SEMESTER I

Discipline: ELECTRONICS AND COMMUNICATION

Stream : EC3
 CODE COURSE NAME CATEGORY L T P CREDIT
ADVANCED
DISCIPLINE
221TEC100 ENGINEERING 3 0 0 3
CORE
MATHEMATICS

Preamble: The purpose of this course is to expose students to the basic theory of linear
algebra and probability.

Course Outcomes: The COs shown are only indicative. For each course, there can be 4 to 6
COs. After the completion of the course the student will be able to

CO 1 To analyze distributions of random variables and make computations based on that

evaluate average behaviour of random variables, and analyze their converging
CO 2
behviours
CO 3 To analyze behaviour of random processes and explain basis of vector spaces.

CO 4 To evaluate properties of linear transformations

To evaluate if a linear tranformaion is diagonalizable and decompose it using
CO 5
spectral decomposition theorem.

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO4 PO5 PO6 PO7
CO 1 3 3 3 3
CO 2 3 3 3 3
CO 3 3 3 3 3
CO 4 3 3 3 3
CO 5 3 3 3 3

Assessment Pattern

Bloom’s Category End Semester Examination

Apply 20
Analyse 20
Evaluate 20
Create
Mark distribution

Total Marks CIE ESE ESE Duration

100 40 60 2.5 hours

Continuous Internal Evaluation Pattern:

Continuous Internal Evaluation: 40 marks

● Problem assignments including unsolved exercise problems from reference text
books: 20 marks
● Quiz: 10 marks
● Test paper (1 number): 10 marks
Quiz shall include topics from at least 50% of the syllabus. Test paper shall include minimum
80% of the syllabus

End Semester Examination Pattern:

End Semester Examination: 60 marks

There will be two parts; Part A and Part B

● Part A will contain 5 numerical/short answer questions with 1 question from each
module, having 5 marks for each question. Students should answer all questions.
● Part B will contain 7 questions with minimum one question from each module of which
student should answer any five. Each question can carry 7 marks.
Model Question Paper

A P J ABDUL KALAM TECHNOLOGICAL UNIVERSITY

M.TECH DEGREE EXAMINATION
SEMESTER:
Branch:
ADVANCED ENGINEERING MATHEMATICS

Time: 2.5 Hours Marks: 60

Part A

Answer ALL Questions. Each question carries 5 marks

𝑘𝑘
1. Given that 𝑓𝑓(𝑥𝑥) = is a probability distribution of a random variable that can take
2𝑥𝑥
on the values 𝑥𝑥 = 0,1,2,3,∧ 4. Find 𝑘𝑘. Find the cumulative distribution function.
2. State and prove weak law of large numbers.
3. Show that (1,3,2, −2), (4,1, −1,3), (1,1,2,0), (0,0,0,1) is a basis for 𝑅𝑅 4 .
4. Let 𝑇𝑇: 𝑉𝑉 → 𝑊𝑊be a linear transformation defined by 𝑇𝑇(𝑥𝑥, 𝑦𝑦, 𝑧𝑧) = (𝑥𝑥 + 𝑦𝑦, 𝑥𝑥 − 𝑦𝑦, 2𝑥𝑥 +
𝑧𝑧). Find the range, null space, rank and nullity of 𝑇𝑇.
5. Describe an inner product space. If V is an inner product space, then for any vectors
𝛼𝛼, 𝛽𝛽in V prove that ‖𝛼𝛼 + 𝛽𝛽‖ ≤ ‖𝛼𝛼‖ + ‖𝛽𝛽‖.

Part B
Answer ANY FIVE Questions, one from each module
(5 x 7 marks = 35marks)

6. If the probability mass function of a RV X is given by 𝑃𝑃(𝑋𝑋 = 𝑥𝑥) = 𝑘𝑘𝑥𝑥 3 , 𝑥𝑥 = 1,2,3,4.

1 3
Find the value of 𝑘𝑘, 𝑃𝑃 ��2 < 𝑋𝑋 < 2�⁄𝑋𝑋 > 1�,mean and variance of 𝑋𝑋.
7. If the moment generating function of a uniform distribution for a random variable X is
1
(𝑒𝑒 5𝑡𝑡 − 𝑒𝑒 4𝑡𝑡 ).Find E(X).
𝑡𝑡
8. Consider the Markov chain with three states, s={1,2,3} that has the following
1 1 1
⎡2 4 4⎤
⎢1 2⎥
transition matrix 𝑃𝑃 = ⎢3 0 3⎥
Draw the state diagram for the chain. If 𝑃𝑃(𝑋𝑋1 =
⎢1 1 ⎥
⎣2 2
0⎦
1
1) = 𝑃𝑃(𝑋𝑋2 = 2) = 4, find 𝑃𝑃(𝑋𝑋1 = 3, 𝑋𝑋2 = 2, 𝑋𝑋3 = 1).

2 2 1
9. Find the eigen values and eigen vectors of 𝐴𝐴 = �1 3 1�.
1 2 2
 1 2
10. Find the least square solution to the equation 𝐴𝐴𝐴𝐴 = 𝑏𝑏, where 𝐴𝐴 = �13� and 𝑏𝑏 =
0 0
4
�5�, Obtain the projection matrix 𝑃𝑃 which projects 𝑏𝑏 on to the column space of 𝐴𝐴.
6
3 2
11. Let T be the linear transformation from R to R defined by T(x,y,z) =(x+y, 2z-x). Let
B1, B2 be standard ordered bases of R3 and R2 respectively. Compute the matrix of T
relative to the pair B1, B2.
12. Let V be a finite-dimensional complex inner product space, and let T be any linear
operator on V. Show that there is an orthonormal basis for V in which the matrix of T
is upper triangular.

**********************************
 Syllabus

Module 1 Axiomatic definition of probability. Independence. Bayes’ theorem and

applications. Random variables. Cumulative distribution function, Probability Mass
Function, Probability Density function, Conditional and Joint Distributions and densities,
Independence of random variables. Functions of Random Variables: Two functions
of two random variables. Pdf of functions of random variables using Jacobian.

Module 2 Expectation, Fundamental theorem of expectation, Moment generating functions,

Characteristic function. Conditional expectation. Covariance matrix. Uncorrelated
random variables. Pdf of Jointly Gaussian random variables, Markov and Chebyshev
inequalities, Chernoff bound. Central Limit theorem. Convergence of random variables.
Weak law of large numbers, Strong law of large numbers.

Module 3 Random Processes. Poisson Process, Wiener Process, Markov Process, Birth-
Death Markov Chains, Chapman- Kolmogorov Equations,

Groups, Rings, homomorphism of rings. Field. Vector Space. Subspaces. direct sum. Linear
independence, span. Basis. Dimension. Finite dimensional vector spaces. Coordinate
representation of vectors. Row spaces and column spaces of matrices.

Module 4 Linear Transformations. Four fundamental subspaces of a linear transformation.

Rank and Rank-nullity theorem. Matrix representation of linear transformation. Change
of basis transformation. System of linear equations. Existence and uniqueness of
solutions. Linear functionals. Dual, double dual and transpose of a linear transformation.

Module 5 Eigen values, Eigen vectors, Diagonizability.

Inner product. Norm. Projection. Least-squares solution. Cauchy-Schwartz inequality.

Orthonormal bases. Orthogonal complement. Spectral decomposition theorem.
Course Plan

Topic No. of
No
Lectures
Module I
Axiomatic definition of probability. Independence. Bayes’
1.1 2
theorem and applications.
Random variables. Cumulative distribution function, Probability
1.2 1
Mass Function,
Probability Density function, Conditional and Joint Distributions
1.3 2
and densities, Independence of random variables.
Functions of Random Variables: Two functions of two random
1.4 2
variables. Pdf of functions of random variables using jacobian.
Module II
Expectation, Fundamental theorem of expectation, Conditional
2.1 1
expectation.
2.2 Moment generating functions, Charectristic function. 1
Covariance matrix. Uncorrelated random variables. Pdf of Jointly
2.3 2
Guassian random variables,
Markov and Chebyshev inequalities, Chernoff bound. Central
2.4 2
Limit theorem.
Convergence of random variables. Weak law of large numbers,
2.5 2
Strong law of large numbers.
3 Module III
3.1 Random Processes. Poisson Process, Wiener Process, 2
Markov Process, Birth-Death Markov Chains, Chapman-
3.2 2
Kolmogorov Equations,
Groups, Rings, homomorphism of rings. Field. Vector Space.
3.3 2
Subspaces. direct sum.
Linear independence, span. Basis. Dimension. Finite dimensional
3.4 2
vector spaces.
Coordinate representation of vectors. Rowspaces and column
3.5 1
spaces of matrices.
4 Module IV
Linear Transformations. Four fundamental subspaces of a linear
4.1 2
transformation. Rank and Rank-nullity theorem.
Matrix representation of linear transformation. Change of basis
4.2 1
transformation.
4.3 System of linear equations. Existence and uniqueness of solutions. 2
Linear functionals. Dual, double dual and transpose of a linear
4.4 2
transformation.
 5 Module V
5.1 Eigen values, Eigen vectors, Diagonizability. 2
Inner product. Norm. Projection. Least-squares solution. Cauchy-
5.2 2
Schwartz inequality.
Orthonormal bases. Orthogonal complement. Spectral
5.3 2
decomposition theorem.

Reference Books

1. Hoffman Kenneth and Kunze Ray, Linear Algebra, Prentice Hall of India.

2. Jimmie Gilbert and Linda Gilbert, Linear Algebra and Matrix Theory, Elsevier

3. Henry Stark and John W. Woods "Probability and Random Processes withApplications to
Signal Processing", Pearson Education, Third edition.

4. Athanasios Papoulis and S. Unnikrishna Pillai. Probability, Random Variables and

Stochastic Processes, TMH
 CODE COURSE CATEGORY L T P CREDIT
ADVANCED DIGITAL PROGRAM
221TEC003 3 0 0 3
SIGNAL PROCESSING CORE 1

Preamble: The course is intended to impart comprehensive knowledge in the domain of

advanced digital signal processing

Prerequisite: Digital Signal Processing

Course Outcomes: After the completion of the course the student will be able to

CO 1 Assimilate the Non-parametric and parametric methods for spectral estimation

CO 2 Distinguish between forward and backward linear prediction

CO 3 List the utilities of adaptive filters

CO 4 Illustrate the use of Levinson Durbin algorithm for the solution of normal equations
Compare and contrast LMS algorithm and RLS Algorithm for Adaptive Direct form
CO 5
FIR filters
Develop the efficient realization of QMF filter bank using polyphase decomposition
CO 6
and multirate identities

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 3 3 2 2
CO 2 3 3 2 2
CO 3 3 2 1 - -
CO 4 3 3 3 3 2
CO 5 3 3 2 2
CO 6 3 2 2 3 2

Assessment Pattern

Bloom’s Category End Semester Examination

Apply 80%
Analyse 20%
Evaluate -
Create -
Mark distribution

Total ESE
CIE ESE
Marks Duration
100 40 60 2.5 hours

CORE COURSES

Evaluation shall only be based on application, analysis or design based questions (for both
internal and end semester examinations).

Continuous Internal Evaluation: 40 marks

Micro project/Course based project: 20 marks

Course based task/Seminar/ Quiz: 10 marks

Test paper, 1 no: 10 marks The project shall be done individually. Group projects not
permitted. Test paper shall include minimum 80% of the syllabus.

End Semester Examination: 60 marks

The end semester examination will be conducted by the University. There will be two parts;
Part A and Part B. Part A contain 5 numerical questions (such questions shall be useful in the
testing of knowledge, skills, comprehension, application, analysis, synthesis, evaluation and
understanding of the students), with 1 question from each module, having 5 marks for each
question. Students shall answer all questions. Part B contains 7 questions (such questions
shall be useful in the testing of overall achievement and maturity of the students in a course,
through long answer questions relating to theoretical/practical knowledge, derivations,
problem solving and quantitative evaluation), with minimum one question from each module
of which student shall answer any five. Each question can carry 7 marks. Total duration of
the examination will be 150 minutes.
Model Question Paper

APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY

M.TECH DEGREE EXAM

221TEC003 ADVANCED DIGITAL SIGNAL PROCESSING

TIME :2.5 HRS MAX MARKS:60

PART A

ANSWER ALL QUESTIONS(5*5=25)

1.Prove that the energy density spectrum of a deterministic signal can be obtained from the
Fourier transform of that signal?

2.Differentiate between the MDL criterion and CAT criterion towards the selection of an AR
Model order?

3.Use the Levinson Durbin algorithm to solve the normal equations recursively for an m-step
forward predictor?

4.In adaptive filtering ,iterative schemes are preferred over linear estimations. Justify?

5.Obtain the frequency domain characterization of a sampling rate converter of a rational

factor 2 /3 ?

PART B

ANSWER ANY FIVE.EACH QUESTION CARRIES 7 MARKS(7*5=35)

6.With necessary equations ,substantiate the use of the Welch method-Averaging Modified
Periodogram for spectrum estimation?

7.Ilustrate, the performance characteristics & Computational requirements of non-parametric

methods for spectral estimation?

8.With necessary equations, explain the Yule‐Walker method for AR model parameters in the
context of parametric spectral estimation?

9.Illustrate how Schur Algorithm can be utilized for the Solution of the Normal Equations?

10.Describe an adaptive equalizer using RLS algorithm?

11.What is the PR condition for a filterbank? Design the analysis filter bank for a perfect
reconstruction two channel Quadrature Mirror filterbank?

12.a.Realize a 3 band FIR filter of length N=15 using polyphase decomposition?

b.Express the output y[n] of the multirate system given below, as a function of the input x[n].
Syllabus and Course Plan(For 3 credit courses, the content can be for 40 hrs and for 2 credit
courses, the content can be for 26 hrs)

Syllabus

Power spectrum estimation, Non-parametric methods for spectral estimation: Parametric

spectral estimation, Linear Prediction and optimum linear filters, Adaptive filters for adaptive
channel equalization,adaptive noise cancellation, Adaptive Direct form FIR filters: Multirate
Signal Processing, The Polyphase decomposition-Applications to sub band coding, Fourier
transform,Short-time(windowed) Fourier transform. The discrete wavelet transform.

Course Plan

No. of
No Topic
Lectures
1 Power spectrum estimation
Estimation of spectra from finite duration observation of signals:
Computation of Energy density spectrum-Estimation of the
1.1 2
Autocorrelation and power spectrum of random signals ‐The
periodogram,Use of DFT in power spectrum estimation -
Non-parametric methods for spectral estimation:
1.2 Barlett method-Averaging Periodogram, Welch method-Averaging 3
Modified Periodogram
Blackman and Tukey Method‐Performance characteristics &
1.3 Computational requirements of non-parametric methods for 3
spectral estimation:
2 Parametric spectral estimation
Parametric spectral estimation:
2.1 2
Relationship between Autocorrelation and Model parameters
Yule‐Walker method for AR model parameters, Burg method for
2.2 3
AR model parameters
Selection of AR model order‐ MA and ARMA models for power
2.3 4
spectrum estimation
3 Linear Prediction and optimum linear filters
 Linear Prediction : Forward and Backward Linear Prediction
3.1 Optimum reflection coefficients for the Lattice Forward and 3
Backward Predictors.
Solution of the Normal Equations: Levinson Durbin Algorithm,
3.2 3
Schur Algorithm
3.3 Properties of Linear Prediction Filters 2
4 Adaptive filters
Adaptive filters for adaptive channel equalization,adaptive noise
4.1 3
cancellation and Linear Predictive Coding of Speech Signals
Adaptive Direct form FIR filters:Minimum mean square
4.2 2
criteria,LMS algorithm
Adaptive Direct form filters:The RLS algorithm,Fast RLS
4.3 3
Algorithm,Properties of Direct Form RLS algorithm
5 Multirate Signal Processing
Mathematical description of sampling rate converters- Interpolator
5.1 2
and Decimator,Multirate Identities
The Polyphase decomposition-Applications to sub band coding -
5.2 3
Two Channel QMF filer bank-PR condition.
Fourier transform,Short-time(windowed) Fourier transform,The
5.3 discrete wavelet transform-Wavelet- admissibility condition.MRA 3
Axioms,scaling and wavelet function

Text books:

1. John G. Proakis, Dimitris G. Manolakis, “Digital Signal Processing”, Pearson

Education,India, 2007

2. Monson H. Hayes, “Statistical Digital Signal Processing and Modeling”, John Wiley and
Sons Inc., New York, 2006.

3. P. Vaidyanathan, “Multirate Systems and Filter Banks”, Pearson Education

References:

1. Simon Haykin, “Adaptive Filter Theory”, Prentice Hall, Englewood Cliffs, NJ1986.
2.Steven M Kay,” Modern spectrum Estimation theory and application”, Pearson
India,January 2009
3.D.G. Manolakis, V.K. Ingle and S.M. Kogon: Statistical and Adaptive Signal Processing,
McGraw Hill, 2000
4.Sophoncles J. Orfanidis, “Optimum Signal Processing “, McGraw-Hill, 2000
5. S K Mitra,”Digital Signal Processing: A computer based approach”, Tata-McGraw Hill
4. C S Burrus, R A Gopinath, H. Guo, “Introduction to Wavelets and Wavelet
Transforms: A primer”, Prentice Hall.
 CODE COURSE NAME CATEGORY L T P CREDIT
TOPICS IN MACHINE PROGRAM
221TEC004 3 0 0 3
LEARNING CORE 2

Preamble: Machine learning is a subfield of artificial intelligence, which is broadly defined

as the capability of a machine to imitate intelligent human behavior. Artificial intelligence
systems are used to perform complex tasks in a way that is similar to how humans solve
problems. Machine learning allows software applications to become more accurate at
predicting outcomes without being explicitly programmed to do so. Machine learning
algorithms use historical data as input to predict new output values. Students will be able to
learn machine learning fundamentals, understand the different types of algorithms in machine
learning, develop in-depth knowledge of machine learning tasks such as regression,
classification, clustering etc.

Prerequisite: A sound knowledge of the fundamentals and basics of probabilty, statistics

and algorithms.

Course Outcomes: After the completion of the course the student will be able to
Understand and apply the fundamentals, concepts and terminologies in machine
CO 1 learning, Deep learning and artificial intelligence.

Understand and analyse the principles of supervised and unsupervised learning and
CO 2 illustrate the functionalities of the supervised and unsupervised learning algorithms.

Understand and analyse the principles of semi-supervised and reinforcement learning

CO 3 and illustrate the functionalities of the semi-supervised and reinforcement learning
learning algorithms.

Analyze and evaluate the performance of artificial neural networks and deep learning
CO 4 neural architectures.

Create and evaluate critically the domain specific applications of Machine learning.
CO 5

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 2 3 2

CO 2 2 3 2

CO 3 3 2 2 2 3

CO 4 2 2 3

CO 5 3 3 2
Assessment Pattern

Bloom’s Category End Semester Examination

%
Apply 10
Analyse 40
Evaluate 30
Create 20

Mark distribution

Total
CIE ESE ESE Duration
Marks

100 40 60 2.5 hours

Continuous Internal Evaluation Pattern:

Evaluation shall only be based on application, analysis or design based questions (for
both internal and end semester examinations).

Continuous Internal Evaluation: 40 marks

Micro project/Course based project: 20 marks
Course based task/Seminar/Quiz: 10 marks
Test paper, 1 no: 10 marks

The project shall be done individually. Group projects not permitted. Test paper shall
include minimum 80% of the syllabus.

End Semester Examination: 60 marks

The end semester examination will be conducted by the University. There will be two parts;
Part A and Part B. Part A contain 5 numerical questions (such questions shall be useful in the
testing of knowledge, skills, comprehension, application, analysis, synthesis, evaluation and
understanding of the students), with 1 question from each module, having 5 marks for each
question. Students shall answer all questions. Part B contains 7 questions (such questions
shall be useful in the testing of overall achievement and maturity of the students in a course,
through long answer question relating to theoretical/practical knowledge, derivations,
problem solving and quantitative evaluation), with minimum one question from each module
of which student shall answer any five. Each question can carry 7 marks. Total duration of
the examination will be 150 minutes.
Model Question Paper

APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY

FIRST SEMESTER M.TECH DEGREE EXAMINATION, (Model Question Paper)

Discipline: ELECTRONICS & COMMUNICATION

ENGINEERING
Stream: EC3 (Signal Processing, Signal Processing
& Embedded Systems, Communication
Engineering & Signal Processing)
Course Code: 221TEC004

Course Name: TOPICS IN MACHINE LEARNING

Max. Marks: 60 Duration: 2.5Hours

PART A
Answer ALL Questions. Each Question Carries 5 marks.

A sample x1, ..., xn has been obtained from a probability model

1 specified by mass or density function fX (x; θ) depending on CO1
parameter(s) θ lying in parameter space Θ. Estimate the
maximum likelihood.

Illustrate with an example, the k- nearest neighbour algorithm for

2 pattern classification. How do you compute the distance between CO2
two patterns ?

3 Explain how semi-supervised learning is different from CO2

supervised learning. Illustrate self learning.

4 Demonstrate gradient descent algorithm for error back CO2

propagation in ANNs.

Suggest a strategy based on machine learning techniques for

5 sentiment analysis in social networks. Explain with neat CO3
schematics.

PART – B
Answer any 5 full questions ; each question carries 7 marks.

6. a) With neat schematics, explain machine learning process flow. Explain the 5 CO4
process involved in model building and validation, interpretaion of model
and data visualization.
6. b) What do you mean by hyperparameters ? Explain the process of fine 2 CO3
tuning the hyper parameters.
7.a) Discuss the principles of density based clustering and explain the basic 3 CO1
algorithm. What do you mean by ε -neighbourhood ?
7. b) With the help of relevant equation explain how do you compute cosine 4 CO5
similarity between two feature vectors.

8.a) State the dual problem of optimization for SVM for linearly non- 2 CO2
separable patterns. Obtain the equation of optimal hyperplane and
optimum weight vectors wo and optimum bias bo for linearly non-
separable patterns. Give expression for finding the label of a test example
xt.
8.b) Explain what do you mean by hard and soft margins. Derive the 5 CO4
expression for the margin of separation for SVM.
9. a) Explain the concept of error back propagation learning used in MLFFNN. 3 CO4
Obtain the weight updation equation for MLFFNN with one hidden layer.
9.b) Discuss the significance of activation functions. With the help of a neat 3 CO1
sketch explain sigmoidal activation function. Explain how spread factor
(β) value is selected
9.c) Discuss the functionalities of different layers in a CNN. 1 CO2

10.a) Compare and contrast between agglomerative and divisive techniques for 2 CO1
clustering. Give the step-by-step algorithm for agglomerative clustering.
What are the proximity measures between clusters ?
10.b) With the help of suitable illustrations explain the selection of number of 5 CO3
clusters using silhouette analysis.
11.a) Discuss the basic concepts of deep learning. With a neat schematic 3 CO1
explain the principles of convolutional neural networks.
11.b) Discuss the need for pooling layer in a CNN architecture. Distinguish 4 CO3
between max pooling and average pooling.
12.a) Illustrate the use of machine learning algorithms in medical image 3 CO5
segmentation for brain tumor detection. Draw neat schematics for the
system and expalin the functionality of each module.
12.b) Explain the use of machine learning algorithms and tools for rain 4 CO5
forecasting in Kerala. Explain the data collection process, data refinement
and models used for prediction. How do you validate the model?
 Syllabus

Module - 1 (Basics of Machine Learning )

Introduction to machine learning, artificial intelligence and deeplearning. Learning
algorithms - over fitting and under fitting, hyperparameters and validation sets, estimators,
bias and variance, Maximum Likelihood Estimation. Machine learning process flow- define
problem, objective, data acquisition and preprocessing, feature engineering, model building
and validation.

Module -2 (Supervised and Unsupervised Learning)

Supervised Learning- Basic principles of linear regression, logistic regression. Classification-
Supervised algorithms-Decision trees, k-Nearest Neighbour, Naive Bayes, support vector
machines, ensemble learning techniques. Unsupervised Learning- Basic principles of
clustering, clustering algorithms-hierarchical algorithms-agglomerative, divisive algorithms.
Partitioning algorithms- k-means, k medoids algorithms, density based algoritms.

Module -3 (Semi-supervised and Reinforcement Learning)

Semi-supervised learning – Types of semi-supervised learning- Self learning, graph based
SSL-label propagation. Reinforcement Learning-Taxonomy, Reinforcement Learning
Algorithms-Value based, Policy based and model based algorithms. Characteristics and types
of reinforcement learning. Reinforcement learning models- Markov decision process, Q-
learning.

Module -4 (Artificial Neural Networks and Deep Learning)

Artificial neural networks- Basic principles of Back propagation, Gradient Descent, Training
Neural Network, Initialisation and actvation functions. Deep learning principles and
achitectures-Dropout, Batch normalisation, Ensemble learning, Data augmentation, Transfer
learning, Convolutional Neural Networks, Recurrent Neural Networks, LSTM, Data
augmentation-GAN.

Module- 5 (Applications of Machine Learning)

Machine learning applications for prediction-weather, sales of a store, eligibility of loan.
Medical diagnoses, Financial industry and trading, image classification, recognition and
segmentation, speech recognition, automatic language translation and auto corrections,
recommendation engines.
Course Plan

No. of Lectures
No Topic
[40Hrs]
1 Basics of machine learning.
1.1 Introduction to machine learning. 1
1.2 Artificial intelligence and deeplearning. 1
Learning algorithms-over fitting and under fitting,
1.3 2
hyperparameters and validation sets.
1.4 Estimators, bias and variance, Maximum Likelihood Estimation. 1
Machine learning process flow- define problem, objective, data
acquisition and preprocessing, feature engineering, model building 2
and validation.
2 Semi-supervised and Reinforcement Learning
Supervised Learning- Basic principles of linear regression, logistic
2.1 2
regression.

Classification-Supervised algorithms-Decision trees, k-Nearest

2.2 2
Neighbour, Naive Bayes.

2.3 Support vector machines, ensemble learning techniques. 2

2.4 Unsupervised Learning- Basic principles of clustering. 1

Clustering algorithms-hierarchical algorithms-agglomerative,

2.5 divisive algorithms. Partitioning algorithms- k-means, k medoids 2
algorithms, density based algoritms.

Partitioning algorithms- k-means, k medoids algorithms, density

2.6 based algoritms. Case study in clustering- Medical image 2
segmentation.
3 Semi-supervised and Reinforcement Learning
Semi-supervised learning – Types of semi-supervised learning-
3.1 2
Self learning, graph based SSL-label propagation.
Reinforcement Learning-Taxonomy, Reinforcement Learning
Algorithms-Value based, Policy based and model based
3.2 algorithms. Characteristics and types of reinforcement learning. 3
Reinforcement learning models- Markov decision process, Q-
learning.
Characteristics and types of reinforcement learning.
3.3 Reinforcement learning models- Markov decision process, Q- 3
learning.
4 Artificial Neural Networks and Deep Learning
 Artificial neural networks- Basic principles of Back propagation,
4.1 2
Gradient Descent.
4.2 Training Neural Network, Initialisation and actvation functions.

Deep learning principles and achitectures-Dropout, Batch

4.3 normalisation, Ensemble learning, Data augmentation, 2
Transferlearning.

Convolutional Neural Networks, Recurrent Neural Networks,

4.4 LSTM, Data augmentation-GAN. 3

5 Applications of Machine Learning

Machine learning applications for prediction-weather, sales of a
5.1 2
store, eligibility of loan.
5.2 Medical diagnoses, Financial industry and trading. 2
5.3 Image recognition, classification and segmentation. 1
Speech recognition, automatic language translation and auto
5.4 2
corrections, recommendation engines.

Text Books

1. Machine Learning: The Art and Science of Algorithms that Make Sense of Data, Peter A
Flach, Cambridge University Press, ISBN-10 1107422221, 2012.
2. Applied Machine Learning, 2nd Edition, M. Gopal, Mc Graw Hill Education, ISBN-10 :
‎9789353160258, 2018.
3. Neural Networks and Learning Machines, Simon S. Haykin, 3rd Edition, Pearson-Prectice
Hall, ISBN-10: 0-13-147139-2, 2009.

Reference Books

1. An Introduction to Machine Learning, Miroslov Kubat, Springer, ISBN-10 3030819345,

2021.

2. Machine Learning, 1st Edition, Saika Dutt, Subramanian Chandramouli, Amit Kumar Das,
Pearson Education, ISBN-10 9353066697, 2018.

3. Machine Learning: A First Course for Engineers and Scientists, Andreas Lindholm, Niklas
Wahlstom, Fredrik Lindsten et al., Cambridge University Press, ISBN-10 1108843603,
2022.
4. Handbook of Reinforcement Learning and Control, Kyriakos G. Vamvoudakis, Yan Wan,
et al., ISBN-10 3030609898, Sprimger, 2021.

5. Neural Networks and Deep Learning, Charu C. Aggarwal, Springer, ISBN: 978-3-319-
94463-0, 2018.
 CODE COURSE NAME CATEGORY L T P CREDIT
LABORATORY
221LEC001 SIGNAL PROCESSING LAB 1 0 0 2 1
1

Preamble: To experiment the concepts introduced in the topics : Linear Algebra, Random
processes, Advanced Signal Processing and Machine Learning

Course Outcomes: After the completion of the course the student will be able to

CO 1 Apply knowledge of Linear algebra, Random processes,Advanced Signal Processing

and Machine Learning in various signal processing applications.
CO 2 Develop the student’s ability on analysing observations of experiments/ simulations,
interpreting them and preparing reports
CO 3 Apply the fundamental principles of linear algebra and random processes
CO 4 Familiarize the basic operations of filter banks through simulations
CO 5 Implement the basic algorithms learned in Machine learning
CO 6 Implement a mini project pertaining to an application of Signal Processing in real life

Assessment Pattern

Bloom’s Category CIE

Apply 40
Analyse 30
Evaluate 15
Create 15

Mark distribution

Total
CIE ESE
Marks

100 100 --

Continuous Internal Evaluation Pattern:

Tools :
Numerical Computing Environment – MATLAB or any other equivalent tool.
Syllabus

No Topics
1 Linear Algebra
1.1 Row Reduced Echelon Form: To reduce the given mxn matrix into Row reduced
Echelon form
1.2 Gram-Schmidt Orthogonalization: To find orthogonal basis vectors for the given
set of vectors. Also find orthonormal basis.
1.3 Least SquaresFit to a Sinusoidal function
1.4 Least Squares fit to a quadratic polynomial
1.5 Eigen Value Decomposition
1.6 Singular Value Decomposition
1.7 Karhunen- Loeve Transform
2 Advanced DSP
2.1 Sampling rate conversion: To implement Down sampler and Up sampler and
study their characteristics
2.2 Two channel Quadrature Mirror Filterbank: Design and implement a two channel
Quadrature Mirror Filterbank
3 Random Processes
3.1 To generate random variables having the following probability distributions (a)
Bernoulli(b) Binomial(c) Geometric(d) Poisson(e)Uniform,(f)
Gaussian(g)Exponential (h) Laplacian
3.2 Central Limit Theorem: To verify the sum of sufficiently large number of
Uniformly distributed random variables is approximately Gaussian distributed
and to estimate the probability density function of the random variable.
4 Machine Learning
4.1 Implementation of K Nearest Neighbours Algorithm with decision region plots
4.2 Implementation of K Means Algorithm with decision region plots
4.3 Implementation of Perceptron Learning Algorithms with decision region plots
4.4 Implementation of SVM algorithmfor classification applications
5 Implement a mini project pertaining to an application of Signal Processing
in real life, make a presentation and submit a report
 SEMESTER I
PROGRAM ELECTIVE I
 CODE COURSE NAME CATEGORY L T P CREDIT
ADVANCED DIGITAL PROGRAM
221EEC012 3 0 0 3
COMMUNICATION ELECTIVE 1

Preamble:Digital communications is a broad term that incorporates all procedures and

forms of transmission of data or information. This course imparts mathematical modelling
about various modulation schemes, channels and multipath mitigation techniques.

Course Outcomes: After the completion of the course the student will be able to

CO 1 Represent digitally modulated signals in signal space.

CO 2 Design of Optimum receiver for AWGN channel.
CO 3 Design of Equalizers for optimum detection in presence of ISI
CO 4 Analyse Multi Channel and Multi Carrier Systems
CO 5 Evaluate Digital Communication through Fading Multipath Channels
CO 6 Analysis of CDMA systems

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 3 3 2 3 3
CO 2 3 3 2 3 3 3 2
CO 3 3 2 2 3 3
CO 4 3 2 2 3 3 3 2
CO 5 3 2 2 3 3 3 2
CO 6 3 3 2 3 3 3 2

Assessment Pattern

Bloom’s Category End Semester Examination

%
Apply 20
Analyse 50
Evaluate 30
Create

Mark distribution

Total ESE
CIE ESE
Marks Duration

100 40 60 2.5 Hrs

Evaluation of Elective Courses

Evaluation shall only be based on application, analysis or design based questions (for both
internal and end semester examinations).

Continuous Internal Evaluation: 40 marks

Preparing a review article based on peer reviewed Original publications (minimum 10

publications shall be referred) : 15 marks

Course based task/Seminar/Data collection and interpretation : 15 marks

Test paper, 1 no. : 10 marks

Test paper shall include minimum 80% of the syllabus.

End Semester Examination: 60 marks

The end semester examination will be conducted by the respective College.

There will be two parts; Part A and Part B.

Part A will contain 5 numerical/short answer questions with 1 question from each module,
having 5 marks for each question (such questions shall be useful in the testing of knowledge,
skills, comprehension, application, analysis, synthesis, evaluation and understanding of the
students). Students should answer all questions.

Part B will contain 7 questions (such questions shall be useful in the testing of overall
achievement and maturity of the students in a course, through long answer questions relating
to theoretical/practical knowledge, derivations, problem solving and quantitative evaluation),
with minimum one question from each module of which student should answer any five.
Each question can carry 7 marks.

Note: The marks obtained for the ESE for an elective course shall not exceed 20% over the
average ESE mark % for the core courses. ESE marks awarded to a student for each elective
course shall be normalized accordingly. For example if the average end semester mark % for
a core course is 40, then the maximum eligible mark % for an elective course is 40+20 = 60
%.
Model Question Paper

APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY

M.TECH DEGREE EXAMINATION
First Semester

Branch:

ADVANCED DIGITAL COMMUNICATION

Time: 2.5 Hours Marks: 60

Part A

Answer ALL Questions. Each question carries 5 marks

1. Represent 8-PSK as a linear combination of two orthonormal signal waveforms.

Using Gray encoding label the corresponding signal points
2. Derive the impulse response of a matched filter.
3. Explain Nyquist criteria for zero ISI. Illustrate the pulse shape for zero ISI.
4. Design a 4 bit PN sequence generator with shift registers.
5. Explain RAKE receiver.

Part B

Answer ANY FIVE Questions. Each question carries 7 marks

6. Three messages are transmitted over an AWGN channel with noise power spectral
density N0/2 The messages are

1 ; 0 ≤ 𝑡𝑡 ≤ 𝑇𝑇
S1(t) = �
0 ; 𝑜𝑜𝑜𝑜ℎ𝑒𝑒𝑒𝑒 𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤
1 ; 0 ≤ 𝑡𝑡 ≤ 𝑇𝑇/2
𝑇𝑇
S2(t)= −S3(t) = � −1; ≤ 𝑡𝑡 ≤ 𝑇𝑇
2
0; 𝑜𝑜𝑜𝑜ℎ𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒

i. Find the dimensionality of the signal space.

ii. Find and draw an appropriate basis for the signal space.
7. Derive the expression for error probability of binary QPSK modulation scheme in
AWGN channel.
8. In an un-equalized linear filter channel the noise free output of the demodulator when
a ‘1’ is transmitted is
0.3; m = 1

xm = 0.9; m = 0
0.3; m = −1
0 otherwise

1; m = 0
Design a three tap zero forcing equalizer so that the output is qm = 
0; m = ±1
9. Explain an FFT based multi-carrier System.
 10. Explain the Delay locked loop for tracking DS spread spectrum signal.
11. A multipath fading channel has a multipath spread of Tm=1 sec and a Doppler
spread Bd = 0.01 Hz. The total channel bandwidth at band-pass available for signal
transmission is W=5 Hz. To reduce ISI the selected pulse duration is T= 10 sec.
(i)Determine the coherence Bandwidth and Coherence Time?
(ii)Is the channel is frequency selective?
12. Explain the ALOHA system and protocols in multiuser communication system.

Syllabus

Module I

Characterization of Communication Signals and Systems: Overview of Digital

Communication systems,Communication Channels and Mathematical models,Representation
of band pass signals and systems, Signal spacerepresentation,Representation of digitally
modulated signals, Spectral Characteristics of Digitally Modulated Signals.

Module II

Optimum receiver for AWGN channel:Correlationdemodulator, matched filter demodulator,

optimum detector,Performance of optimum receiver for memoryless modulationtechniques,
Probability of error for binary modulation and M-ary orthogonal signals, Probability of
errorQPSK, QAM.

Module III

Communication through Band-Limited Linear Filter channelsOptimum receiver for channels

with ISI andAWGN,Equalization techniques: Linear equalization, Decisionfeedback
equalization,Adaptive equalization Algorithms(ZF and LMS).

Module IV

Multi Channel and Multi Carrier Systems: Multichannel Digital communication in AWGN
channels,Multicarrier communication- Discrete implementation of multicarrier modulation.
FFT based multi carrier system,Spread spectrum principles, Generation of PN sequences,
DSSS & FHSS, Synchronization of Spread Spectrum signals.

Module V

Digital Communication through Fading Multipath Channels: Characterisation of fading

multipath channel,Frequency-nonselective slowly fading channel, Diversity
techniques,Digital signalling over a frequencyselectiveslowly fading channel, RAKE
receiver,Multiple access techniques- CDMA signal and channel models.
Course Plan

No Topic No. of
Lectures
1 Characterization of Communication Signals and Systems:
1.1 Overview of Digital Communication systems. 1
1.2 Communication Channels and Mathematical models 2
1.3 Representation of band pass signals and systems, Signal
2
spacerepresentation.
1.4 Representation of digitally modulated signals 2
1.5 Spectral Characteristics of Digitally Modulated Signals. 2

2 Optimum receiver for AWGN channel:

2.1 Correlationdemodulator, matched filter demodulator, optimum
3
detector.
2.2 Performance of optimum receiver for memoryless
modulationtechniques: probability of error for binary modulation 3
and M-aryorthogonal signals
2.3 Probability of errorQPSK, QAM. 2

3 Communication through Band-Limited Linear Filter channels

3.1 Optimum receiver for channels with ISI and
3
AWGN.
3.2 Equalization techniques: Linear equalization, Decision
3
feedback equalization,
3.3 Adaptive equalization: Algorithms(ZF and LMS) 2

4 Multi Channel and Multi Carrier Systems

4.1 Multichannel Digital communication in AWGN channels 2
4.2 Multicarrier communication: Discrete implementation of
2
multicarrier modulation. FFT based multi carrier system
4.3 Spread spectrum principles, Generation of PN sequences, 2
4.4 Direct sequence spread spectrum (DSSS), Frequency Hopping
SpreadSpectrum (FHSS), Synchronization of Spread Spectrum 3
signals.

5 Digital Communication through Fading Multipath Channels

5.1 Characterisation of fading multipath channel 1
5.2 Frequency-nonselective slowly fading channel 2
5.3 Diversity techniques for Fading Multipath channels 2
5.4 Digital signalling over a frequencyselectiveslowly fading channel.
2
RAKE receiver
5.5 Multiple access techniques- CDMA signal and channel models,
3
Random access Methods
Text Books

1. John G.Proakis, Digital Communications, 4/e, McGraw-Hill

Reference Books

1. Edward. A. Lee and David. G. Messerschmitt, “Digital Communication”, Allied

Publishers (second edition).
2. Viterbi, A. J., and J. K. Omura. Principles of Digital Communication and Coding. NY:
McGraw-Hill, 1979. ISBN: 0070675163.
3. Marvin K Simon, Sami M Hinedi, William C Lindsey - Digital Communication -
Techniques –Signal Design & Detection, PHI.
4. Bernard Sklar,” Digital Communications: Fundamentals and applications “, Prentice
Hall 2001.
5. Andrea Goldsmith,” Wireless Communications”, Cambridge University Press 2005.
 CODE COURSE NAME CATEGORY L T P CREDIT
PROGRAM
221EEC013 PATTERN ANALYSIS 3 0 0 3
ELECTIVE 1

Preamble : Pattern analysis is the use of machine learning algorithms to identify and
categorise patterns. It classifies data based on statistical information or knowledge gained
from patterns and their representation. The popular pattern analysis tasks are pattern
recognition, classification, clustering and retrival. Students will be able to learn pattern
analysis fundamentals, understand the different types of algorithms in pattern analysis,
develop in-depth knowledge of pattern analysis tasks such as classification, clustering,
matching, retrieval etc.

Prerequisite: A sound knowledge of the fundamentals and basics of probabilty, statistics

and algorithms.

Course Outcomes: After the completion of the course the student will be able to
Understand and apply the fundamentals, concepts and terminologies in pattern
CO 1 analysis.

Understand and analyse the principles of feature extraction and optimization and
CO 2 illustrate the functionalities of the feature extraction and optimization algorithms.

Understand and analyse the principles of supervised models for pattern analysis and
CO 3 illustrate the functionalities of the supervised pattern analysis algorithms.

Understand and analyse the principles of unsupervised models for pattern analysis and
CO 4 illustrate the functionalities of the unsupervised pattern analysis algorithms.

Create and evaluate critically the domain specific applications of pattern analysis.
CO 5

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 2 3 3

CO 2 2 2 2

CO 3 3 2 2 3 3

CO 4 2 2 2

CO 5 3 3 2
Assessment Pattern

Bloom’s Category End Semester Examination

Apply 20 %
Analyse 40 %
Evaluate 20 %
Create 20 %

Mark distribution

Total
CIE ESE ESE Duration
Marks

100 40 60 2.5 hours

Continuous Internal Evaluation: 40 marks

Preparing a review article based on peer reviewed

Original publications (minimum 10 Publications shall be referred): 15 marks

Course based task/Seminar/Data Collection and interpretation: 15 marks

Test paper, 1 no.: 10 marks

Test paper shall include minimum 80% of the syllabus. include minimum 80% of the syllabus.

End Semester Examination: 60 marks

The end semester examination will be conducted by the respective College. There will be two
parts; Part A and Part B. Part A will contain 5 numerical/short answer questions with 1
question from each module, having 5 marks for each question (such questions shall be useful
in the testing of knowledge, skills, comprehension, application, analysis, synthesis, evaluation
and understanding of the students). Students should answer all questions. Part B will contain
7 questions (such questions shall be useful in the testing of overall achievement and maturity
of the students in a course, through long answer questions relating to theoretical/practical
knowledge, derivations, problem solving and quantitative evaluation), with minimum one
question from each module of which student should answer any five. Each question can carry
7 marks.

Note: The marks obtained for the ESE for an elective course shall not exceed 20% for each
elective course shall be normalized accordingly. For example if the average end semester
mark % for a core course is 40, then the maximum eligible mark % for an elective course is
40+20 = 60 %.
Model Question paper

APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY

FIRST SEMESTER M.TECH DEGREE EXAMINATION, (Model Question Paper)

Discipline: ELECTRONICS & COMMUNICATION ENGINEERING

Stream: EC3 (Signal Processing, Signal Processing
& Embedded Systems, Communication
Engineering & Signal Processing)
Course Code: 221EEC013

Course Name: PATTERN ANALYSIS

Max. Marks: 60 Duration: 2.5Hours

PART A
Answer ALL Questions. Each Question Carries 5 marks.

Explain the concept of discriminant functions. Derive the

1 discriminant function for a two class classification problem. CO1
Assume patterns are drawn from normal distribution

2 Demonstrate dimensionality reduction of feature vectors using CO2

Fisher score based method.

Illustrate with an example, the k- nearest neighbour algorithm for

3 pattern classification. How do you compute the distance between CO2
two patterns ?

With a suitable example demostrate the k-means clustering

4 algorithm. Discuss the significance of silhoutte score to select CO2
optimal k.

What do you mean by out lier detection? Explain an

5 experimental set up for outlier detection from banking CO3
transaction data.

PART – B
Answer any 5 full questions ; Each question carries 7 marks.

6. a) The minimum error rate classification can be achieved by the use of 5 CO4
discriminant functions gi(x) = ln p(x|wi) + ln P(wi). The densities
p(x|wi) are assumed to be multivariate Gaussian. Determine the
discriminant function for (i) Σi = σ^2*I (ii) Σi = Σ and Σi =
arbitrary. Discuss the nature of decision surface in each case.
6. b) Illustrate the working of LDA with a suitable toy dataset. Also 2 CO3
explain how predictions are made.
7.a) Explain Bayes decision theory. Give expression for Bayes decision 3 CO1
rule. Discuss the significance.
7. b) Give the expresion for Bayesian decision function. Use Bayes 4 CO5
decision rule to find the answer to the following problem: Suppose a
drug test is 99% sensitive and 99% specific. That is, the test will
produce 99% true positive results for drug users and 99% true
negative results for non-drug users. Suppose that 0.5% of people are
users of the drug. If a randomly selected individual tests positive,
what is the probability that they are a user?

8.a) Compare and contrast between the use of DFT and DCT for 2 CO2
extracting transform based features from image data.
8.b) Discuss the significance of dimensionality reduction. Explain step- 5 CO4
by step algorithm for PCA
9. a) Discuss gradient descent algorithm used in MLFFNN. Also 3 CO4
comment on local and global minimum
9.b) Explain the concept of error back propagation learning used in 3 CO1
MLFFNN. Obtain the weight updation equation for MLFFNN with
one hidden layer.
9.c) Compare MLFFNN and perceptron based classifiers for pattern 1 CO2
analysis. Comment on the stopping criterion for both.
10.a) Explain what do you mean by hard and soft margins. Derive the 2 CO1
expression for the margin of separation for SVM.
10.b) State the dual problem of optimization for SVM for linearly non- 5 CO3
separable patterns. Obtain the equation of optimal hyperplane and
optimum weight vectors wo and optimum bias bo for linearly non-
separable patterns. Give expression for finding the label of a test
example xt.
11.a) Discuss the principles of fuzzy of k-means clustering 3 CO1

11.b) Illustrate hierarchical clustering for text classification of English 4 CO3

sentences. Clearly explain the methodology used for generating
feature vectors from sentences, similarity measures selected,
formation of similarity matrix and formation of clusters. Also
discuss the termination criteria.
12.a) List at least 3 pattern analysis tasks that we frequently encounter in 3 CO5
our real world life.
12.b) Illustrate the functionality of Music Information Retrieval system 4 CO5
with neat sketches. Design a MIR system. Explain the functionality
of each module.

Syllabus

Module - 1 ( Introduction to Pattern Analysis ) : Introduction - features, feature vectors

and classifiers, Supervised versus unsupervised pattern recognition. Classifiers based on
Bayes Decision theory- Review of Probability Theory, Conditional Probability and Bayes
rule. Bayesian classification for normal distributions, Bayes classifier, Case I, Case II and
Case III. Linear classifiers- Linear and quadratic discriminant functions and decision hyper
planes.

Module -2 (Feature extraction and optimization)

Review of Linear Algebra. Linear Transformations-KLT, SVD, ICA, DFT, DCT, DST,
Hadamard Transform, Wavelet Transform and transform based features. Feature engineering-
Feature extraction- Global and local features, features for shape and characterization, typical
features for object recognition in images, typical features for speech and audio classification,
Feature reduction and optimization- Dimensionality reduction techniques, Principal
component analysis.
Module 3 (Supervised Models for Pattern Classification )
K-Nearest-Neighbor Classification, selection criterion for k. The perceptron learning
algorithm, classifier using perceptron. Multi layer perceptrons- Multilayer feed forward
neural networks, Training neural networks, back propagation, gradient descent algorithm,
activation functions and performance analysis of MLFFNN. Introduction to CNN. Support
vector machines- SVM for linearly separable and nonlinearly separable patterns. Concept of
maximum margin, kernels for SVM, kernel trick.

Module -4 (Unsupervised Models for pattern Classification)

Clustering - Vector Quantization, k-means clustering, Silhouette score, Cluster analysis,

Proximity measures, Clustering Algorithms - Sequential algorithms, Hierarchical algorithms -
Agglomerative algorithms, Divisive algorithms. Clustering Schemes - based on function
optimization -Fuzzy clustering algorithms, Probabilistic clustering, Clustering algorithms
based on graph theory.

Module- 5 (Applications of Pattern Analysis)

Application of pattern analysis in image classification, speech recognition, speaker
identification, multimedia document recognition (MDR), automatic medical diagnosis.
Outlier detection, novelty/anomaly detection using pattern analysis. Music and image/video
retrieval systems.
Course Plan

Topic No. of Lectures

No
[40Hrs]
1 Introduction to Pattern Analysis
Introduction - features, feature vectors and classifiers, Supervised
1.1 2
versus unsupervised pattern recognition.
Classifiers based on Bayes Decision theory- Review of Probability
1.2 2
Theory, Conditional Probability and Bayes rule.
Bayesian classification for normal distributions, Bayes classifier,
1.3 3
Case I, Case II and Case III.
Linear classifiers- Linear and quadratic discriminant functions and
1.4 2
decision hyper planes.
2 Feature extraction and optimization
Review of Linear Algebra. Linear Transformations-KLT, SVD,
2.1 ICA, DFT, DCT, DST, Hadamard Transform, Wavelet Transform 2
and transform based features.

Feature engineering- Feature extraction- Global and local features,

2.2 3
features for shape and characterization.

Typical features for object recognition in images, typical features

2.3 2
for speech and audio classification.

Feature reduction and optimization- Dimensionality reduction

2.4 3
techniques.

2.5 Principal component analysis. 2

3 Supervised Models for Pattern Classification
K-Nearest-Neighbor Classification, selection criterion for k. The
3.1 3
perceptron learning algorithm, classifier using perceptron.
Multi layer perceptrons- Multilayer feed forward neural networks,
Training neural networks, back propagation, gradient descent
3.2
algorithm, activation functions and performance analysis of
MLFFNN. Introduction to CNN.
Support vector machines- SVM for linearly separable and
3.3 nonlinearly separable patterns. Concept of maximum margin, 3
kernels for SVM, kernel trick.
4 Unsupervised Models for pattern Classification
Clustering - Vector Quantization, K-means clustering, Silhouette
4.1 3
score.
Cluster analysis, Proximity measures, Clustering Algorithms -
4.2
Sequential algorithms, Hierarchical algorithms - Agglomerative
 algorithms, Divisive algorithms.

Clustering Schemes based on function optimization-Fuzzy

4.3 clustering algorithms, Probabilistic clustering, 2

4.4 Clustering algorithms based on graph theory. 3

5 Applications of Pattern Analysis
Application of pattern analysis in image classification, speech
5.1 2
recognition, speaker identification.
Multimedia document recognition (MDR), automatic medical
5.2 2
diagnosis.
Outlier detection, novelty/anomaly detection using pattern
5.3 1
analysis.
5.4 Music and image/video retrieval systems. 2

Text Books

1. Pattern classification , Richard O. Duda and Hart P.E, and David G Stork, , 2nd Edn., John
Wiley & Sons Inc., 2001

2. Neural Networks and Learning Machines, Simon S. Haykin, 3rd Edition, Pearson-Prectice
Hall, ISBN-10: 0-13-147139-2, 2009.

3. Pattern Recognition, Sergios Theodoridis, Konstantinos Koutroumbas , Academic Press, 2006.

Reference Books

1. Pattern Recognition and Classification- An Introduction, Geoff Dougherty, ISBN: 978-1-

4614-5323-9, Springer, 2013.

2. Advances in Fuzzy Clustering and its Applications, Jose Valente de Olliveira (Editor),
Witold Pedrycz (Editor), ISBN: 978-0-470-02760-8, Wiley 2017.

3. Digital Pattern Recognition, King Sun Fu, ISBN: 978-3-642-96303-2, Springer, 1976

4. Pattern Recognition and Image Analysis Earl Gose, Richard Johnsonbaugh, and Steve
Jost,, PHI Pvt. Ltd., NewDelhi-1, 1999.

5. Statistical Pattern Recognition, 2nd Edition, Andrew R. Webb, ISBN:9780470845134,

John Wiley & Sons, 2002
 CODE COURSE CATEGORY L T P CREDIT
SPEECH SIGNAL PROGRAM
221EEC014 3 0 0 3
PROCESSING ELECTIVE 1

Preamble: This course aims to develop in-depth understanding of fundamentals of speech

analysis, parametric representations and models of speech and speech processing applications
enabling the students to explore into research and development of speech processing systems.

Course Outcomes: After the completion of the course the student will be able to

Understand the basic concepts of speech production and apply time domain
CO 1
analysis methods for classification of speech sounds
Analyse speech segments using frequency domain techniques - STFT and Cepstral
CO 2
analysis
CO 3 Apply LPC Analysis to speech signals
Analyse and apply speech coding techniques for speech compression, storage and
CO 4
transmission
CO 5 Understand the fundamentals of speech processing applications

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 3 3 3 3
CO 2 3 3 3 3
CO 3 3 3 3 3
CO 4 3 3 3 3
CO 5 3 3 3 3

Assessment Pattern

Bloom’s Category End Semester Examination

Apply 40%
Analyse 40%
Evaluate 20%
Create

Mark distribution

Total ESE
CIE ESE
Marks Duration

100 40 60 2.5 hours

Continuous Internal Evaluation Pattern:
Preparing a review article based on peer reviewed Original publications (minimum10
Publications shall be referred):15 marks

Course based task/Seminar/Data Collection and interpretation: 15 marks

Test paper (Shall include minimum of 80% of the syllabus)1 no.: 10 marks

End Semester Examination Pattern:

The end semester examination will be conducted by the respective College. There will be two
parts; Part A and Part B. Part A will contain 5 numerical/short answer questions with 1
question from each module, having 5 marks for each question (such questions shall be useful
in the testing of knowledge, skills, comprehension, application, analysis, synthesis, evaluation
and understanding of the students). Students should answer all questions. Part B will contain
7 questions (such questions shall be useful in the testing of overall achievement and maturity
of the students in a course, through long answer questions relating to theoretical/practical
knowledge, derivations, problem solving and quantitative evaluation), with minimum one
question from each module of which student should answer any five. Each question can carry
7 marks.

Note: The marks obtained for the ESE for an elective course shall not exceed 20% over the
average ESE mark % for the core courses. ESE marks awarded to a student for each elective
course shall be normalized accordingly. For example if the average end semester mark % for
a core course is 40, then the maximum eligible mark % for an elective course is 40+20 = 60
%.
Model Question Paper

No. of Pages: 2 D
APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY
FIRST SEMESTER M.TECH DEGREE EXAMINATION
Branch: Electronics and Communication Engineering
Stream(s): Signal Processing
Course Code & Name: 221EEC014 - SPEECH SIGNAL PROCESSING
Max. Marks: 60 Duration:
2.5 hours
PART A
Answer all questions. Each question carries 5 marks.

1. How do you differentiate voiced and unvoiced speech segments using short time zero
crossing rate and short time energy?

2. Why do we do short time analysis in the case of speech signals? Distinguish between
narrow band and wide band spectrograms.
3. Explore the use of AR models in the analysis of speech signals.
4. How is perceptual irrelevancy removal used in speech compression?
5. Investigate the challenges in speech segmentation?

PART B
Answer any five questions. Each question carries 7 marks.

6. The waveforms and autocorrelation plots of two phonemes segmented from

continuous speech are given in Fig.1(a) and Fig.1(b). Analyze the figures and state

your inferences.
Fig.1(a)
 Fig.1(b)
7. How can we use cepstral analysis to separate source and filter characteristics of a
speech signal? Derive the steps involved in obtaining cepstral coefficients.
8. Formulate the Filter Bank Summation (FBS) method of STFT synthesis. Derive the
FBS constraint.
9. How can we use Levinson Durbin algorithm to find the LPC coefficients?
10. How can we use sub band coding for speech compression?
11. How can we convert any given text to speech ?
12. How can we automatically verify the identity of a given speaker?
13.
Syllabus

MODULE I
Speech Production and Short-Time Speech Analysis:Acoustic theory of speech production,
Excitation, Vocal tract model for speech analysis, Formant structure, Pitch, Articulatory
Phonetics, and Acoustic Phonetics, Time domain analysis (Short time energy, short time zero
crossing Rate, ACF),

MODULE II
Frequency domain analysis: Filter Banks, STFT, Spectrogram, Formant Estimation &
Analysis, Cepstral Analysis, MFCC,

MODULE III
Parametric representation of speech: AR model, ARMA model, LPC model, Autocorrelation
method, Covariance method, Levinson-Durbin Algorithm, Lattice form, Sinusoidal Model,
GMM, Hidden Markov Model,

MODULE IV
Speech coding: Phase Vocoder, LPC, Sub-band coding, Adaptive Transform Coding,
Harmonic Coding, Vector Quantization based Coders, CELP

MODULE V
Applications of speech processing: Fundamentals of Speech recognition, Speech
segmentation, Text-to-speech conversion, speech enhancement, Speaker Verification,
Language Identification
Course Plan

No. of
No Topic
Lectures
1 MODULE I
1.1 Acoustic theory of speech production, Excitation, Vocal tract 2
model for speech analysis
1.2 Formant structure, Pitch, Articulatory Phonetics, and Acoustic 2
Phonetics
1.3 Time domain analysis (Short time energy, short time zero crossing 2
Rate, ACF).
2 MODULE II
2.1 Filter Banks, STFT 3
2.2 Spectrogram, Formant Estimation & Analysis 2
2.3 Cepstral Analysis, MFCC 2
3 MODULE III
3.1 AR model, ARMA model, LPC Analysis - LPC model, 3
Autocorrelation method
3.2 Covariance method, Levinson-Durbin Algorithm, Lattice form 3
3.3 Sinusoidal Model, GMM, Hidden Markov Model 3
4 MODULE IV
4.1 Phase Vocoder, LPC, Sub-band coding 3
4.2 Adaptive Transform Coding, Harmonic Coding 3
4.3 Vector Quantization based Coders, CELP 3
5 MODULE V
5.1 Fundamentals of Speech recognition, Speech segmentation. 3
5.2 Text-to-speech conversion, speech enhancement 3
5.3 Speaker Verification, Language Identification 3

Reference Books

1. Thomas F. Quatieri, Discrete-Time Speech Signal Processing: Principles and Practice,

Prentice Hall; ISBN: 013242942X; 1st edition
2. Rabiner and Juang, Fundamentals of Speech Recognition, Prentice Hall, ISBN:
0130151572.
3. Douglas O'Shaughnessy, Speech Communications: Human & Machine, IEEE Press,
Hardcover 2nd edition, 1999; ISBN: 0780334493.
4. Digital Processing of Speech Signals, 1st edition, ISBN: 97881317051314.
 CODE COURSE NAME CATEGORY L T P CREDIT
ADVANCED EMBEDDED PROGRAM
221EEC015 3 0 0 3
PROCESSORS ELECTIVE 1

Preamble:This course is intended to impart thorough knowledge in embedded processors. It

also helps to develop skills in designing complex systems using different processor
architectures.

Course Prerequisites: Basic knowledge in digital electronics and Microprocessors at UG

level.

Course Outcomes: After the completion of the course the student will be able to

CO 1 Understand the basics of an embedded system

CO 2 Familiarize ARM Architecture
CO 3 Understand and analyse the structure and design of an Embedded System
CO 4 Analyse Product Enclosure, Design and Development
CO 5 Compare standard I/O interfaces.
CO 6 Design an Embedded System

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 3 2 2
CO 2 3 2 2 3
CO 3 3 3 2 2
CO 4 3 3 3 2
CO 5 3 2 3 3
CO 6 3 3 3 3 3

Assessment Pattern

Bloom’s Category End Semester Examination

Apply 20%
Analyse 50%
Evaluate
Create 30%
Mark distribution

Total Marks CIE ESE ESE Duration

100 40 60 2.5 hours

Evaluation shall only be based on application, analysis or design-based questions (for both
internal and end semester examinations).

Continuous Internal Evaluation Pattern:

Continuous Internal Evaluation: 40 marks

Preparing a review article based on peer reviewed original publications

15 marks
(minimum 10 publications shall be referred)

Course based task/Seminar/Data collection and interpretation 15 marks

Test paper, 1 no. (Test paper shall include minimum 80% of the
10 marks
syllabus.)
TOTAL 40 marks

End Semester Examination Pattern:

The end semester examination will be conducted by the respective College. There will be two
parts; Part A and Part B.

Part A will contain 5 numerical/short answer questions with 1 question from each module,
having 5 marks for each question (such questions shall be useful in the testing of knowledge,
skills, comprehension, application, analysis, synthesis, evaluation and understanding of the
students). Students should answer all questions.

Part B will contain 7 questions (such questions shall be useful in the testing of overall
achievement and maturity of the students in a course, through long answer questions relating
to theoretical/practical knowledge, derivations, problem solving and quantitative evaluation),
with minimum one question from each module of which student should answer any five.
Each question can carry 7 marks.

Note: The marks obtained for the ESE for an elective course shall not exceed 20% over the
average ESE mark % for the core courses. ESE marks awarded to a student for each elective
course shall be normalized accordingly. For example, if the average end semester mark % for
a core course is 40, then the maximum eligible mark % for an elective course is 40+20 = 60
%.
Model Question Paper

APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY

FIRST SEMESTER M. TECH DEGREE EXAMINATION

Program: M.Tech.in EC3 (Signal Processing, Signal Processing & Embedded Systems,
Communication Engineering & Signal Processing)

Course Code: 221EEC015

Course Name: ADVANCED EMBEDDED PROCESSORS

Max. Marks: 60 Duration: 150 Minutes

Part A

Answer All Questions. Each Carries 5 mark.

1. Mention two application of embedded system in health care and how can it be used
for monitoring purpose.
2. Discuss about some of the deviation of ARM architecture from pure RISC nature.
3. Explain the linear waterfall model of EDLC.
4. Discuss any two product enclosure development techniques
5. What is cross compilation? What are the files generated on cross compilation?
Part B

Answer any five questions: Each question carries 7 marks.

6. With a neat diagram explain the parts or elements of an embedded system.

7. Explain the 5-stage pipeline ARM organization with the help of a neat diagram.
8. Discuss the different phases of EDLC with the help of classic Embedded Product
Development Life Cycle Model
9. Discuss and describe the communication interfacing technique used in USB and
IEEE 1394.
10. Explain the external communication interfaces in detail.
11. Describe the boundary scan technique for testing the interconnection among the
various chips in a complex hardware board.
12. Explain the SPI bus architecture with the help of a neat bus architecture diagram
 Syllabus

Module 1: 8 hours

Introduction to Embedded systems: Embedded system examples, Parts of Embedded

System, Typical Processor architecture, Power supply, clock, memory interface, interrupt, I/O
ports, Buffers, Programmable Devices, ASIC, etc. Simple interfacing examples, Memory
Technologies, EPROM, Flash, OTP, SRAM, DRAM, SDRAM etc

Module 2: 11 hours

ARM architecture: ARM organization and Implementation, Memory Hierarchy, ARM

Instruction Set and Thumb Instruction set.High- Level Language Programming, System
Development using ARM. Digital Signal Processing on ARM, Peripheral Programming and
system design for a specific ARM processor (ARM7/9).

Module 3: 7 hours

Embedded System product Development: Embedded System product Development Life

cycle (EDLC). Specifications, Component selection, Schematic Design, PCB layout,
fabrication and assembly.

Module 4: 6 hours

Product enclosure design and development: Concept of firmware, operating system and
application programs. Power supply Design, External Interfaces.

Module 5: 8 hours

Embedded System Development Environment: IDE, Cross compilation,

Simulators/Emulators. Hardware Debugging. Hardware testing methods like Boundary Scan,
In Circuit Testing (ICT) etc., Serial EEPROM, PWM, Analog to digital converter Bus
architecture like I2 C, SPI, AMBA, CAN etc.
COURSE PLAN

No. of
No Topic
Lectures
1 Introduction to Embedded systems
Embedded system examples, Parts of Embedded System, Typical
1.1 Processor architecture, Power supply, clock, memory interface, 4
interrupt, I/O ports, Buffers, Programmable Devices, ASIC, etc
Simple interfacing examples, Memory Technologies, EPROM, Flash,
1.2 4
OTP, SRAM, DRAM, SDRAM etc

2 ARM architecture

ARM organization and Implementation, Memory Hierarchy, ARM

2.1 4
Instruction Set and Thumb Instruction set

High- Level Language Programming, System Development using

2.2 4
ARM
Digital Signal Processing on ARM, Peripheral Programming and
2.3 3
system design for a specific ARM processor (ARM7/9).
3 Embedded System product Development

3.1 Embedded System product Development Life cycle (EDLC) 3

Specifications, Component selection, Schematic Design, PCB layout,

3.2 4
fabrication and assembly.

4 Product enclosure design and development

4.1 Concept of firmware, operating system and application programs 3

4.2 Power supply Design, External Interfaces. 3

5 Embedded System Development Environment
5.1 IDE, Cross compilation, Simulators/Emulators 3

Hardware Debugging. Hardware testing methods like Boundary Scan,

5.2 2
In Circuit Testing (ICT) etc.,
Serial EEPROM, PWM, Analog to digital converter, Bus architecture
5.3 3
like I2C, SPI, UART, AMBA, CAN etc.
Text Books

1. Shibu K.V. Introduction to Embedded Systems, Tata McGraw Hill, 2009.

2. Steve Furber, ARM System-on-chip Architecture, Second Edition Pearson
Education, 2007.
Reference Books

1. Van Ess, Currie and Doboli, Laboratory Manual for Introduction to Mixed-Signal,
Embedded Design, Alphagraphics, USA
2. William Hohl, ARM Assembly Language Programming, CRC Press, 2009.
3. Andrew Sloss, Dominic Symes, Christ Wright, ARM System Developer’s guide –
Designing and optimizing software, Elsevier Publishers, 2008.
 CODE COURSE NAME CATEGORY L T P CREDIT
INFORMATION HIDING AND PROGRAM
221EEC016 3 0 0 3
DATA ENCRYPTION ELECTIVE 1

Preamble: The course is designed to provide an insight to various data encryption and
information hiding techniques and applying these techniques in various security applications.
The course also aims to develop skills in analysing the strengths and weakness of various
techniques used for information security.

Course Outcomes: After the completion of the course the student will be able to

CO 1 Apply various techniques for data encryption and analyse the performance (K3).
CO 2 Identify and apply information hiding and digital watermarking techniques for given
problems in security systems (K3).
CO 3 Analyse publications related to encryption and information hiding in journals and
conferences and submit report (K4).
CO 4 Choose and solve a research problem in the area of data encryption /and information
hiding (K5).

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 3 2
CO 2 3 3 2
CO 3 3 3 3 2
CO 4 3 3 3 3 3 2 2

Assessment Pattern

Continuous
End Semester
Bloom’s Category Internal
Examination
Evaluation
Understand 10
Apply 40 10
Analyse 10 15
Evaluate 15
Create

Mark distribution

Total Marks CIE ESE ESE Duration

100 40 60 2.5 hours

Continuous Internal Evaluation Pattern:

Preparing a review article based on peer reviewed

original publications (minimum 10 Publications 15 marks
shall be referred):
Course based task/Seminar/Data Collection and
15 marks
interpretation:
Test paper, 1 no.:
(Test paper shall include minimum 80% of the 10 marks
syllabus.)

End Semester Examination Pattern:

The end semester examination will be conducted by the respective College. There will be two
parts; Part A and Part B. Part A will contain 5 numerical/short answer questions with 1
question from each module, having 5 marks for each question (such questions shall be useful
in the testing of knowledge, skills, comprehension, application, analysis, synthesis, evaluation
and understanding of the students). Students should answer all questions. Part B will contain
7 questions (such questions shall be useful in the testing of overall achievement and maturity
of the students in a course, through long answer questions relating to theoretical/practical
knowledge, derivations, problem solving and quantitative evaluation), with minimum one
question from each module of which student should answer any five. Each question can carry
7 marks.

CO level Assessment Questions

CO1: Apply various techniques for data encryption and analyse the performance
1. Distinguish between a synchronous and a nonsynchronous stream cipher.
2. In the DSS scheme, if Eve can find the value of r (random secret), can she forge a
message? Explain.
3. Prove that (x3 + x2 + 1) is an irreducible polynomial of degree 3.

CO2: Identify and apply information hiding and digital watermarking techniques for
given problems in secure communication systems
1. Choose a digital watermarking technique for protecting an image and justify
2. Derive a method to hide an image in a video with optimal bandwidth.
3. Describe the hiding techniques in spatial and temporal transform domains and their
applications.

CO3: Analyse publications related to encryption and information hiding in journals and
conferences and submit report.
1. Review any 10 conference papers on digital watermarking techniques, analyse and
prepare a report.
2. Analyse at least 10 journal papers on video steganography and present.
CO4: Choose and solve a research problem in the area of data encryption /and
information hiding.
1. Encrypt any text data using symmetric and asymmetric methods of encryption and
compare the performance.
2. Protect an audio file which has to be protected and transmitted through a vulnerable
and noisy channel using image steganography.
3. Hide an image using any one of the conventional steganographic techniques and also
using network steganography (Protocol) and compare your results.
Model Question Paper
Course Code: 221EEC016
Course Name: INFORMATION HIDING AND DATA ENCRYPTION
Max. Marks: 60 Duration: 2.5 Hours
PART A
Answer all Questions. Each question carries 5 marks

Sl. Question CO
No.
1. Find the remainder when 1653 is divided by 7 1
2. Describe the parameters used for measuring the capability of hiding 1
techniques
3. Identify the techniques used for tamper detection in image and audio and 1
explain how the detection is carried out.
4. Suggest a mathematical model for protecting the bio-medical signals and 2
justify.
5. Explain detection theoretic approach for steganalysis. 2

PART B
Answer any one full question from each module. Each question carries 7 marks
6. Derive a general formula to calculate the number of each kind of
transformation (SubBytes, ShiftRows, MixColumns, and
AddRoundKey) and the number of total transformations for AES192 1
and AES256. The formula should be parametrized on the number of
rounds.
7. Write down algorithm for embedding and retrieval of text data using
2
spread spectrum technique.
8. Derive a mathematical model for hiding an image and explain. 2
9. Differentiate between adaptive and non-adaptive techniques for
information hiding. Which are the adaptive techniques used for hiding 2
audio?
10. Write down the encoding and decoding process for image hiding in the
2
DCT domain and explain.
11. Discuss temporal and transform domain techniques used in video using
2
relevant examples.
12. Illustrate SVM method for steg-analysis 2
 Syllabus

Module1:

Review of Number Theory: Elementary Number theory, Algebraic Structures- Groups,

Rings and Finite Fields, Polynomials over Finite Fields (Fq), Introduction to Complexity
theory.

Data Encryption Methods: Introduction to Cryptography, Classical Cryptography, Stream

Ciphers, Public Key Cryptography based on Knapsack problem, AES. Digital Signature, Zero
Knowledge Proofs.

Module2:
Introduction to Information Hiding: Steganography. Objectives, difference, requirements,
Types – Fragile, Robust. Parameters and metrics - BER, PSNR, WPSNR, Correlation
coefficient, MSE, and Bit per pixel.

Information Hiding Approaches: LSB, additive and spread spectrum methods.

Module3:
Digital Watermarking: Algorithms, Types of Digital Watermarks, Applications, Audio
Watermarking

Applications of Information Hiding: Authentication, annotation, tamper detection and

Digital rights management. Hiding text and image data, mathematical formulations.

Module4:
Information Hiding in 1D signals: Time and transform techniques, hiding in Audio,
biomedical signals, HAS Adaptive techniques.

Information Hiding in 2D signals: Spatial and transform techniques-hiding in images, ROI

images.

Module5:
Information Hiding in video: Temporal and transform domain techniques, Bandwidth
requirements, HVS Adaptive techniques.

Steg analysis: Statistical Methods, HVS based methods, SVM method, Detection theoretic
approach.
Course Plan

Topic No. of
No
Lectures
1 Review of Number Theory and Data Encryption Methods:
1.1 Elementary Number theory 1
Algebraic Structures- Groups, Rings and Finite Fields,
1.2 2
Polynomials over Finite Fields (Fq)
1.3 Introduction to Complexity theory. 1
Introduction to Cryptography, Classical Cryptography, Stream
1.4 2
Ciphers
1.5 Public Key Cryptography based on Knapsack problem 2
1.6 AES. Digital Signature, Zero Knowledge Proofs. 2
2 Information Hiding:
2.1 Steganography. Objectives, difference, requirements 2
Types – Fragile, Robust. Parameters and metrics - BER, PSNR,
2.2 3
WPSNR, Correlation coefficient, MSE, and Bit per pixel.
Information Hiding Approaches: LSB, additive and spread
2.3 3
spectrum methods.
3 Digital Watermarking and applications of Image Hiding:
3.1 Digital Watermarking Algorithms 2
3.2 Types of Digital Watermarks and Applications 1
3.3 Audio Watermarking 1
Applications of Information Hiding: Authentication, annotation,
3.4 2
tamper detection and Digital rights management
3.5 Hiding text and image data, mathematical formulations. 2
4 Information Hiding in 1D and 2D signals:
Information Hiding in 1D signals: Time and transform techniques,
4.1 3
hiding in Audio, biomedical signals, HAS Adaptive techniques.
Information Hiding in 2D signals: Spatial and transform
4.2 techniques-hiding in images, ROI images, HVS Adaptive 4
techniques.
5 Information Hiding in video and Steg analysis:
Information Hiding in video: Temporal and transform domain
5.1 3
techniques, Bandwidth requirements.
Steg analysis: Statistical Methods, HVS based methods, SVM
5.2 4
method, Detection theoretic approach.

Reference Books

1. Neal Koblitz, A Course in Number Theory and Cryptography, 2nd Edition, Springer
2. Stefan Katzenbeisser, Fabien A. P. Petitcolas, Information Hiding Techniques for
Steganography and Digital Watermarking, Artech House Publishers, 2000.
3. Neil F Johnson et al Kluwer, Information hiding: Steganography and Watermarking -
Attacks and Countermeasures,Springer, 2001.
4. Ingemar J Cox, Digital Watermarking, The Morgan Kaufman Series in Multimedia
Information and Systems, 2001
5. Ira S Moskowits, Information Hiding, Proceedings, 4th International Workshop, IH
2001, Pittsburg, USA, April 2001, Eds:2. AVISPA package homepage,
http:/www.avispaproject.org/
6. Handbook of Applied Cryptography, AJ Menezes, CRC Press, 2001.
 CODE COURSE NAME CATEGORY L T P CREDIT
PROGRAMMING TOOLS
PROGRAM
221EEC017 FOR MODELING AND 3 0 0 3
ELECTIVE 1
SIMULATION

Preamble:This course is about learning the programming languages used in the development
of embedded systems. Learners can use the concepts learned in this course for the
development of processor based systems.

Course Prerequisites: Basic knowledge in programming, and embedded systems.

Knowledge on ARM processors or any other processors and their architecture is a
requirement.

Course OutcomesAfter the completion of the course the student will be able to

CO 1 To understand the Linux system and command level programming

CO 2 To understand the c programming basics
CO 3 To apply the c programming language in ARM programming
CO 4 To understand the basics of python language and its constructs
CO 5 To apply the python language in embedded applications

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 3 3 3 3
CO 2 3 3 3 3
CO 3 3 3 3 3
CO 4 3 3 3 3
CO 5 3 3 3 3

Assessment Pattern

End Semester Examination

Bloom’s Category
(Marks)
Apply 20
Analyse 20
Evaluate
Create 20
Mark distribution

Total Marks CIE ESE ESE Duration

100 40 60 2.5 hours

Evaluation shall only be based on application, analysis or design based questions(for both
internal and end semester examinations).

Continuous Internal Evaluation Pattern:

Continuous Internal Evaluation: 40 marks

Preparing a review article based on peer reviewed original publications (minimum 10
publications shall be referred) : 15 marks
Course based task/Seminar/Data collection and interpretation: 15 marks
Test paper, 1 no. : 10 marks
Test paper shall include minimum 80% of the syllabus.

All course based assignments/tasks shall be of programming examples/ programming

implementations of embedded systems. Review article can be prepared on embedded
application implementations, embedded processors,

End Semester Examination Pattern:

The end semester examination will be conducted by the respective College. Therewill be two
parts; Part A and Part B.
Part A will contain 5 numerical/short answerquestions with 1 question from each module;
having 5 marks for each question (suchquestions shall be useful in the testing of knowledge,
skills, comprehension,application, analysis, synthesis, evaluation and understanding of the
students).Students should answer all questions.
Part B will contain 7 questions (such questionsshall be useful in the testing of overall
achievement and maturity of the students in acourse, through long answer questions relating
to theoretical/practical knowledge,derivations, problem solving and quantitative evaluation),
with minimum one questionfrom each module of which student should answer any five. Each
question can carry7 marks.
Note: The marks obtained for the ESE for an elective course shall not exceed20% over the
average ESE mark % for the core courses. ESE marks awarded to astudent for each elective
course shall be normalized accordingly. For example if theaverage end semester mark % for a
core course is 40, then the maximum eligiblemark % for an elective course is 40+20 = 60 %.
Model Question Paper

APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY

FIRST SEMESTER M. TECH DEGREE EXAMINATION, (Model Question Paper)

Program: M.Tech. in EC3 (Signal Processing, Signal Processing & Embedded Systems,
Communication Engineering & Signal Processing)

Course Code: 22EEC017

Course Name: PROGRAMMING TOOLS FOR MODELING AND SIMULATION

Max. Marks: 60 Duration: 150 Minutes

PART A

Answer ALL Questions. Each Carries 5 mark.

1. Write a C program to read an English Alphabet through keyboard and display

whether the given Alphabet is in upper case or lower case.
2. With a suitable example, explain how pointers can help in changing the content of a
single dimensionally array passed as an argument to a function in C.
3. Write a python program to print odd numbers from 1 to N.
4. How will you plot magnitude spectrum in Matplotlin in Python?
5. Explain virtual desktop. How do you share a program across different virtual
desktops under Linux?
PART – B

Answer any 5 questions; each question carries 7 marks.

6. Write a Linux X window program to create a window with a little black square in it
and exits on a key press.
7. Explain the features of the Linux system? Why Linux is regarded as a more secure
operating system than other operating systems?
8. Write a C program to find the transpose of a matrix.
9. Write an embedded C ARM I/O program to display a message on the LCD using 8-
bit mode and delay.
10. Write Python code to multiply two matrices using nested loops and also perform
transpose of the resultant matrix.
11. Write a note on the image processing function in Python.
12. Write a NumPy program to create an element-wise comparison (equal, equal within
a tolerance) of two given arrays
 Syllabus

Module I (8 Hrs.)

Linux: Introduction, The shell, Shell script and programming, Shell configuration, Linux
files, directories and archives. The X window system, Xorg, and Display managers, Gnome,
KDE, Linux software management.

Module II (8 Hrs.)

Embedded Programming: C programming, Constants, variables and data types, operators and
Expressions, I/O operations, Control flow statements (if else, switch, loops), Arrays and
strings, Functions, structures and unions, Pointers, file management, Dynamic memory
allocation and Linked lists.

Module III (8 Hrs.)

Embedded C Programming using embedded IDE. ARM I/O programming, LED, LCD,
Keypad interfacing, UART,SPI,I2C programming, Timer programming, Interrupt and
Exception programming, ADC, DAC, Sensor interfacing.

Module IV (8 Hrs.)

Python Programming basics: variables, Input, Output, Basic operations, String manipulation,
Loops, functions, Lists, Dictionary.

Module V (8 Hrs.)

Python programming- Advanced python, Formatting, class, files, exceptions.

Course Plan

No. of
No Topic
Lectures
Linux: Introduction, The shell, Shell script and programming, Shell configuration,
1 Linux files, directories and archives. The X window system, Xorg, and Display
managers, Gnome, KDE, Linux software management.

Linux : Introduction, distributions, accessing linux system,

1.1 deskop and command line interface 3

The shell, Shell script and programming, Shell configuration,

1.2 Linux files, directories and archives. 3

The X window system, Xorg, and Display managers, Gnome,

1.3 KDE, Linux software management. 2

Embedded Programming: C programming, Constants, variables and data types,

2
operators and Expressions, I/O operations, Control flow statements (if else, switch,
 loops), Arrays and strings, Functions, structures and unions, Pointers, file
management, Dynamic memory allocation and Linked lists.
Embedded Programming: C programming, Constants, variables
2.1 3
and data types, operators and Expressions, I/O operations,
Control flow statements (if else, switch, loops), Arrays and strings,
2.2 3
Functions, structures and unions,
Pointers, file management, Dynamic memory allocation and
2.3 2
Linked lists.
Embedded C Programming using embedded IDE. ARM I/O programming, LED,
3 LCD, Keypad interfacing, UART,SPI,I2C programming, Timer programming,
Interrupt and Exception programming, ADC, DAC, Sensor interfacing
Embedded C Programming using embedded IDE. ARM I/O
3.1 3
programming, LED, LCD, Keypad interfacing
3.2 UART,SPI,I2C programming, Timer programming, 3
Interrupt and Exception programming, ADC, DAC, Sensor
3.3 2
interfacing
Python Programming basics : variables, Input, Output, Basic operations, String
4
manipulation, Loops, functions, Lists, Dictionary
Python Programming basics : variables, Input, Output, Basic
4.1 3
operations,
4.2 String manipulation, Conditional instructions, Loops, functions, 3
4.3 2
Python programming- Advanced python, Formatting, class, files, exceptions, linux
5 commands, web server, Signal plotting and processing (numpy and matplotlib),
Graphics, Computer vision. Programming examples of systems.
Python programming- Advanced python, Formatting, class, files,
5.1 3
exceptions
linux commands, web server, Signal plotting and processing
5.2 3
(numpy and matplotlib),
5.3 Graphics, Computer vision. Programming examples of systems. 2

Text Books

1. E. Balagurusamy, Programming in ANSI C, Tata McGraw-Hill, 6th Edition 2012

2. Muhammad Ali Mazidi, Shujen Chen, Sarmad Naimi, SepehrNaimi, “Freescale
ARM Cortex-M Embedded Programming using C language”, Mazidi and
Naimi,2014
3. Simon Monk, “Raspberry Cook Book Software and Hardware Problems and
Solutions”, 2nd Edition, O’ Reilly Media Inc., 2016.
4. Richard Petersen, "Linux: The Complete Reference", 2017, Sixth Edition, McGraw
Hill Education
Reference Books

1. David Russell, “Introduction to Embedded systems Using ANSI C and the Arduino
development Environment”, 2010, 1rd edition, Morgan & Claypool Publishers.
2. E.I. Horvath, E.A. Horvath, “Learning IoT with Python and Raspberry Pi”, Learning
IoT LLC, 2019.
3. Sarmad Naimi, Muhammad Ali Mazidi, SepehrNaimi, “The STM32F103 Arm
Microcontroller and Embedded Systems: Using Assembly and C”, MicroDigitalEd,
2020
 SEMESTER I
PROGRAM ELECTIVE II
 CODE COURSE NAME CATEGORY L T P CREDIT

DSP PROCESSORS AND PROGRAM

221EEC018 3 0 0 3
ARCHITECTURE ELECTIVE 2

Preamble: The aim of the course is to give an overview of the commonly used
DSPalgorithms, their applications and various techniques for the algorithmic and
architecturelevel optimisations through various algorithm to architecture mapping which can
lead to efficient hardware implementations. The course also introduces the basic features in
DigitalSignal Processors,Micro controllers with DSP extensions, DSP Architecture with case
studies, the latest architectural trends in DSPsand their programming tools.

Course Outcomes: After the completion of the course the student will be able to

Analyse the basic resource constraints in a practical DSP system and solve them
CO 1 using various techniques/transformations that map the DSP algorithms to efficient
architectures.
Apply the knowledge of various single core and multicore Digital Signal Processor
CO 2 architectures in identifying the optimal processor for solving real life signal
processing problems.
Evaluate the DSP algorithms implemented in dedicated DSP processors and the
CO 3
micro controllers with DSP extensions
Create algorithms to solve signal processing problems using the latest hardware
CO 4
platforms and software tools.

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 1 3 3 3 1
CO 2 1 3 3 3 1
CO 3 1 3 3 3 1
CO 4 2 3 3 3 2

Assessment Pattern

Bloom’s Category End Semester Examination

Apply 20
Analyse 15
Evaluate 15
Create 10
Mark distribution

Total ESE
CIE ESE
Marks Duration

100 40 60 2.5 hours

Continuous Internal Evaluation Pattern:

Continuous Internal Evaluation: 40 marks

Preparing a review article based on peer reviewed Original publications (minimum 10)
Publications shall be referred): 15 marks

Course based task/Seminar/Data Collection and interpretation: 15 marks

Test paper, 1 no.: 10 marks

Test paper shall include minimum 80% of the syllabus.

End Semester Examination Pattern:

The end semester examination question paper consists of two parts; Part A and Part B. Part
A will contain 5 numerical/short answer questions with 1 question from each module, having
5 marks for each question (such questions shall be useful in the testing of knowledge, skills,
comprehension, application, analysis, synthesis, evaluation and understanding of the
students). Students should answer all questions. Part B will contain 7 questions (such
questions shall be useful in the testing of overall achievement and maturity of the students in
a course, through long answer questions relating to theoretical/practical knowledge,
derivations, problem solving and quantitative evaluation), with minimum one question from
each module of which student should answer any five. Each question can carry 7
marks.

Syllabus

Module 1: Basics of DSP Algorithm Representation to Architecture Mapping

DSP Algorithm representations –Block Diagram, Signal Flow Graph, Data Flow Graph,
Dependence Graph;
Introduction to Filter structures- Recursive, Non-recursive and Lattice structures;
Fundamentals of DSP algorithm to architecture mapping - Loop bound, Iteration Bound,
Critical Path;
Algorithms for computing Iteration Bound – Longest Path Matrix Algorithm, Minimum
Cycle Mean Algorithm.
Module 2: Transformations for Improved DSP Architectures
VLSI performance measures - area, power, and speed;
Pipelining - Pipelining of FIR filters, Concept of Fine Grain Pipelining;
Parallel Processing – Designing Parallel FIR systems;
Pipelining and Parallel Processing for low power;
Folding and Unfolding Transformations and its applications.

Module 3: Single Core DSP Architectures

Introduction to General Purpose Processors (GPP), Microcontroller Units (MCU), Digital
Signal Processors (DSP) - comparison and Applications;
The key features of a Digital Signal Processors – Dedicated hardware units, Circular Buffers,
Modified bus structures and Memory access schemes;
Introduction to Harvard, Super Harvard DSP architectures, Concept of Instruction Level
Parallelism, VLIW Architecture and Single Instruction Multiple Data (SIMD) processor
architecture;
Case Study 1: Introduction to a popular DSP from Texas Instruments, The TMS320C67xx
Series Processor- CPU Architecture - CPU Data Paths and Control - Timers – Multichannel
Buffered Serial Ports (McBSPs)- Internal Data/ Program Memory - External Memory
Interface.

Case Study 2: Introduction to ARM Cortex-M Based Microcontrollers with DSP extensions -
ARMv7E-M architecture

Module 4: Homogeneous Multicore DSPs

Introduction to multicore processors and their applications, A brief comparison between DSP
SoCs, Field-Programmable Gate Arrays (FPGAs), Graphic Processors and CPUs;
Introduction to Multicore DSP Architectures: The TMS320C66x
architecture: The CPU, Overview of the peripherals, Overview of the memory organization.

Module 5: Programming the DSPs

Introduction to Code Composer Studio (CCS) software development tool and the
TMS320C6XX EVM kit. Introduction to Keil Development tool, CMSIS DSP software
library and ARM Cortex-M4 development board;
Introduction to Open MP Application Programming Interface (API) and Open Computing
Language (OpenCL);
Implementation of simple DSP algorithms;
Latest architectural trends in digital signal processing: Introduction to Heterogeneous
Multicore DSP Architecture and FPGA SoCs.
Course Plan

No. of
No Topic
Lectures
1 Basics of DSP Algorithm Representation to Architecture Mapping
DSP Algorithm representations –Block Diagram, Signal Flow
1.1 2
Graph,Data Flow Graph, Dependence Graph.
Introduction to Filter structures- Recursive, Non-recursive and Lattice
1.2 1
structures.
Fundamentals of DSP algorithm to architecture mapping - Loop
1.3 2
bound,Iteration Bound, Critical Path.
Algorithms for computing Iteration Bound – Longest Path Matrix
1.4 2
Algorithm, Minimum Cycle Mean Algorithm.

2 Transformations for Improved DSP Architectures

2.1 VLSI performance measures - area, power, and speed 1
Pipelining - Pipelining of FIR filters, Concept of Fine Grain
2.2 2
Pipelining.
2.3 Parallel Processing – Designing Parallel FIR systems. 2
2.4 Pipelining and Parallel Processing for low power. 1
2.5 Folding and Unfolding Transformations and its applications. 2

3 Single Core DSP Architectures

Introduction to General Purpose Processors (GPP), Microcontroller
3.1 Units (MCU), Digital Signal Processors (DSP) - comparison and 1
Applications.
The key features of a Digital Signal Processors – Dedicated hardware
3.2 units, Circular Buffers, Modified bus structures and Memory access 1
schemes.
Introduction to Harvard, Super Harvard DSP architectures, Concept
3.3 of Instruction Level Parallelism, VLIW Architecture and Single 1
Instruction Multiple Data (SIMD) processor architecture.
Case Study 1: Introduction to a popular DSP from Texas Instruments, 3
The TMS320C67xx Series Processor- CPU Architecture - CPU Data
Paths and Control - Timers – Multichannel Buffered Serial Ports
3.4
(McBSPs)- Internal
Data/ Program Memory - External Memory Interface.

Case Study 2: Introduction to ARM Cortex-M Based Microcontrollers

3.5 3
with DSP extensions - ARMv7E-M architecture

4 Homogeneous Multicore DSPs

Introduction to multicore processors and their applications, A brief
4.1 comparison between DSP SoCs, Field-Programmable Gate Arrays 1
(FPGAs), Graphic Processors and CPUs.
 Introduction to Multicore DSP Architectures: The TMS320C66x
4.2 architecture: The CPU, Overview of the peripherals, Overview of the 4
memory organization.

5 Programming the DSPs

Introduction to Code Composer Studio (CCS) software development
tool and the TMS320C6XX EVM kit
5.1 3
Introduction to Keil Development tool, CMSIS DSP software library
and ARM Cortex-M4 development board
Introduction to Open MP Application Programming Interface (API)
5.2 2
and Open Computing Language (OpenCL).
5.3 Implementation of simple DSP algorithms 2
Latest architectural trends in digital signal processing: Introduction to
5.4 1
Heterogeneous Multicore DSP Architecture and FPGA SoCs.

Text Books

1. Keshab K. Parhi, "VLSI Signal Processing Systems, Design and Implementation",

John Wiley &Sons,1999
2. NaimDahnoun, “Multicore DSP: from algorithms to real-time implementation on the
TMS320C66x SoC”. John Wiley & Sons, 2018.
3. Steven W. Smith, “The Scientist and Engineer’s Guide to Digital Signal Processing”
Second Edition, California Technical Publishing , 1999.
4. Reference Link for Overview of Latest Processor Architectures– Digital signal
processors (DSPs) | Overview | Processors | TI.com,
https://training.ti.com/system/files/docs/c66x-corepac-instruction-set-
referenceguide.pdf
5. Joseph Yiu “The Definitive Guide to ARM® Cortex®-M3 and Cortex®-M4
Processors”, Elsevier, 2014

Reference Books

1.RulphChassaing, “Digital Signal Processing and Applications with the C6713 and
C6416 DSK”, John Wiley & Sons, 2005.
2.Sen M. Kuo, Woon-Seng S. Gan, Digal Signal Processors: Architectures,
Implementations, and Applications, Prentice Hall, 2004.
3.Lars Wanhammar, DSP Integrated Circuits, Academic Press, 1999.
4.B Venkataramani, M Bhaskar, “Digital Signal Processors: Architecture, Programming
and Applications”, 2nd Ed., Tata McGraw-Hill Education, 2002.
5. A. Kharin, S. Vityazev and V. Vityazev, "Teaching multi-core DSP implementation on
EVM C6678 board," 2017 25th European Signal Processing Conference (EUSIPCO),
2017, pp. 2359-2363, doi: 10.23919/EUSIPCO.2017.8081632
6.Donald S. Reay.. “Digital Signal Processing Using the ARM Cortex M4”, (1st. ed.).
Wiley Publishing, 2015
7.CemÜnsalan, M. ErkinYücel, H. DenizGürhan, “Digital Signal Processing Using Arm
Cortex-M Based Microcontrollers: Theory and Practice”, ARM Education Media, 2018.
Model Question Paper
APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY
FIRST SEMESTER M. TECH DEGREE EXAMINATION, (Model Question Paper)
Course Code: 221EEC018
Course Name: DSP PROCESSORS AND ARCHITECTURE
Max. Marks: 60 Duration: 2.5 Hours
PART A
Answer all Questions. Each Carries 5 mark.
1 Differentiate between Signal Flow Graph (SFG) and Data Flow Graph (DFG) with
example.
2 What is pipelining? Explain with an example, how it helps in reducing the critical
path delay in implementing the DSP systems.
3 In what way the Super Harvard architecture-based DSPs differs from the normal
microprocessors?
4 What is the concept of Heterogeneous Multicore DSP Architecture? Quote an
example processor?
5 Quoting a suitable example, explain the architectural advantages of an FPGA SoC.
PART – B
(Answer any five questions, each carries 7 mark.)

6 Explain the Longest Path Matrix (LPM) Algorithm for computing the iteration bound
of a DFG.
For the DFG shown in figure below, the computation times of the nodes are shown in
parentheses. Compute the iteration bound of this DFG using the LPM algorithm.

7 For the following transfer function given, Derive the basic lattice filter and draw its
structure
8 Consider a direct-form implementation of the FIR filter
y(n) = ax(n) + bx(n-2) +cx(n-3)
Assume that the time required for 1 multiply-add operation is T
i. Pipeline this filter such that the clock period is approximately T
ii. Draw block filter architecture for a block size of three. Pipeline this
block filter such that clock period is about T. What is the system sample
9 The TMS320C6713 processor is used for an application where, it has to read the
audio data inputted through the codec and has to send the data which is band limited
to 1 KHz, to another external device for further processing. If the processor is
connected to the audio codec through the McBSPs of the TMS320C6713 processor.

a)Draw the interconnection diagram showing all the necessary signals, for inputting
an analog signal to the processor for the processing and to send the result there after,
with the entire data transfer initiated through the McBSPs.

b)What are the various registers need to be programmed in order to affect the data
transfer? Explain the role and functionality of each.
10 Draw a neat block schematic of the architecture of TMS320C66x series of
processor. Briefly explain the role of each block.
11 Give an overview of the memory organisation in TMS320C66xx series of
processors. Explain the role of various memory controllers and interfaces in
relieving the CPU load.
12 Give an overview of the latest architectural trends for implementing DSP
algorithms. How will you compare FPGA SoCs and DSP SoCs?.
 CODE COURSE NAME CATEGORY L T P CREDIT
PROGRAM
221EEC019 CODING THEORY 3 0 0 3
ELECTIVE 1

Preamble: This course aims at a rigorous analysis of various error correction codes starting
from the earliest Hamming code to the latest polar codes used in 5G

Course Outcomes: After the completion of the course the student will be able to

CO 1 Analyse block codes and various bounds governing their construction

CO 2 Analyse LDPC coding and hard decision/soft decision decoding
CO 3 Illustrate coding and decoding of BCH/RS codes
CO 4 Review convolutional encoding and decoding them using BCJR algorithm
CO 5 Illustrate coding and decoding of Turbo codes
CO 6 Discuss polar codes and their applications in 5G

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 3 3 2
CO 2 3 2 2
CO 3 3 3 3
CO 4 3 2 3
CO 5 3 3 2
CO 6 3 3 3
Assessment Pattern

Bloom’s Category End Semester Examination

Apply 10
Analyse 25
Evaluate 20
Create 5
Mark distribution

Total Marks CIE ESE ESE Duration

100 40 60 2.5 hours

Continuous Internal Evaluation Pattern:

Continuous Internal Evaluation: 40 marks

● Preparing a review article based on peer reviewed Original publications
(minimum 10 publications shall be referred) : 15 marks
● Course based task/Seminar/Data collection and interpretation:15 marks
● Test paper, 1 no. :10 marks
Test paper shall include minimum 80% of the syllabus.

End Semester Examination Pattern:

There will be two parts; Part A and Part B

● Part A will contain 5 numerical/short answer questions with 1 question from each module,
having 5 marks for each question. Students should answer all questions.
● Part B will contain 7 questions with minimum one question from each module of which
student should answer any five. Each question can carry 7 marks.

Model Question Paper

A P J ABDUL KALAM TECHNOLOGICAL UNIVERSITY

M.TECH DEGREE EXAMINATION
FIRST SEMESTER
CODING THEORY
Time: 2.5 Hours Max. Marks: 60

PART A
Answer All Questions 5 x 5 marks = 25 marks
1. Explain Hamming bound and Gilbert-Varshamov bound
2. Analyse the effect of short-cycles in the decoding of LDPC codes
3. Show that the dual of an RS code is also an RS code (and hence MDS)
4. Construct the Generator matrix of a RM(1,3) code
5. Discuss the encoding of Turbo codes
PART B
Answer any 5 questions 5 x 7 marks = 35marks

6.The parity check bits of a (8,4) block code are generated by, p0 =m0 +m1 +m3 , p1 =m0 +m1
+m2 ,p2 =m0 +m2 +m3, p3 =m1 +m2 +m3 where m0 , m1, m2 and m3 are the message digits.
a) Find the generator matrix and the parity check matrix for this code in the form
G=[P:Ik ]
b) Find the minimum weight of this code.
c) Find the error-detecting capabilities of this code.
d) Check whether 11101000 and 11100000 are valid codewords using H matrix.

7. a)Explain hard decision decoding of LDPC codes using Bit-flipping algorithm in BSC.
b) .Explain BCJR decoding of convolutional codes

8. For a binary, narrow sense, triple error correcting BCH code of length 15, constructed
using the polynomial x4+x+1
(a) Compute a generator polynomial for this code
(b) Determine the rate of the code
(c) Construct the parity check matrix and generator matrix for this code
9. Form the generator matrix of the first order RM code RM (1,3) of length 8. What is the
minimum distance of the code? Determine its parity check sums and devise a majority logic
decoder for the code. Decode the received vector r = (01000101)
10. Describe the basic ideas of polarization. Analyse mathematically channel polarisation for
N=2 channel

11.Analyze the iterative soft decoding of Turbo codes

12. Differentiate between the BCH Viewpoint and Vandermonde viewpoints of Reed
Solomon Codes
Syllabus

Module 1 Linear Block Codes and Bounds: Course Overview-Relevance of Error

correction schemes in communication systems, Repetition coding, concepts of Code rate,
Hamming Distance, Minimum Distance, Error detecting and correcting capability, Review of
Group, ring, fields and vectorspaces, Hamming bound, Singleton bound, Plotkin bound,
Gilbert-Varshamov bound

Module 2 LDPC Codes Regular and Irregular Low Density Parity Check Codes-Tanner
graph, Message Passing decoding-Hard decision and Soft decision, Bit flipping algorithm for
decoding, Bit flipping algorithm for decoding, Belief Propagation decoding: Sum Product
algorithm

Module 3 BCH and RS codes Galois Fields -- Irreducible and Primitive Polynomials- BCH
Codes - Design, BCH Bound, Decoding BCH codes – Decoding BCH – the general outline,
computation of the syndrome, error locator polynomial, Chien Search algorithm, Finding the
error locator polynomial. Berlekamp Massey Algorithm. Burst-error correction capability of
BCH codes. Reed Solomon Codes – BCH code viewpoint. Vandermonde matrix view point

Module 4 Convolutional Codes and Turbo codes Review of convolutional codes :

Encoding, state diagram, trellis diagram , Viterbi Decoding, BCJR algorithm, Turbo Codes:
Turbo encoder, Parallel concatenation decoding

Module 5 Reed-Muller Codes and Polar Codes Reed Muller Codes, Encoding and
decoding of RM (1, m) codes. Majority-logic decoding of Reed-Muller codes. Polar Codes –
Introduction, polarization of BEC channels, Polar transform and frozen bits.

Course Plan

No. of
No Topic
Lectures
1 Linear Block Codes and Bounds
Course Overview-Relevance of Error correction schemes in
1.1 1
communication systems
Repetition coding, concepts of Code rate, Hamming Distance,
1.2 Minimum Distance, Error detecting and correcting capability. 1
1.3 Review of Group,ring, fields and vector spaces 2
Linear Block codes-Generator and parity check matrices,encoding,
1.4 standard array and syndrome decoding, Hamming codes 2

1.5 Maximum Distance Separable (MDS) codes 1

Bounds on size of codes: Hamming bound, Singleton bound,
1.6 2
Plotkin bound, Gilbert-Varshamov bound
2 LDPC Codes
Regular and Irregular Low Density Parity Check Codes-Tanner
2.1 1
graph
Message Passing decoding-Hard decision and
2.2 1
Soft decision
2.3 Message passing decoding in a BEC 1
2.4 Bit flipping algorithm for decoding 1
2.5 Belief Propagation decoding: Sum Product algorithm 2
3 BCH and RS codes
Galois Fields -- Irreducible and Primitive Polynomials, Primitive
3.1 elements, Field extension, Conjugate elements and Minimal 2
Polynomials. Cyclotomiccosets.

BCH Codes - Design, BCH Bound, Decoding BCH codes –

Decoding BCH – the general outline, computation of the
3.2 syndrome, error locator polynomial, Chien Search algorithm, 4
Finding the error locator polynomial. Berlekamp Massey
Algorithm. Burst-error correction capability of BCH codes.

Reed Solomon Codes – BCH code viewpoint. Vandermonde

matrix view point. MDS property. Generalized Reed-Solomon
3.3 codes. Application of BCH decoding algorithms to Reed-Solomon 4
decoding. Sudan’s algorithm for decoding. Use of RS codes in
disks and cloud storage.

4 Convolutional Codes and Turbo codes

Review of convolutional codes: Encoding, state diagram, trellis
4.1 1
diagram
4.2 Review of convolutional codes: Viterbi Decoding 1
4.3 Decoding of convolutional codes: BCJR algorithm 2
4.4 Turbo Codes: Turbo encoder, Parallel concatenation decoding 3
5 Reed-Muller Codes and Polar Codes
Reed Muller Codes, Encoding and decoding of RM (1, m) codes.
5.1 Majority-logic decoding of Reed-Muller codes. 4

Polar Codes – Introduction, polarization of BEC channels, Polar

5.2 4
transform and frozen bits.
Text Books

1. Shu Lin, D. J Costello Jr. Error Control Coding: Fundaments and Applications, Prentice
Hall

Reference Books:

1. Ron M Roth, Introduction to Coding Theory, Cambridge University Press

2. T. Richardson, R. Urbanke, Modern Coding Theory, Cambridge University Press

3. A. Thangaraj, LDPC and Polar Codes in 5G Standard, NPTEL

 CODE COURSE NAME CATEGORY L T P CREDIT
MULTIRATE SIGNAL
PROGRAM
221EEC020 PROCESSING AND 3 0 0 3
ELECTIVE 2
WAVELETS

Preamble: Multirate systems play a central role in many areas of signal processing, such as
filter bank theory and multiresolution theory. This course imparts a comprehensive
knowledge of topics in multirate signal processing and wavelets, essential in some of the
standard signal processing techniques such as signal analysis, denoising and other
applications.

Course Outcomes: After the completion of the course the student will be able to

Apply sampling rate conversions, decimation and interpolation as part of Signal

CO 1
Processing techniques
CO 2 Design a perfect reconstruction filter bank system
Analyze the signal decomposition and reconstruction using tree structured filter
CO 3
banks
CO 4 Design an orthogonal/biorthogonal wavelet system according to the application.
CO 5 Implement the wavelet based decomposition using appropriate filter bank structure
CO 6 Apply wavelets for signal analysis and other applications

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 3
CO 2 3 2
CO 3 3
CO 4 3
CO 5 3
CO 6 3 2

Assessment Pattern

Bloom’s Category End Semester Examination

Apply 30
Analyse 20
Evaluate 10
Create
Mark distribution

Total ESE
CIE ESE
Marks Duration

100 40 60 2.5 hours

Continuous Internal Evaluation Pattern:

Preparing a review article based on peer reviewed original publications 15 marks

(minimum 10 Publications shall be referred):
Course based task/Seminar/Data Collection and interpretation: 15 marks
Test paper, 1 no.: 10 marks
(Test paper shall include minimum 80% of the syllabus.)

End Semester Examination Pattern:

The end semester examination will be conducted by the respective College. There will be two
parts; Part A and Part B. Part A will contain 5 numerical/short answer questions with 1
question from each module, having 5 marks for each question (such questions shall be useful
in the testing of knowledge, skills, comprehension, application, analysis, synthesis, evaluation
and understanding of the students). Students should answer all questions. Part B will contain
7 questions (such questions shall be useful in the testing of overall achievement and maturity
of the students in a course, through long answer questions relating to theoretical/practical
knowledge, derivations, problem solving and quantitative evaluation), with minimum one
question from each module of which student should answer any five. Each question can carry
7 marks.
Model Question Paper

Name:
Reg. No: APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY E
FIRST SEMESTER M.TECH DEGREE EXAMINATION

MULTIRATE SIGNAL PROCESSING AND WAVELETS

Subject: 221EEC020
Time: 2.5 Hours Maximum: 60 Marks

Part A

Answer all questions. Each question carries 5 marks (5x5=25 Marks)

1. Obtain the type-I polyphase decomposition of decimation filter with transfer

function 𝐻𝐻(𝑧𝑧) = 1 + 5𝑧𝑧 −1 + 10𝑧𝑧 −2 + 10𝑧𝑧 −3 + 5𝑧𝑧 −4 + 𝑧𝑧 −5 .

2. The prototype filter of a 2 channel QMF is given as

𝐻𝐻0 (𝑧𝑧) = 𝑎𝑎 + 𝑏𝑏𝑧𝑧 −1 + 𝑐𝑐𝑧𝑧 −2 + 𝑑𝑑𝑧𝑧 −3 − 𝑒𝑒𝑧𝑧 −5 . Design the remaining filtersof the
filter bank.

3. Obtain the time frequency variation of the function(𝑡𝑡) = 𝑒𝑒 −𝑎𝑎|𝑡𝑡| .

4. Differentiate the concepts of orthogonality and biorthogonality.

5.Explain how wavelets can be used for facial recognition.

Part B

Answer any FIVE full questions; each question carries 7 marks. (5x7=35)

6 a) Consider the following multirate system. Obtain the expression for its
output. (3 Marks)

b) Derive the frequency domain input-output relation of a decimator. (4 Marks)

7. a) Design a sampling rate converter to convert a digital audio signal of 48kHz
sampling rate to 32kHz sampling rate. (4 Marks)

b) Develop a computationally efficient realization of a factor of 3 interpolator

employing a length 15 linear phase FIR filter. (3 Marks)

8. a) Design a 3 channel perfect reconstruction filter bank given its polyphase

components as
 (4 Marks)

b) For a two channel QMF, let the analysis filter

𝐻𝐻0 (𝑧𝑧) = 2 + 5𝑧𝑧 −1 + 3𝑧𝑧 −2 − 8𝑧𝑧 −3 + 8𝑧𝑧 −4 .
Design remaining analysis filters and synthesis filters which result in perfect
reconstruction. Calculate the alias component matrix for the filter bank. (3 Marks)

9 a) Derive the constraint equations for the design of a Daubechies 6-tap orthogonal wavelet
system. (7 Marks)

10 a) Explain B-spline scaling functions. Derive the B-spline scaling function of

order 2 from Haar scaling function. (3 Marks)

b) Derive the filter bank structure to obtain the fine scale coefficients sj+1(k) of
any signal f(t) from its smooth (sj (k)) and detail coefficients (dj (k)) at coarse
resolution, and hence explain it with the required block diagram. (4 Marks)

11 a) Derive the analysis and synthesis side equations for a biorthogonal wavelet system.
(7 Marks)

12. Deduce the method of decomposing a given sequence into wavelet packets using Haar
wavelets and draw the decomposition tree with an example. (7 marks)

Syllabus

Module 1:Multi-rate System Fundamentals

Review of basic multi-rate operations: up sampling and down sampling, time domain and
frequency domain analysis, Need for antialiasing and anti-imaging filters.Interpolator and
decimator design, Noble identities.Type 1 and Type 2 polyphase decomposition, 2-channel
and N-channelpolyphase decomposition .Efficient structures for decimation and interpolation
filters, efficient structures for fractional sampling rate conversion.

Module 2:Multi-rate Filter Banks

Overview of Maximally decimated filter banks and non-maximally decimated filter bank.
Uniform DFT filter banks - design, polyphase implementation.Two-channel critically
sampled filter banks.Amplitude-Complementary 2-Channel Filter Bank
Example - Two channel Haar Filter bank and its polyphase decomposition, Quadrature mirror
filter (QMF) bank, Errors in the QMF bank, conditions for perfect reconstruction, polyphase
implementation. Design of perfect reconstruction M- channel Filter Banks, Overview of
Uniform and non-uniform tree structured filter banks, Dyadic filter bank.
Module 3:Continuous and Discrete Wavelet Transform

Short Time Fourier Transform (STFT), STFT as a bank of filters, Choice of window function
and time frequency trade-off.The Uncertainty Principle and Time Frequency
Tiling,Continuous wavelet transform (CWT) and inverse CWT,Properties of Wavelets used
in CWT, Admissibility condition.Concept of orthogonal and orthonormal basis functions,
function spaces. Discrete Wavelet Transform.Haar Scaling Function, Nested Spaces, Haar
Wavelet Function, Orthogonality of scaling and translate functions, Normalization of Haar
Bases at different Scales, Refinement Relation with respect to normalized bases.Support of a
Wavelet System, Daubechies Wavelets.

Module 4:Design of Orthogonal & Biorthogonal wavelets systems

Designing Orthogonal Wavelet systems - a direct approach, Frequency domain approach for
designing wavelets. Implementation using tree structured QMF bank and equivalent M-
channel filter bank.Designing Biorthogonal Wavelet systems: Biorthogonality in Vector
Space, Biorthogonal Wavelet Systems.Signal Representation Using Biorthogonal Wavelet
System,Biorthogonal Analysis and Biorthogonal Synthesis.Construction of Biorthogonal
Wavelet Systems-B-splines,Computation of the discrete wavelet transform using Mallat
Algorithm and Lifting Scheme

Module 5:Wavelet Packet Analysis and applications of wavelets

Wavelet Packet Transform – Signal representation using wavelet packet analysis
Applications of Wavelets and Wavelet Packets in Signal and Image compression
Wavelet based signal denoising.

Course Plan
Topic No. of
No
Lectures
1 Multi-rate System Fundamentals:
Review of basic multi-rate operations: up sampling and down
sampling, time domain and frequency domain analysis, Need for
1.1 3
antialiasing and anti-imaging filters.
Interpolator and decimator design, Noble identities.
Type 1 and Type 2 polyphase decomposition, 2-channel and N-
1.2 2
channelpolyphase decomposition
Efficient structures for decimation and interpolation filters,
1.3 3
efficient structures for fractional sampling rate conversion.
2 Multi-rate Filter Banks
Overview of Maximally decimated filter banks and non-maximally
decimated filter bank. Uniform DFT filter banks - design,
polyphase implementation.
2.1 Two-channel critically sampled filter banks 3
Amplitude-Complementary 2-Channel Filter Bank
Example - Two channel Haar Filter bank and its polyphase
decomposition.
 Quadrature mirror filter (QMF) bank , Errors in the QMF bank,
2.2 2
conditions for perfect reconstruction, polyphase implementation
2.3 Design of perfect reconstruction M- channel Filter Banks 2
Overview of Uniform and non-uniform tree structured filter banks. 1
2.4 Dyadic filter bank.

3 Continuous and Discrete Wavelet Transform

Short Time Fourier Transform (STFT), STFT as a bank of filters,
3.1 2
Choice of window function and time frequency trade-off.
3.2 The Uncertainty Principle and Time Frequency Tiling 1
Continuous wavelet transform (CWT) and inverse CWT.
3.3 2
Properties of Wavelets Used in CWT, Admissibility condition.
Concept of orthogonal and orthonormal basis functions, function
spaces. Discrete Wavelet Transform.
Haar Scaling Function, Nested Spaces.
3.4 Haar Wavelet Function, Orthogonality of scaling and translate 4
functions, Normalization of Haar Bases at different Scales,
Refinement Relation with respect to normalized bases.
Support of a Wavelet System, Daubechies Wavelets.
4 Design of Orthogonal & Biorthogonal wavelets systems
Designing Orthogonal Wavelet systems - a direct approach,
Frequency domain approach for designing wavelets.
4.1 3
Implementation using tree structured QMF bank and equivalent
M-channel filter bank.
Designing Biorthogonal Wavelet systems: Biorthogonality in
Vector Space, Biorthogonal Wavelet Systems.
4.2 Signal Representation Using Biorthogonal Wavelet System, 3
Biorthogonal Analysis and Biorthogonal Synthesis.
Construction of Biorthogonal Wavelet Systems-B-splines
Computation of the discrete wavelet transform using Mallat
4.3 3
Algorithm and Lifting Scheme
5 Wavelet Packet Analysis and applications of wavelets
Wavelet Packet Transform – Signal representation using wavelet
5.1 2
packet analysis
Applications of Wavelets and Wavelet Packets in Signal and
5.2 3
Image compression
5.3 Wavelet based signal denoising. 2

Reference Books

1.P. P. Vaidyanathan, Multirate Systems and Filter Banks, Pearson Education, 2006.
2.Fredric J Harris, Multirate Signal Processing for Communication Systems, 1st Edition,
Pearson Education, 2007

3.Sanjit K. Mitra, Digital Signal Processing: A Computer based Approach, Special Indian
Edition,McGraw Hill, 2013.
4. Spectral Audio Signal Processing, Julius O. Smith III, W3K Publishing, 2011.

5. N.J. Fliege, Multirate Digital Signal Processing, John Wiley, 1994.

6. K. P. Soman, K. I. Ramachandran, N. G. Resmi, PHI, Insight into wavelets :From
theory topractice

7. G.Strang& T. Nguyen , Wavelets and Filter bank, Wellesly-Cambridge

8. M. Vetterli& J. Kovacevic, Wavelets and sub band coding, Prentice Hall

 CODE COURSE NAME CATEGORY L T P CREDIT
ADAPTIVE SIGNAL PROGRAM
221EEC021 3 0 0 3
PROCESSING ELECTIVE 2

Preamble: This course introduces the design and analysis of signal processing algorithms
which can automatically change the system parameters to get a desired output, when a
stationary random signal is applied to it.

Course Outcomes: After the completion of the course the student will be able to

CO 1 Analyse stationary random signals and adaptive systems.

CO 2 Analyse performance of the gradient search algorithms.
CO 3 Evaluate the effect of gradient noise in weight vector solution.
CO 4 Analyse LMS algorithms, adaptive recursive filters and Kalman filters.
CO 5 Apply the adaptive systems for applications in system modelling and inverse
adaptive modelling.

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 1
CO 2 2 2 2
CO 3 2 2 2
CO 4 2 2 3 2
CO 5 2 2 3 3 2

Assessment Pattern

Bloom’s Category End Semester Examination

Understand 20%
Apply 70%
Analyse 10%
Evaluate
Create

Mark distribution

Total ESE
CIE ESE
Marks Duration

100 40 60 2.5 hours

Continuous Internal Evaluation Pattern:

Preparing a review article based on peer reviewed original 15 marks

publications (minimum 10 Publications shall be referred):
Course based task/Seminar/Data Collection and interpretation: 15 marks
Test paper, 1 no.: 10 marks
(Test paper shall include minimum 80% of the syllabus.)

End Semester Examination Pattern:

The end semester examination will be conducted by the respective College. There will be two
parts; Part A and Part B. Part A will contain 5 numerical/short answer questions with 1
question from each module, having 5 marks for each question (such questions shall be useful
in the testing of knowledge, skills, comprehension, application, analysis, synthesis, evaluation
and understanding of the students). Students should answer all questions. Part B will contain
7 questions (such questions shall be useful in the testing of overall achievement and maturity
of the students in a course, through long answer questions relating to theoretical/practical
knowledge, derivations, problem solving and quantitative evaluation), with minimum one
question from each module of which student should answer any five. Each question can carry
7 marks.
 Model Question Paper Subject: 221EEC021

Name:
Reg. No: E

APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY

FIRST SEMESTER M. TECH DEGREE EXAMINATION

ADAPTIVE SIGNAL PROCESSING

Time: 2.5 Hours Maximum: 60 Marks

PART A (Answer all questions)

1 Prove that the autocorrelation of a non-zero wide sense stationary random 5 marks
sequence is positive definite. Compute the 2x2 auto correlation matrix of
, where a and are constants, N(.) refers to normal
distribution.

2 Compare and contrast various configurations of adaptive systems. Give 5 marks

one example for each.

3. Evaluate the expression for MSE performance surface of a system with 5 marks
at the weight vector
Calculate .

4. Obtain the expression for weight updation in the LMS algorithm. What is 5 marks
the condition for convergence in the LMS algorithm?

5. Illustrate the method for adaptive modelling of earth’s impulse response 5 marks
for geophysical exploration

PART B (Answer any five question)

6 Compute the most significant two Eigen vectors of the correlation 7 marks

matrix

7. Cos πk/7 7 marks

z-1
Sin πk/7

W0 W1
-
+ εk
+
Consider the adaptive system given in figure. Write an expression for the
 performance surface. Determine a range for the adaptive gain constant.
Calculate W* and .

Module 2

8. Determine the autocorrelation matrix, performance penalty and 7 marks

perturbation of a two-weight system with performance surface
and δ=1/4.

9. Derive the prediction and correction equations of the discrete Kalman 7 marks
Filter.

10 a. Derive the expressions for the weight updating in adaptive recursive 5 marks
filter.

10 b. Identify major challenges in adaptive recursive filter implementation. 2 marks

11. Explain in detail about the adaptive modelling of multipath 7marks

communication channels.

12 Derive the expression for the covariance of noise in the gradient 7marks
estimation in the case of steepest descent method?
 Syllabus

(For 3 credit courses, the content can be for 40 hrs)

Review of discrete time stochastic process and auto correlation matrix - Introduction to
adaptive systems - performance functions - Gradient search methods -Gradient estimation
and its effects on the adaptation - LMS algorithm - Adaptive recursive filters - discrete
Kalman filter- application of adaptive filtering.

Course Plan

No. of
No Topic
Lectures
1 Review of discrete time stochastic process and auto correlation matrix
1.1 Univariate and Multivariate random sequences, Gaussian noise 3
1.2 Autocorrelation matrix of stationary process and its properties 2
1.3 Eigen decomposition, properties of Eigen vectors, whitening 3
2 Introduction to adaptive systems
Introduction to adaptive systems – definitions– characteristics–
2.1 2
configurations – applications
Adaptive linear combiner – MSE performance function –
2.2 2
Wiener-Hopf equation
Searching the MSE performance function – Newton's method
2.3 for searching the performance function, stability and 2
convergence
Steepest descent algorithm – Stability and convergence –
2.4 2
Learning curve
3 Gradient estimation and its effects on adaptation
3.1 Performance penalty and perturbation 2
3.2 Effects on weight vector solution, covariance of weight vector 3
3.3 Excess MSE, mis adjustment and time constants 2
4 Adaptive algorithms and structures
LMS algorithm, Convergence of weight vector and learning
4.1 2
curve
4.2 Adaptive recursive filters 2
4.3 Discrete Kalman Filter, filtering example 4
5 Applications of adaptive filtering
5.1 Adaptive modelling of multipath communication channel 2
5.2 Adaptive modelling for FIR filter synthesis 2
5.3 Adaptive equalization of telephone channels 2
5.4 Adapting poles and zeros for IIR digital filter synthesis 2
Reference Books

1. Bernard Widrow and Samuel D. Stearns, Adaptive signal processing, Pearson

education.
2. Simon Hykins, Adaptive filter theory, Fifth edition, Pearson education.
3. A. Papaulis and U. Pillai, Probability Random Variables and Stochastic Processes,
4th Edition, McGraw Hill Education.
4. S. Thomas Alexander, Adaptive Signal Processing – Theory and Applications,
SpringerVerlag
 CODE COURSE NAME CATEGORY L T P CREDIT
PROGRAM
221EEC022 INTERNET OF THINGS 3 0 0 3
ELECTIVE 2

Preamble: The core modules of this elective course include introduction to Internet of
Things (IoT), IoT protocol and software, IoT point to point communication technologies, IoT
security and IoT Platform. This course aims to teach the student to understand the concepts of
IoT and its applications.

Prerequisites: NIL

Course Outcomes: After the completion of the course the student will be able to

CO 1 Identify the IoT networking components with respect to OSI layer.

CO 2 Understand the communication networks and protocols used in IoT

CO 3 Understand the cloud resources, data analysis and applications.

CO 4 Understand IoT security and threat models analysis.

CO 5 Understand various IoT applications and its variants.

CO 6 Design and develop prototype models for various IoT applications.

Mapping of course outcomes with program outcomes

PO 1 PO 2 PO 3 PO 4 PO 5 PO 6 PO 7
CO 1 3 3 3 3
CO 2 3 3 3 3
CO 3 3 3 3 3
CO 4 3 3 3 3
CO 5 3 3 3 3
CO 6 3 3 3 3

Assessment Pattern

Bloom’s Category End Semester Examination

Apply 20
Analyse 20
Evaluate
Create 20
Mark distribution

Total Marks CIE ESE ESE Duration

100 40 60 2.5 hours

Evaluation shall only be based on application, analysis or design-based questions (for both
internal and end semester examinations).
Continuous Internal Evaluation Pattern:

Continuous Internal Evaluation: 40 marks

Preparing a review article based on peer reviewed original publications (minimum 10
publications shall be referred): 15 marks
Course based task/Seminar/Data collection and interpretation: 15 marks
Test paper, 1 no. : 10 marks
Test paper shall include minimum 80% of the syllabus.

End Semester Examination Pattern:

The end semester examination will be conducted by the respective College. There will be two
parts; Part A and Part B.
Part A will contain 5 numerical/short answer questions with 1 question from each module;
having 5 marks for each question (such questions shall be useful in the testing of knowledge,
skills, comprehension, application, analysis, synthesis, evaluation and understanding of the
students). Students should answer all questions.
Part B will contain 7 questions (such questions shall be useful in the testing of overall
achievement and maturity of the students in a course, through long answer questions relating
to theoretical/practical knowledge, derivations, problem solving and quantitative evaluation),
with minimum one question from each module of which student should answer any five.
Each question can carry 7 marks.
Note: The marks obtained for the ESE for an elective course shall not exceed 20% over the
average ESE mark % for the core courses. ESE marks awarded to a student for each elective
course shall be normalized accordingly. For example, if the average end semester mark % for
a core course is 40, then the maximum eligible mark % for an elective course is 40+20 = 60
%.
 Model Question Paper

APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY

FIRST SEMESTER M. TECH DEGREE EXAMINATION, (Model Qtn Paper)

Program: M.Tech. in EC3 (Signal Processing, Signal Processing & Embedded Systems,
Communication Engineering & Signal Processing)

Course Code: 221EEC022

Course Name: INTERNET OF THINGS

Max. Marks: 60 Duration: 150 Minutes

PART A

Answer All Questions. Each Carries 5 marks.

1. How is data link addressing different from network addressing?

2. Explain the CoAP message format.

3. Discuss the components in the open source cloud platform OpenStack..

4. List the standard encryption protocols used in IoT. Explain any one algorithm with
example.

5. With the help of block diagram explain the IoT application in healthcare. Explain each
modules of the network.

PART B

Answer any 5 questions: Each question carries 7 marks.

6. What is internet protocol suit? How is it differ from th ISO-OSI model?

7. Design an IoT based water quality monitoring system using MQTT protocol. Explain the
publish subscribe model in detail.

8. Design a LoRa based smart street lighting system for smart cities (block diagram only).
Discuss the security features of LoRaWAN.

9. Explain the requirements of privacy and security, vulnerabilities from threats, and need of
threat-analysis in IoT

10. Discuss on the cryptographic functions and their security services.

11. How IoT is related to big data analytics? Explain with an example.

12. Design an M2M Smart meter which allow you to track energy consumption in real-time.
Draw the block diagram of the network and explain each modules
 SYLLABUS

Module 1: 8 hours

Introduction to IoT: Basics of Networking- network types,layered network models,

addressing,TCP/IP Transport layer, Emergence of IoT- evolution of IoT, IoT networking
components ,IoT network architecture and design-comparing IoT architectures, simplified
IoT architecture,Smart objects: the things in IoT-sensors, actuators and smart objects, sensor
networks

Module 2: 8 hours

IoT protocols and Software: Connecting smart objects: communications criteria, IoT
connectivity technologies(IEEE 802.15.4, Zigbee,Sigfox,LoRA,NB-IoT,Wi-Fi,Bluetooth),IoT
communication technologies: infrastructure protocols (Ipv6,6LoWPAN), Data protocols
(MQTT,CoAp, AMQP, XMPP, SOAP, REST, WebSocket)

Module 3: 8 hours

Introduction to Cloud computation and Big data analytics: cloud computing:

introduction, cloud models, cloud implementation-cloud simulation,open source cloud
(OpenStack),commercial cloud:AWS,introduction to data analytics for IoT, machine learning,
Big data analytics tools and technology – Hadoop,Edge streaming analytics,Network
analytics.

Module 4: 8 hours

IoT security: common challenges in IoT security,fundamentals of cryptography:

cryptographic algorithms and their security services, the lightweight features of cryptographic
algorithms, Quadruple Trust Model for IoT-A – Threat Analysis and model for IoT-A, Cloud
security.

Module 5: 8 hours

IoT application and its Variants: Case studies: IoT for smart cities, health care, agriculture,
smart meters.M2M, Web of things, Cellular IoT, Industrial IoT, Industry 4.0, IoT standards.
Course Plan

No. of
No Topic
Lectures

1 Introduction to IoT:
Basics of Networking- network types,layered network models,
1.1 2
addressing,TCP/IP Transport layer
1.2 Emergence of IoT- evolution of IoT,IoT networking components 3
IoT network architecture and design-comparing IoT
1.3 2
architectures,simplified IoT architecture.
Smart objects: the things in IoT-sensors, actuators and smart
1.4 1
objects,sensor networks
2 IoT protocols and Softwares:
2.1 Connecting smart objects: communications criteria 2
IoT connectivity technologies(IEEE 802.15.4,Zigbee,Sigfox,LoRA,NB-
2.2 3
IoT,Wi-Fi,Bluetooth)
IoT communication technologies: infrastructure
2.3 protocols(IPv6,6LoWPAN),Data 3
protocols(MQTT,CoAp,AMQP,XMPP,SOAP,REST,WebSocket)
3 Introduction to Cloud computation and Big data analytics:
3.1 cloud computing: introduction, cloud models 1
cloud implementation-cloud simulation,open-source cloud(OpenStack),
3.2 2
commercial cloud:AWS
3.3 introduction to data analytics for IoT, machine learning 1
3.4 Big data analytics tools and technology - Hadoop 2
3.5 Edge streaming analytics, Network analytics 2
4 IoT security:
4.1 common challenges in Iot security 1

fundamentals of cryptography: cryptographic algorithms and their

4.2 2
security services, the lightweight features of cryptographic algorithms

Quadruple Trust Model for IoT-A – Threat Analysis and model for IoT-
4.3 2
A,
4.4 Cloud security 3
5 IoT application and its Variants:
 5.1 Case studies: IoT for smart cities, health care, agriculture, smart meters. 3

5.2 M2M, Web of things, Cellular IoT 3

5.3 Industrial IoT, Industry 4.0, IoT standards. 2

Text Books

1. S.Misra, A. Mukherjee, and A.Roy, 2020. Introduction to IoT. Cambridge

University Press.
2. David Hanes, Gonzalo Salgueiro, Patrick Grossetete, Rob Barton and Jerome Henry,
―IoT Fundamentals: Networking Technologies, Protocols and Use Cases for
Internet of Things, Cisco Press, 2017.
3. Rajkamal, “Internet of Things : Architecture and Design Principles”, McGraw Hill
(India) Private Limited
4. Chuan Kun Wn,”Internet of Things Security: Architectures and Security
Measures”,Springer 2021.
5. Alessandro Bassi, Martin Bauer, Martin Fiedler, Thorsten Kramp, Rob van
Kranenburg, Sebastian Lange, Stefan Meissner, “Enabling things to talk – Designing
IoT solutions with the IoT Architecture Reference Model”, Springer Open, 2016

Reference Books

1. Jan Holler, Vlasios Tsiatsis, Catherine Mulligan, Stamatis Karnouskos, Stefan

Avesand, David Boyle, “From Machine to Machine to Internet of Things”, Elsevier
Publications, 2014.
2. LuYan, Yan Zhang, Laurence T. Yang, Huansheng Ning, The Internet of Things:
From RFID to the Next-Generation Pervasive Network, Aurbach publications,
March,2008.
3. Vijay Madisetti , Arshdeep Bahga, Adrian McEwen (Author), Hakim Cassimally
“Internet of Things A Hands-on-Approach” Arshdeep Bahga & Vijay Madisetti,
2014.
4. Asoke K Talukder and Roopa R Yavagal, “Mobile Computing,” Tata McGraw Hill,
2010.
5. Barrie Sosinsky, “Cloud Computing Bible”, Wiley-India, 2010
6. RonaldL. Krutz, Russell Dean Vines,Cloud Security: A Comprehensive Guide to
Secure Cloud Computing,Wiley-India, 2010
 CODE COURSE NAME CATEGORY L T P CREDIT
PROGRAM
221EEC023 OPTIMIZATION TECHNIQUES 3 0 0 3
ELECTIVE 2

Preamble: This course aims to enable the students to apply suitable optimization techniques for
various applications.
Prerequisite: Nil

Course Outcomes: After the completion of the course the student will be able to
CO1 Outline the mathematical building blocks of optimization

CO2 Model and solve linear programming problems

CO3 Apply principles and techniques for solving nonlinear programming models

CO4 Investigate and assess constrained convex optimization problems

CO5 Appreciate prominent heuristic optimization algorithms
CO 6 Solve optimization problem through optimization software

Mapping of course outcomes with program outcomes

PO1 PO2 PO3 PO4 PO5 PO6 PO7

CO1 3 2 3 3 1
CO2 3 1 2 3 3 1 1
CO3 3 1 2 3 3 1 1
CO4 3 1 2 3 3 1 1
CO5 3 2 3 3 1 1
CO 6 3 1 2 3 3 1 1

Assessment Pattern

Bloom’s Category End Semester Examination

Apply 30
Analyze 20
Evaluate 10
Create

Mark distribution

Total Marks CIE ESE ESE Duration

100 40 60 2.5hours
Continuous Internal Evaluation Pattern:
• Preparing a review article based on peer reviewed original publications (minimum 10
publications shall be referred): 15 marks
• Course based Task/ Seminar/ Data: 15 marks
• Test paper 1 no. (Test paper shall include minimum 80% of syllabus):10marks

End Semester Examination Pattern:

The end semester examination will be conducted by respective College. There will be two parts; Part A
and Part B. Part A will contain 5 numerical/short answer questions with 1 question from each module,
having 5 marks for each question (such questions shall be useful in the testing of knowledge, skills,
comprehension, application, analysis, synthesis, evaluation and understanding of the students). Students
should answer all questions. Part B will contain 7 questions (such questions shall be useful in the
testing of overall achievement and maturity of the students in a course, through long answer questions
relating to theoretical/practical knowledge, derivations, problem solving and quantitative evaluation),
with minimum one question from each module of which student should answer any five. Each question
can carry 7 marks.
Note: The marks obtained for the ESE for an elective course shall not exceed 20% over the average
ESE mark % for the core courses. ESE marks awarded to a student for each elective course shall be
normalized accordingly. For example, if the average end semester mark % for a core course is 40, then
the maximum eligible mark % for an elective course is 40+20 = 60 %.

SYLLABUS
Module 1:
Mathematical Background:
Vector norm, Matrix norm, Inner product, Norm ball, Interior point, Closure and boundary,
Complement, scaled sets, and sum of sets, Supremum and infimum, Vector subspace, Function,
Continuity of function, Derivative and gradient, Hessian, Convex sets and convex functions.
Introduction to optimization - Optimal problem formulation, Engineering applications of optimization,
Optimization techniques - Classification.

Module 2:
Linear Programming:
Linear Programming - Formulation of the problem, Graphical method, Simplex method, Artificial
variable techniques, Duality Principle, Dual simplex method.

Module 3:
Non-linear programming:
Unimodal Function, Elimination methods – Fibonacci method, Golden section method, Direct search
methods – Random walk, Grid search method, Indirect search methods – Steepest descent method,
Newton’s method.
Module 4:
Convex optimization:
Standard form of convex optimization problems, Global optimality, An optimality criterion for
differentiable convex function, Lagrange dual function and conjugate function, Lagrange dual
problem, Karush–Kuhn–Tucker (KKT) optimality conditions, Lagrange dual optimization.

Module 5:
Optimization algorithms:
Genetic algorithm, Neural network-based optimization, Ant colony optimization, Particle swarm
optimization. Optimization Libraries in Python: scipy.optimize, CVXPY, CVXOPT.

Text Books
1. Chong-Yung-Chi, Wei-Chiang Li, Chia-Hsiang Lin, Convex Optimization for Signal Processing
and Communications – From fundamentals to applications, CRC press.
2. Sukanta Nayak, Fundamentals of Optimization Techniques with Algorithms, Academic press.
3. Singiresu S. Rao, Engineering Optimization: Theory and Practice, John Wiley and Sons.

Reference Books
1. Igor Griva, ArielaSofer, Stephen G Nash,Linear and Nonlinear Optimization, Second edition,
SIAM.
2. Kalyanmoy Deb,Optimization for Engineering: Design Algorithms and Examples, Second
edition,Prentice Hall.
3. David G Luenberger, Linear and Nonlinear Programming, Second edition, Addison-Wesley.
No.
No Topic ofLectures
1 Mathematical Background:
1.1 Vector norm, Matrix norm, Inner product, Norm ball 1
1.2 Interior point, Closure and boundary 1
1.3 Complement, scaled sets, and sum of sets, Supremum and infimum 1
1.4 Vector subspace, Function, Continuity of function, 1
1.5 Derivative, gradient and Hessian 1
1.6 Convex sets and convex functions 1
1.7 Introduction to optimization - Optimal problem formulation 1
Engineering applications of optimization, Optimization techniques
1.8 1
Classification
2 Linear Programming:
Linear Programming - Formulation of the problem, Graphical
2.1 2
method
2.2 Simplex method 2
2.3 Artificial variable techniques, Duality Principle 2
2.4 Dual simplex method 2
3 Non-linear programming:
3.1 Uni-modal Function 1
3.2 Elimination Methods: (1) Fibonacci Method 1
3.3 Elimination Methods: (2) Golden Section Method 1
3.4 Direct Search Methods: (1) Random Walk 1
3.5 Direct Search Methods: (2) Grid Search Method 1
3.6 Indirect Search Method: (1) Steepest Descent Method 1
3.7 Indirect Search Method: (2) Newton’s Method 2
4 Convex optimization:
4.1 Standard form of convex optimization problems 1
Global optimality, An optimality criterion for differentiable convex
4.2 2
function
4.3 Lagrange dual function and conjugate function 1
4.4 Lagrange dual problem 2
4.5 Karush–Kuhn–Tucker (KKT) optimality conditions 2
5 Optimization algorithms:
5.1 Genetic algorithm 1
5.2 Neural network-based optimization 2
5.3 Ant colony optimization 2
5.4 Particle swarm optimization. 1
Optimization Libraries in Python: scipy.optimize, CVXPY,
5.5 2
CVXOPT
 Model Question paper

APJ ABDUL KALAM TECHNOLOGICAL UNIVERSITY

FIRST SEMESTER M. TECH DEGREE EXAMINATION

Electronics and Communication Engineering

Streams: Signal Processing, Signal Processing and Embedded

Systems, Communication Engineering & Signal Processing

Course Code and Course Name: 221EEC023 Optimization Techniques

Max.Marks: 60 Duration:2.5 Hours

PARTA
Answer ALL Questions. Each Carries 5 marks.

1. Define the gradient of the function. Demonstrate its importance in the multi-variable
optimization
2 State and prove complementary slackness theorem.

3 Using Newton’s method minimize f=(3x1-1)3 + 4x1x2 + 2x22 by taking initial point as (1,2)

4 State and prove Kuhn – Tucker conditions in non-linear programming

5 Differentiate supervised and supervised learning.

PART–B
Answer any FIVE full questions; each question carries 7 marks.
6.a) 0.5000 0.5000 0.5000 0.5000
0.6533 0.2706 −0.2706 −0.6533
Given, Ѱ1-1= � �,
0.5000 −0.5000 −0.5000 0.5000
0.2706 −0.6533 0.6533 −0.2706
1 1 1 1 4
1 −𝑗𝑗 −1 𝑗𝑗 5 4 marks
Ѱ2-1= � � and 𝑥𝑥 = � �
1 −1 1 −1 5
1 𝑗𝑗 −1 −𝑗𝑗 4
Let Ѱ1 andѰ2 be two change of basis matrices. Obtain the representation of the
given vector 𝑥𝑥 in terms of Ѱ1 andѰ2. Calculatel0, l1 and l∞ norms of the
representation.
6.b) Discuss the convexity and concavity of the following functions
a) 𝑓𝑓(𝑥𝑥) = (𝑥𝑥1 + 𝑥𝑥2 )𝑒𝑒 (𝑥𝑥 1 +𝑥𝑥 2 ) 𝑥𝑥1 > 0, 𝑥𝑥2 > 0
b) 𝑓𝑓(𝑥𝑥) = 𝑥𝑥1 𝑓𝑓1 (𝑥𝑥) + 𝑥𝑥2 𝑓𝑓2 (𝑥𝑥)𝑥𝑥1 ≥ 0, 𝑥𝑥2 ≥ 0 and both 𝑓𝑓1 (𝑥𝑥) and 𝑓𝑓2 (𝑥𝑥) are 3 marks
convex functions

7.a) Congratulations! Upon graduating from college, you have immediately been
offered a high-paying position as president of the Lego Furniture Company.
Your company produces chairs (each requiring 2 square blocks and 1 3 marks
rectangular block) as well as tables (each requiring 2 square blocks and 2
rectangular blocks) and has available resources consisting of 8 rectangular
 blocks and 6 square ones. Assume chairs and tables each sell for 5 and 7
respectively, and that your company sells all of what it produces.
(i) Set up an LP whose objective is to maximize your company’s revenue.
(ii) Represent it in the standard form and matrix form.
7.b) Solve the following LPP using dual simplex method
min z = 3x1+2x2
subject to 2x1+ 3x2 ≥ 30
-x1+2x2 ≤ 6 4 marks
x1+3x2> 20
x1, x2 ≥ 0

8a) Solve the following optimization problem using Simplex algorithm. 3 marks

8 b) Consider an LPP 4 marks

max z = -x1 - x2
subject to -x1+ x2 ≥ 1
2x1- x2 ≤ 2
x1 , x 2 ≥ 0
Find the dual to the problem. Solve the primal and the dual graphically, and
verify that the results of the strong duality theorem hold.

9a) Prove that in Fibonacci search algorithm, at the end of (n-1) iterations, the 4 marks
length of the interval of uncertainty is reduced from (b1-a1) to (b1-a1) / Fn.
Moreover, show that Fibonacci method is more efficient than Golden section
search algorithm.
9 b) Use Newton’s method to solve, 3 marks
minimize 𝑓𝑓(x1,x2) = 5x14+6x24-6x12+2x1x2+5x22+15x1-7x2+13
T
Use the initial guess (1, 1) .

10a) Use the steepest descent method to solve 4 marks

minimize 𝑓𝑓(x1,x2) = 4x1 +4x1x2+2x22-3x1,
2
starting from the point T
(2, 2) .
Perform three iterations.

10 b) Describe Grid search method with a suitable example. 3 marks

 11a) Consider the problem 4 marks
min 𝑓𝑓(x1,x2,x3,x4) = x12+x22+x32+x42
subject to g1(x1,x2,x3,x4) = x1+2x2+3x3+5x4-10 = 0
g2(x1,x2,x3,x4) = x1+2x2+5x3+6x4-15 = 0
Solve the problem using Lagrangian method and compute the Lagrange
multipliers.

11 b) Formulate the Karush-Kuhn-Tucker (KKT) necessary conditions for the 3 marks

following optimization problem

min z = x1 + x2
subject to g1(x1,x2) = x13-x2 ≥ 0
g2(x1,x2) = x1 ≥ 0
g3(x1,x2) = x2 ≥ 0

12 a) Explain Ant Colony Optimization algorithm in detail. 4 marks

12 b) Describe Particle Swam Optimization algorithm. List out its advantages, 3 marks
disadvantages and applications.