Skip to main content


EEE3030 : Signal Processing and Machine Learning

  • Offered for Year: 2023/24
  • Module Leader(s): Professor Jeffrey Neasham
  • Lecturer: Dr Kabita Adhikari
  • Owning School: Engineering
  • Teaching Location: Newcastle City Campus

Your programme is made up of credits, the total differs on programme to programme.

Semester 1 Credit Value: 10
Semester 2 Credit Value: 10
ECTS Credits: 10.0
European Credit Transfer System


To develop in-depth knowledge of discrete-time signal processing algorithms, and approaches to measure deterministic and random signals in the frequency domain.

To measure the computational cost of different algorithms used in time/frequency transformation.

To gain proficiency in methods to distinguish the desired signals from noise using appropriate digital filters.

Using problem-based learning to gain skills in the design and implementation of digital signal processing systems for real world applications.

This module also aims to provide underlying mathematical, statistical, and theoretical concepts of Machine Learning along with essential programming skills and expertise to design, build, and implement appropriate Machine Learning techniques for various engineering applications. The module introduces classical regression, classification and clustering models and modern deep learning models. The module includes relevant programming exercises to complement the theoretical concepts, which will allow students to gain valuable conceptual and programming skills to build, optimise and implement these models into a range of practical engineering challenges.

Outline Of Syllabus

Fundamentals of digital signals
Sampling theory, discrete time signals, Nyquist limit, ADC/DAC performance, quantisation.

Deterministic Signals
Describing the deterministic signals, transformation of deterministic-time signal into frequency domain using DFT (discrete Fourier transform) and FFT (fast Fourier transform), comparison of DFT and FFT computational loads, derivation of the DFT and matrix interpretation of the DFT, spectral leakage in DFT/FFT and mitigation approaches by windowing.

Random Signals
Describing random sequences, statistical properties related to random sequences, Wiener- Khintchine theorem.

Digital Filters
Importance and advantages of digital filters, implementation of digital filters, design of FIR filters, FIR filter design by Impulse Response Truncation, optimality of IRT method, Gibb's phenomenon, FIR filter design using window functions. Design of IIR filters by bilinear z- transform, frequency transformations, finite word length effects in IIR filters.

Multirate DSP
Decimation/interpolation, multistage approaches to digital filtering and computational efficiency, polyphase filters.

Adaptive Filters
Concept of adaptive filtering, Wiener filter, Concept of Steepest Descent Algorithm, LMS Algorithm, direct and inverse system models, adaptive noise cancellation, adaptive equalisation.

Case Studies
Examples of DSP applied to real-world problems e.g. biomedical instrumentation, radar/sonar, communication.

Introduction of Machine Learning
Machine Learning Definition and Examples. Brief introduction on: Supervised Learning, Unsupervised Learning, Data Representation, Brief introduction of Regression, Classification and Clustering Methods

Review of Linear Algebra
Vectors and Matrices, Scalar Multiplication, Matrix-Vector Multiplication, Vector inner product, Vectorization, Matrix Multiplication Properties, Inverse and Transpose, computing and plotting data, and control statements – for, while, if statements

Supervised Learning – Linear regression
Linear regression with one variable, model representation, cost function, gradient descent for linear regression, multivariate linear regression

Supervised Learning - Logistic Regression
Classification, hypothesis representation, decision boundary, cost function, regularized logistic regression, dealing with overfitting and underfitting problems, Support Vector Machines (SVM), non-linear decision boundary, SVM with Kerels

Unsupervised Learning
Clustering, K-Means Algorithm, random initialization, optimsation methods, choosing size of clusters

Neural Networks
Non-Linear Hypothesis, Neurons and the Brain, Cost Function, Backpropagation, Artificial Neural Networks

Dimensionality Reduction
Feature extraction methods, data compression and Visualisation, Principal Component Analysis, reconstruction from compressed representation

Ethical Issues
Machine Learning Ethics, Legislation and Fairness, privacy, and security of data-driven Machine Learning models

Teaching Methods

Teaching Activities
Category Activity Number Length Student Hours Comment
Structured Guided LearningLecture materials270:209:00Non-synchronous: 20 mins pre-recorded videos consisting of theory on DSP topics & MATLAB demonstrations
Scheduled Learning And Teaching ActivitiesLecture102:0020:00Discussions and Q&A sessions + Computing lab sessions for programming exercises (Sem II)
Structured Guided LearningLecture materials300:2010:00Pre-recorded video lectures, students’ Asynchronous online study (Sem II)
Guided Independent StudyAssessment preparation and completion12:002:00Written Exam (Sem II)
Guided Independent StudyAssessment preparation and completion101:3015:00Revision for final exam (Sem II)
Guided Independent StudyAssessment preparation and completion115:0015:00Coursework consisting of a DSP system design & verification in MATLAB.
Guided Independent StudyAssessment preparation and completion117:0017:00Completion and review of formative assessment (online Quizzes) (Sem II)
Scheduled Learning And Teaching ActivitiesPractical112:0022:00Present in person seminar and practical lab session – each session will review lecture materials, DSP exercises and students’ questions, then students will work on MATLAB based DSP exercise under supervision of lecturer.
Guided Independent StudyDirected research and reading270:209:00Student study time of non-synchronous pre-recorded material.
Guided Independent StudyIndependent study145:0045:00Reviewing lecture notes, online materials, general background reading and work on MATLAB exercises.
Guided Independent StudyIndependent study361:0036:00Lectures follow up: Reviewing lecture materials, building understanding, and creating comments on provided lecture documents (Sem II)
Teaching Rationale And Relationship

Non-synchronous sessions provide the fundamental concepts of the course while the lab sessions and MATLAB based exercises provide an opportunity to develop skills in design, application and testing of signal processing algorithms.

The module is delivered using mixture of teaching methods: asynchronous recorded lecture videos and scheduled in-person Q&A/discussion sessions that cover theoretical foundations of Machine Learning. In addition, practical lab sessions provide opportunity to build, test and implement those Machine Learning concepts via multiple programming exercises that will be issued throughout the course.

Assessment Methods

The format of resits will be determined by the Board of Examiners

Description Length Semester When Set Percentage Comment
Written Examination1202A75Final exam covering machine learning, optimal and adaptive filters.
Other Assessment
Description Semester When Set Percentage Comment
Written exercise1M25MATLAB based signal analysis/filtering task assessed by written report.
Formative Assessments

Formative Assessment is an assessment which develops your skills in being assessed, allows for you to receive feedback, and prepares you for being assessed. However, it does not count to your final mark.

Description Semester When Set Comment
Prob solv exercises1MWeekly exercises to gain familiarity with signal processing methods and MATLAB programming.
Prob solv exercises2MStudents will complete the set programming exercises to design and verify Machine learning models
Digital Examination2MStudents will complete online quizzes to check their understanding of the taught materials
Assessment Rationale And Relationship

The summative assignment (25%) will help students to demonstrate core understanding of course material, analysis/design skills applied to realistic DSP problems and their ability to simulate and verify algorithm performance in MATLAB. The formative MATLAB exercises will provide weekly feedback on their understanding of the DSP topics and their readiness to take on the assignment.

The written exam primarily assesses students’ knowledge and understanding on optimum filtering and fundamental Machine Learning principles. In the scheduled computing labs, students will work on set programming tasks; students will be assessed on their practical skills (formulation, design, testing and analysis skills) on those exercises.

Reading Lists