.
DOI: 10.5220/0012602900003739
Paper published under CC license (CC BY-NC-ND 4.0)
In Proceedings of the 1st International Conference on Artificial Intelligence for Internet of Things: Accelerating Innovation in Industry and Consumer Electronics (AI4IoT 2023), pages 520-526
ISBN: 978-989-758-661-3
Proceedings Copyright © 2024 by SCITEPRESS – Science and Technology Publications, Lda.

2 MATERIALS AND METHODS
The Wireless Sensor Network Security Laboratory at
SSE (Saveetha School of Engineering), SIMATS
(Saveetha Institute of Medical and Technical
Sciences), conceptualised and refined the proposed
research project. Within this recommended CRN
System Design, there are two distinct groups. Group
1 is termed the Energy Detection Technique, while
Group 2 is named the Pilot Radio Based Detection
Technique. For each group, a sample size of 140 was
repeatedly determined (Ghasemi and Sousa 2008).
After sourcing the WSN network dataset from an
online platform, data pre-processing techniques were
applied to remove redundant and unnecessary data.
Subsequently, the EE rate of the novel Energy
Detection Technique, as well as the Pilot and Radio
Based Detection Technique, was examined and
compared against the relevant datasets.
On an experimental front, an online dataset was
procured and utilised in the current research project.
The creation of CRN systems was facilitated using
Python programming tools. Among various software
options, Python stands out as a preferred tool for
developing and analysing WSN results. This software
boasts a plethora of tools and inbuilt library features
that cater to the comprehensive processing of WSNs.
2.1 Energy Detection Technique
The Energy Detection (ED) approach is a
fundamental detection method that doesn't require
prior knowledge of the PU signal. This makes the
detection method advantageous in terms of ease of
use and computational simplicity. It gauges a specific
spectrum segment based on the received energy. To
ascertain the channel's availability, the sensor
compares the perceived energy to a threshold level.
However, this method's extended sensing time, aimed
at improving the SNR, leads to higher power
consumption. Furthermore, fluctuations in ambient
noise profoundly influence the detection's efficacy.
The determination metric for ED can be expressed as
per Kabalci and Kabalci (2019).
𝐸𝐷 =


∑


‖𝑦(𝑘)‖

(1)
The presence or absence of the PU is ascertained
by contrasting the observed energy with the set
threshold. This approach essentially differentiates
between two scenarios: either the PU signal is absent,
denoted by H
0
, or it is present, represented by H
1
. The
signal received by the SU is denoted by Y(n), while
S(n) represents the signal transmitted by the PU.
Furthermore, W(n) characterises the additive white
Gaussian noise (AWGN) that maintains a zero mean.
The sensor's responsibility is to evaluate the
relationship as delineated by Kockaya and Develi
(2020).
𝐻

: 𝑌
(
𝑛
)
= 𝑊
(
𝑛
)
, : 𝑃𝑟𝑖𝑚𝑎𝑟𝑦𝑈𝑠𝑒𝑟𝐴𝑏𝑠𝑒𝑛𝑡
𝐻

: 𝑌
(
𝑛
)
= 𝑆
(
𝑛
)
+ 𝑊
(
𝑛
)
: 𝑃𝑟𝑖𝑚𝑎𝑟𝑦𝑈𝑠𝑒𝑟𝑃𝑟𝑒𝑠𝑒𝑛𝑡
The fundamental block diagram for energy
sensing is displayed in Figure 1.
2.2 Pseudocode
Initialization:
- Set up variables to store data for Energy
Detection Technique (EDT) and Pilot and Radio
Based Detection Technique (PRDT)
- Input primary and secondary user signals
EDT:
- Calculate the energy of the received signal
- Compare the energy with a threshold to
determine if a primary user signal is present
- If a primary user signal is detected, secondary
user waits for the primary user to finish
PRDT:
- Cross-correlate the received signal with a
known primary user signal
- Determine if a primary user signal is present
based on the correlation result
- If a primary user signal is detected, secondary
user waits for the primary user to finish
Compare Results:
- Compare the performance of EDT and PRDT
in terms of accuracy, computational complexity, and
latency
Output:
- Report on the comparison of the efficiency of
EDT and PRDT in detecting primary user signals in
cognitive radio networks
2.3 Pilot and Radio Based Detection
Technique
Cyclostationary features, derived from pilot
structures, prove to be effective in spectrum sensing.
Taking the scattered pilot mode of PP1 into
consideration allows us to evaluate these attributes.
With pilot sub-bands in each OFDM symbol spaced
at intervals of 12 subcarriers, we can expect distinct
peak values. If the same pattern and formation are
consistent across all symbols, this observation
remains valid, as indicated by Khoshnevis (2012).
Practical transceivers set aside some of their signal
Enhancing Cognitive Radio Network Design with New Energy Detection versus Pilot and Radio-Based Techniques
521

strength for pilot signals, aiding in detection on the
receiver side. For instance, Digital TV transmissions
encompass pilot signals that are typically set to be 11
dB below the data-carrying signals. The challenge of
identifying the presence or absence of the PU in
spectrum access is frequently framed as a traditional
binary hypothesis testing problem, as elucidated by
Hattab and Ibnkahla (2014).
Pseudocode
Step 1: Start
Step 2: Initialize the radio signal parameters
Step 3: Continuously monitor the radio signals for
any potential pilot signals
Step 4: If a potential pilot signal is detected:
a. Extract the pilot signal from the radio signal
b. Analyze the extracted pilot signal to determine
if it matches the expected characteristics of a valid
pilot signal
c. If the extracted pilot signal is a valid pilot
signal, determine the location of the source of the
pilot signal
Step 5: Repeat steps 3-4 until the desired
detection threshold is reached
Step 6: End
2.4 Statistical Analysis
To compute the Standard Deviation (SD), mean
deviation data, significance point data, and to create
graphical representations, as well as to calculate the
Independent Sample T-Test, the statistical software
tool IBM SPSS version 26.0 is employed. The current
research methodology favoured the use of the SPSS
software for analysing the relevant intrusion dataset.
During a specific experimental phase, two distinct
graphs showcasing different features were crafted,
and the study concentrated on group statistics
practices and independent sample tests regarding the
experimental findings. Ideally, the network dataset
should comprise both training and testing datasets.
The training dataset is derived from extracting the test
dataset from the genuine dataset, given there are 140
records in total. The research encompasses two
independent variables, namely Accuracy and Loss,
and two dependent variables: EDT and PRDT.
3 RESULTS
Figure 1 illustrates the average accuracy of both the
Pilot & Radio Detection Technique (PRDT) and the
Energy Detection Method (EDT) models. The
innovative Energy Detection Technique (EDT)
consistently outperforms the Pilot and Radio
Detection Techniques (PRDT) in terms of mean
efficiency. As detailed in Tables 1 and 2, the average
overall accuracies for EDT and PRDT stand at
95.27% and 93.97%, respectively. The standard
deviation is represented as +/- 2 SD, with the
methodology denoted as +/- Y, and the mean
accuracy as Y.
Table 3 presents a side-by-side accuracy
comparison between the EDT and PRDT. When set
against the PRDT's 93.9702%, the EDT boasts an
impressive accuracy of 95.2713%. This demonstrates
that the PRDT method lags behind the EDT in terms
of effective spectrum allocation.
Table 3 further provides statistical metrics like the
mean, standard error, and both mean and standard
error values for EDT & PRDT. The t-test utilises the
accuracy parameter. The proposed technology's mean
accuracy stands at 95.2713%, juxtaposed against the
PRDT's mean accuracy of 93.9702%. The standard
deviation for EDT is recorded as 5.37334, which is
notably lower (by 4.29716) than that of PRDT. EDT's
average Standard Error clocks in at 1.38739, while
PRDT registers 1.10952.
As for the data in Table 4, it undergoes analysis
via the Independent T-test, maintaining a 95%
confidence level. SPSS computations reveal the
statistical significance value for the Energy Detection
as 0.800 (P>0.05) based on the t-value. This
significant value of 0.800 (p>0.05) underscores the
non-statistical significance in the disparities
concerning accuracy and loss between the two
algorithms, as discerned from the Independent
Sample T-Test evaluation.
4 DISCUSSION
This study juxtaposes the performance reliability of
the energy detection method (EDT) with the pilot and
radio detection (PRDT) techniques. Several examples
were employed to contrast the efficiency of the
proposed EDT against the conventional PRDT for
pinpointing spectrum vacancies. To emulate the
datasets, the UCI Machine Learning Repository
dataset was utilised. Python served as the system's
programming language. The predictions of the
proposed model are scrutinised to ascertain its
capability to yield superior results, and empirical
evaluations are undertaken to set the proposed
technique against any prevailing structure in terms of
precision and efficiency. The prior research utilises
the
Pilot and Radio Based Detection Technique,
which posts an average accuracy rate of 93.9702%.
AI4IoT 2023 - First International Conference on Artificial Intelligence for Internet of things (AI4IOT): Accelerating Innovation in Industry
and Consumer Electronics
522

Table 1: Accuracy of Energy Detection and the Pilot and Radio Detection.
S No.
Accuracy (%)
Energy Detection Pilot and Radio Detection
Sample 1 80.85 80.71
Sample 2 81.89 80.89
Sample 3 83.89 81.15
Sample 4 84.23 83.52
Sample 5 84.89 84.48
Sample 6 86.39 85.09
Sample 7 87.66 86.31
Sample 8 88.17 87.78
Sample 9 88.64 88.06
Sample 10 88.64 89.72
Sample 11 89.08 90.58
Sample 12 90.29 91.66
Sample 13 92.07 92.34
Sample 14 92.07 93.70
Sample 15 93.69 94.00
Table 2: Loss of Energy Detection and Pilot and Radio Detection.
Sample
Loss
Energy Detection Pilot and Radio Detection
Sample 1 19.15 19.29
Sample 2 18.11 19.11
Sample 3 16.11 18.85
Sample 4 15.77 16.48
Sample 5 15.11 15.52
Sample 6 13.61 14.91
Sample 7 12.34 13.69
Sample 8 11.83 12.22
Sample 9 11.36 11.94
Sample 10 11.36 10.28
Sample 11 10.92 09.42
Sample 12 09.71 08.34
Sample 13 07.93 07.66
Sample 14 07.93 06.30
Sample 15 06.31 06.00
Table 3: Comparison of the accuracy and loss in which the Energy Detection has the highest accuracy (95.27%) and the
lowest loss (04.72%) respectively compared to Pilot and Radio Detection has the lowest accuracy (93.97%) and highest loss
(06.02%).
Group
Statistics
Group N Mean
Std.
Deviation
Std. Error
Mean
Accuracy
Energy Detection 15 95.2713 5.37334 1.38739
Pilot Radio Detection 15 93.9702 4.29716 1.10952
Loss
Energy Detection 15 04.7287 5.37334 1.38739
Pilot Radio Detection 15 06.0298 4.29716 1.10952
Enhancing Cognitive Radio Network Design with New Energy Detection versus Pilot and Radio-Based Techniques
523

Table 4: Independent Sample T-Test is applied for the sample collections with a confidence interval as 95%. After applying
the SPSS calculation, it was found that the least square support vector machine has a statistical significance value of 0.800
(P>0.05) shows they are insignificant.
Independent
Samples Test
Levene's
Test for
Equality of
Variances
t-test for
Equality of
Means
F
Sig. t df
Significance
Mean
Difference
Std. Error
Difference
95%
Confidence
Interval of the
Difference
One-Sided p Two-Sided p Lower Upper
Accuracy
Equal
variances
assumed
2.326 0.400 -0.256 28 0.400 0.800 -0.45533 1.77648 -4.09429 3.18363
Equal
variances not
assumed
-0.256 26.709 0.400 0.800 -0.45533 1.77648 -4.10223 3.19157
Loss
Equal
variances
assumed
2.326 0.400 0.256 28 0.400 0.800 0.45533 1.77648 -3.18363 4.09429
Equal
variances
not
assumed
0.256 26.709 0.400 0.800 0.45533 1.77648 -3.19157 4.10223
Figure 1: Fundamental Block Diagram for Energy Detection.
An energy detection approach is developed,
brandishing a mean accuracy of 95.2713%. The
significance level for the Independent Sample T-Test
(Two-Tailed Test) in this exploration was ascertained
at p = 0.800 (p>0.05).
To access the non-continuous band, Liu et al.
(2018) conceived a CR system rooted in transform
domain data transmission. This mechanism discerns
the band's status via spectrum sensing and earmarks
the prospective effective spectrum marker vector. By
harnessing a basis carrier sprouted from the frequency
marking vector, energy can be channelled towards the
dormant sub-channels. Numerical simulations
intimate that an agile design might empower the
envisioned CR system to overshadow the broad band
system, thereby enhancing capacity. Ghasemi and
Sousa (2008) proffer a synopsis of the pertinent
regulations and pivotal challenges entailed in
deploying energy detection capacities in CNR
systems in their research. Furthermore, they elucidate
several design trade-offs that must be orchestrated to
elevate the efficiency of diverse elements. Kabalci
and Kabalci (2019) provide an exhaustive exposition
on the nuances of CR technology, CRNs, and their
adaptive strategies within the framework of Smart
Grid (SG) systems, factoring in applications for
metering, monitoring, and data management.
AI4IoT 2023 - First International Conference on Artificial Intelligence for Internet of things (AI4IOT): Accelerating Innovation in Industry
and Consumer Electronics
524

Figure 2: Comparison of Accuracy of Energy Detection Technique with Pilot and Radio Detection Technique in terms of
mean Accuracy. The mean accuracy of the Energy Detection is higher than the Pilot and Radio Detection and the standard
deviation of the Energy Detection is better than the Pilot and Radio Detection. Bar graph comparison of Pilot and Radio
Detection which has mean accuracy of 93.9702% compared to Energy Detection which has mean accuracy of 95.2713%. X-
axis algorithm, Y-axis mean accuracy with SD  2 SD.
A paramount hurdle for CR lies in identifying and
chronicling every spectrum void prevalent in the
ambient milieu. An enhanced pilot-based spectrum
access method for CR was introduced by Ghaith
Hattab and colleagues in 2014. Contrary to
conventional pilot-based detectors that merely scout
for the presence of pilot signals, the proposed detector
capitalises on the presence of the signal transmitting
actual data. They statistically juxtapose the
performance of their proposed detector with that of
the extant ones, showcasing that the recognition rate
of their detector markedly excels.
5 CONCLUSION
In the rapidly evolving landscape of Cognitive Radio
(CR) technology, efficient spectrum sensing
methodologies are pivotal to ensuring optimal use of
the available spectrum. Amidst the array of available
techniques, the Energy Detection Technique (EDT)
and the Pilot and Radio Detection Technique (PRDT)
have emerged as frontrunners. While both methods
offer significant advantages, the nuanced differences
in their performance, especially in terms of accuracy
and loss, warrant a closer inspection.
Key Points:
● Comprehensive Approach: Both EDT and
PRDT are comprehensive in their approach to
spectrum sensing. They adopt different
methodologies to detect the presence or
absence of primary users in the spectrum.
● Accuracy Metrics: In the allocation of
spectrum holes, PRDT recorded an accuracy
of 93.9702%. However, EDT surpassed this
with a slightly higher accuracy of 95.2713%.
This difference, though seemingly marginal,
can have profound implications in real-world
applications.
● Loss Considerations: PRDT, while
commendable in its precision, also presented a
loss of 06.0298. In contrast, EDT marked a
reduced loss of 04.7287, indicating a more
efficient utilization of resources.
● Operational Efficiency: Beyond mere
numbers, the efficacy of EDT signifies
smoother operational flow and fewer
interruptions, which is vital for seamless
communication.
● Robustness: The resilience and robustness of
EDT, as indicated by its superior performance
metrics, suggest that it is better equipped to
handle diverse and dynamic spectrum
environments.
● Future Implications: Given the current
trajectories, it's plausible to predict that
advancements in EDT might further widen the
performance gap, making it a more preferred
choice for future CR implementations.
Enhancing Cognitive Radio Network Design with New Energy Detection versus Pilot and Radio-Based Techniques
525

In essence, while the Pilot and Radio Detection
Technique has its merits, it's the novel Energy
Detection Technique that appears to be leading the
race in terms of efficiency and accuracy. The inherent
strengths of EDT, as evinced by its superior
performance metrics, make it an auspicious contender
in the quest for the most effective spectrum sensing
methodology. As the telecommunication sector
hurtles towards more data-intensive applications and
crowded spectrums, such innovations and their
meticulous evaluations will be of paramount
importance.
REFERENCE
Ghasemi, A., and E. S. Sousa. 2008. “Spectrum Sensing in
Cognitive Radio Networks: Requirements, Challenges
and Design Trade-Offs.” IEEE Communications
Magazine.
https://doi.org/10.1109/mcom.2008.4481338.
G. Ramkumar, R. Thandaiah Prabu, Ngangbam Phalguni
Singh, U. Maheswaran, Experimental analysis of brain
tumor detection system using Machine learning
approach, Materials Today: Proceedings, 2021, ISSN
2214-7853,
https://doi.org/10.1016/j.matpr.2021.01.246.
Hamdan, N. S., M. H. Mohamad, N. E. Samad, and S. R.
Ab Rashid. “Analysis of Spectrum Holes and Power
Allocation Using Water Filling Model for Underlay
Cognitive Radio.” Accessed December 21, 2022.
http://crim.utem.edu.my/wp-content/uploads/2022/09/
019-39-401.pdf.
Hattab, Ghaith, and Mohammed Ibnkahla. 2014.
“Enhanced Pilot-Based Spectrum Sensing Algorithm.”
2014 27th Biennial Symposium on Communications
(QBSC). https://doi.org/10.1109/qbsc.2014.6841184.
Hu, Hang, Hang Zhang, Hong Yu, Yi Chen, and Javad
Jafarian. 2015. “Energy-Efficient Design of Channel
Sensing in Cognitive Radio Networks.” Computers &
Electrical Engineering 42 (February): 207–20.
Kabalci, Ersan, and Yasin Kabalci. 2019. “Cognitive Radio
Based Smart Grid Communications.” From Smart Grid
to Internet of Energy. https://doi.org/10.1016/b978-0-
12-819710-3. 00006 -5.
Kumar M, M., Sivakumar, V. L., Devi V, S.,
Nagabhooshanam, N., & Thanappan, S. (2022).
Investigation on Durability Behavior of Fiber
Reinforced Concrete with Steel Slag/Bacteria beneath
Diverse Exposure Conditions. Advances in Materials
Science and Engineering, 2022.
Khoshnevis, Hossein. 2012. “Pilot Signature Based
Detection of DVB-T2 Broadcasting Signal for
Cognitive Radio.” 2012 Third International
Conference on The Network of the Future (NOF).
https://doi.org/10.1109/nof.2012.6464011.
Kockaya, Kenan, and Ibrahim Develi. 2020. “Spectrum
Sensing in Cognitive Radio Networks: Threshold
Optimization and Analysis.” EURASIP Journal on
Wireless Communications and Networking 2020 (1).
https://doi.org/10.1186/s13638-020-01870-7.
Liu, Xin, Yongjian Wang, Yanping Chen, Junjuan Xia,
Xueyan Zhang, Weidang Lu, and Feng Li. 2018. “A
Multichannel Cognitive Radio System Design and Its
Performance Optimization.” IEEE Access.
https://doi.org/10.1109/access.2018.2810061.
Maleki, Sina, Geert Leus, Symeon Chatzinotas, and Björn
Ottersten. 2014. “To AND or To OR: How Shall the
Fusion Center Rule in Energy-Constrained Cognitive
Radio Networks?” In 2014 IEEE International
Conference on Communications (ICC), 1632–37.
2015. “Energy-Efficient Distributed Spectrum Sensing
with Combined Censoring and Sleeping.” IEEE
Transactions on Wireless Communications 14 (8):
4508–21.
Mukherjee, T., and A. Nath. 2015. “Cognitive Radio-
Trends, Scope and Challenges in Radio Network
Technology.” Aquatic Microbial Ecology:
International Journal. https://www.researchgate.net/
profile/Asoke-Nath-4/publication/279530663_ijarcsms
_cognitive_radio_tm_an_30_06_2015/links/55956134
08ae5d8f3930ef3c/ijarcsms-cognitive-radio-tm-an-30-
06-2015.pdf.
Neel, J. O. D. 2006. “Analysis and Design of Cognitive
Radio Networks and Distributed Radio Resource
Management Algorithms.”
https://vtechworks.lib.vt.edu/handle/10919/29998.
S. G and R. G, "Automated Breast Cancer Classification
based on Modified Deep learning Convolutional Neural
Network following Dual Segmentation," 2022 3rd
International Conference on Electronics and
Sustainable Communication Systems (ICESC),
Coimbatore, India, 2022, pp. 1562-1569, doi:
10.1109/ICESC54411.2022.9885299.
Sun, Hongjian, Arumugam Nallanathan, Cheng-Xiang
Wang, and Yunfei Chen. 2013. “Wideband Spectrum
Sensing for Cognitive Radio Networks: A Survey.”
IEEE Wireless Communications 20 (2): 74–81.
Wu, Qingying, Benjamin K. Ng, and Chan-Tong Lam.
2022. “Energy-Efficient Cooperative Spectrum Sensing
Using Machine Learning Algorithm.” Sensors 22 (21).
https://doi.org/ 10.3390/s22218230.
AI4IoT 2023 - First International Conference on Artificial Intelligence for Internet of things (AI4IOT): Accelerating Innovation in Industry
and Consumer Electronics
526