�ݺ�ߣ

1
Six Monthly Progress Presentation
on
Design and Implementation of Spectrum sensing Techniques
in Cognitive Radio Systems
(From 15th
December 2020 to 15th
June 2021)
Presented By
Neelam Dewangan
Under the Supervision of
Dr. Arun Kumar (Supervisor), Associate Professor , BIT Durg
Dr. R.N.Patel (Co-Supervisor), Associate Professor, NIT Raipur

Design and Implementation of Spectrum Sensing Techniques in Cognitive Radio Systems Presented by : Neelam Dewangan 2
Scope of the Work
• To conduct an extensive study of the techniques for Spectrum sensing in
Cognitive Radio
• To investigate the scope of performance improvement of Spectrum
sensing and propose intelligent optimization techniques for Spectrum
sensing schemes

Objectives
• The main objective of this research is to design efficient techniques for
cooperative spectrum sensing that are capable of discovering multiple
spectrum access opportunities in a single sensing period to increase the
achievable cooperative gain while limiting the incurred cooperation
overhead and sensing errors
• Second objective is to investigate the performance of optimized technique
over different channels.
• Developing an optimal decision fusion rule that considers the correlation
between the cooperating secondary users’ local decisions.

Flowchart of Methodology
Current work

Sl. No Month & Year No. of days attended Work done by research scholar
1 January 2021 12 (Telephonic/ online
conversations)
Implementation of Python based algorithm (linear regression analysis and
logistic regression) for Machine Learning and verification with standard
datasets
2 February 2021 14 (Telephonic/ online
conversations)
Implementation and verification of results of Paper (a part of paper ) : J.
Tian et al., "A Machine Learning-Enabled Spectrum Sensing Method for
OFDM Systems," in IEEE Transactions on Vehicular Technology, vol. 68, no.
11, pp. 11374-11378, Nov. 2019
3 March 2021 13 (Telephonic/ online
conversations)
Dataset Generation for OFDM based Cognitive Radio
4 April 2021 04 (Telephonic/ online
conversations)
Simulation of Energy Detector in Python with datasets generated through
Monte Carlo Simulations
5 May 2021 11 (Telephonic/ online
conversations)
Simulation of Cyclostationary Detector in Python with datasets generated
through Monte Carlo Simulations
6 June 2021 05 (Telephonic/ online
conversations)
Implementation and verification of results of Paper (a part of paper ) : Y.
Arjoune and N. Kaabouch, "On Spectrum Sensing, a Machine Learning
Method for Cognitive Radio Systems," 2019 IEEE International Conference on
Electro Information Technology (EIT), 2019, pp. 333-338
Six Month Progress (15th
December 2020 to 15th
June 2021)

Python Implementation of Linear Regression Model
• Linear regression is a statistical method for modeling relationship between
a dependent variable with a given set of independent variables.
• we consider dependent variables as response and independent variables
as features for simplicity
• It is assumed that the two variables are linearly related
X 0 1 2 3 4 5 6 7 8 9
Y 1 3 2 5 7 8 8 9 10 12
Table : Data Set for Linear Regression

Cont…
Feature vector x = [x1, x2, …., xn],
Response vector y = [y1, y2, …., yn]
• Regression Line is given as
h(xi) represents the predicted response value for ith
observation.
β0 and β 1 are regression coefficients and represent y-intercept and slope of regression line respectively.
εi is residual error in ith
observation.
• Cost Function or square error is given by :
where SSxy is the sum of cross-deviations of y and x:
and SSxx is the sum of squared deviations of x.

Results
Plot 1 : Scatter Plot of dataset (Linear Regression)
Estimated coefficients: β0 = -0.0586206896552 β1 = 1.45747126437
Plot 2 : Plot of regression line dataset (Linear
Regression)

Python Implementation of Logistic Regression Model
• Logistic regression is basically a supervised classification algorithm.
• In a classification problem, the target variable(or output), y, can take only
discrete values for given set of features(or inputs), X
• As Linear regression assumes that the
data follows a linear function,
Logistic regression models the
data using the sigmoid function.
Table : Data Set for Linear Regression

Cont…
• The dataset has ‘p’ feature variables and ‘n’ observations.
• The feature matrix is represented as:
• Here, denotes the values of feature for observation.
• The observation, can be represented as:

Results
Output :
Plot 3: Regression line dataset (Linear Regression)
Accuracy : 0.89

Implementation and verification of results of Paper : J. Tian et al., "A Machine Learning-Enabled
Spectrum Sensing Method for OFDM Systems," in IEEE Transactions on Vehicular Technology, vol. 68, no.
11, pp. 11374-11378, Nov. 2019, doi: 10.1109/TVT.2019.2943997.
• This paper proposes class assisted prediction method based on Naïve Bayes
classifiers on OFDM systems.
• This paper also studies accuracy of different algorithms against different training
samples.

Implementation of Naïve Bayes Classifiers
• Bayes’ Theorem finds the probability of an event occurring given the probability of another event that has
already occurred. Bayes’ theorem is stated mathematically as the following equation:
where A and B are events and P(B) ? 0.
• Basically, we are trying to find probability of event A, given the event B is true. Event B is also termed
as evidence.
• P(A) is the priori of A (the prior probability, i.e. Probability of event before evidence is seen). The evidence
is an attribute value of an unknown instance(here, it is event B).
• P(A|B) is a posteriori probability of B, i.e. probability of event after evidence is seen.
Why to choose Bayes Classifiers ?
• They are extremely fast for both training and prediction
• They provide straightforward probabilistic prediction
• They are often very easily interpretable
• They have very few (if any) tunable parameters

1: Iterations = 250
2: SNR value from -5dB to -15dB
3: ModType = BPSK (It stores the modulation type. It can have one of the
four values BPSK, QPSK, 16-QAM and 64-QAM).
4: Generate OFDM signal
(i) T = OFDM(ModType)
(ii) S = Power Adjustment(T, SNR value) + WGN
(iii) Assign Signal label
5: Generate Noise
(i) N = WGN
(ii) Assign Noise label
6: Dataset = concatenate(Dataset, S, N)
7: Update ModType
8: Until all ModTypes go to step 4
9: Update SNR value
10: Until all SNR values go to step 3
11: Iterations = Iterations - 1
12: Go to step 2 if iterations 0
Dataset Generation Algorithm
Reference : Hassaan Bin Ahmad et al. “Ensemble Classifier Based Spectrum Sensing in Cognitive Radio Networks”
Volume 2019 |Article ID 9250562 | https://doi.org/10.1155/2019/9250562

Energy Detection
• Energy detection is a non-coherent detection scheme which do not need
prior knowledge about the primary user .
• Energy of the receiving signal in the presence of primary user is calculated
• If the received signal is above threshold then it is deduced that Primary
user is present.
• Also , the output of energy detector is compared with threshold and the
signal is detected .

Fig :Detector performance with 1000 statistics Fig :Detector performance with 10000 statistics
Simulation Results for Energy Detector

Fig :Detector statistics with 1000 statistics Fig :Detector statistics with 10000 statistics

Fig :Detector performance with 1000 statistics Fig :Detector performance with 10000 statistics
Simulation Results for Cyclostationary Detector

Implementation and verification of results of Paper : Y. Arjoune and N. Kaabouch, "On Spectrum
Sensing, a Machine Learning Method for Cognitive Radio Systems," 2019 IEEE International Conference
on Electro Information Technology (EIT), 2019, pp. 333-338, doi: 10.1109/EIT.2019.8834099
• This paper concludes that spectrum sensing methods based on a machine learning theory for
cognitive radio networks provides a more reliable solution as compared to traditional
methods like Energy detection, cyclostationary and matched filter.
• It also states that the random forest model outperforms all the other machine learning
methods.
• Currently working on Random Forest model , implementation of traditional methods has
been done.

Road Map for the duration of 15th
June 2021 to 15th
December 2021
June July August September October November December
Dataset verification
Implementation ML algorithms with
verified datasets
Compilation of results
Preparation and submission of
paper for various journals and
conferences

Thanks

�ݺ�ߣ

Research Presentation for those who are dong M.Tech

Recommended

More Related Content

Similar to Research Presentation for those who are dong M.Tech (20)

More from NeelamDewangan7 (6)

Recently uploaded (20)

Research Presentation for those who are dong M.Tech