Directed Studies Projects and Supervisors

Transparent transistors

Project Supervisor: James Bullock
Offered: Summer Semester, Semester 1 & Semester 2

Transparent oxide electronics, such as those based on indium and zinc oxides, are exciting emerging technology platforms. They can be deposited at low temperatures on flexible/plastic substrates enabling various applications such as augmented reality displays and other wearable electronics. This Directed Studies project will explore the fabrication and characterization of oxide-based field effect transistors on transparent substrates. It will involve work in a laboratory environment to fabricate devices and is an excellent project for those considering a career in electronics research. It will also involve measuring and modelling the behaviour of fabricated devices to identify their potential and opportunities for further optimization. This project would suit someone in the electronics pathway of the EEE Masters course.

Silicon solar cell advanced characterization

Project Supervisor: James Bullock
Offered: Summer Semester, Semester 1 & Semester 2

Silicon solar cells dominate the photovoltaics industry and, in many parts of the world, have become the cheapest source of electricity. When optically or electrically pumped, silicon will emit photons with an energy around 1.1eV. This process is called photo- or electro-luminescence and its magnitude is proportional to the concentration of excess electrons/holes generated within the silicon layer. Measuring the intensity and spectrum of the emitted light can be very useful in assessing the optoelectronic quality of devices. Conveniently, a solar cell’s implied open circuit voltage is also proportional to the excess electron and hole concentration and thus photoluminescence can be used as an excellent proxy for device implied open circuit voltage. This provides us with a way to estimate a solar cell’s device voltage during the fabrication process without having to finish the device. This directed studies module will consist of both a practical and theoretical component. The practical component will be focused on measuring photoluminescence images of large area silicon wafers. The theoretical component will involve building models to convert the measured photoluminescence into an estimate of device implied voltage.

Contact supervisor

Metasurface lasers

Project Supervisor: Kenneth Crozier
Offered: Semester 1 & Semester 2
Keywords: Photonics and nanofabrication

The term ‘metamaterial’ has been adopted to describe any artificially produced, structured material designed to have optical properties not usually exhibited by naturally occurring matter. Metamaterials that are very thin have come to be known as ‘metasurfaces’. Numerous applications have been demonstrated using metasurfaces, including in flat optics, antennas, cloaking devices, holography and 3D imaging, sensing and imaging, and photonic integrated surfaces. In this project, the student will investigate lasers based on metasurfaces. The project will include theory, simulations, and experiments.

Contact supervisor

Resilient Communication Networks of the Future

Project Supervisor: Prof Tansu Alpcan
Offered: Semester 1 & Semester 2

Industries from manufacturing to transport and power grids, heavily rely on wired and wireless communication technologies as enablers. Future networks such as 6G will significantly depend on AI and machine (deep) learning (ML) methods for improved automation and efficiency. Specifically, machine and reinforcement learning methods will be used in 6G networks for tasks such as dynamic resource allocation, routing, and optical transmission parameter optimization.

Today’s communication and classical computing systems are vulnerable as evidenced by multiple malicious attacks successfully targeting critical infrastructures worldwide. Modern machine learning algorithms are vulnerable to intentionally designed adversarial data. Specifically, adversarial attacks can modify the input of deep neural networks during deployment and lead the algorithm to an opposite result. Soon, malicious adversaries will launch attacks on communication systems targeting directly the underlying machine learning algorithms, causing them to misclassify or distort their outputs. Hence, attacks on machine learning algorithms embedded into communication network infrastructure may disrupt power systems, ambulances, traffic lights, and factories. This may lead to loss of personal safety and costly economic outcomes.

This project aims to develop a scientific framework to reduce AI/ML risks in communication networks and develop autonomously self-configuring communication systems. The main research question that we will address is “How to develop novel self-configuring communication systems using communication and game-theoretic first principles to mitigate novel risks originating from AI/ML?”

The project will build upon interdisciplinary knowledge from a range of topics such as optimisation, game theory, information theory, and (adversarial) machine learning, in addition to basic knowledge of communication networks. It is a requirement for the interested student to have background knowledge in (at least some of) these area as well as good Python programming skills for simulation studies supplementing theoretical analysis.

Data-oriented Methods for Manufacturing and Industry 4.0

Project Supervisor: Prof Tansu Alpcan
Offered: Semester 1 & Semester 2

Industrial process and quality control has traditionally relied on statistics-based methods to monitor and improve manufacturing processes. Techniques such as statistical process control charts and process capability analysis have been foundational in detecting variations and ensuring high yields. These methods often require extensive domain knowledge and manual intervention, making them time-consuming and less adaptive to complex, modern manufacturing environments. As the manufacturing landscape evolved, there has been a shift from purely statistical-based methods to ML-integrated statistical approaches. Machine Learning (ML) technologies have the potential to revolutionise industrial manufacturing, significantly enhancing production efficiency and defect reduction.

This project aims to address the research question: “How can we harness the potential of ML to reduce manufacturing defects and improve yields, particularly when faced with the challenges posed by imperfect data in Industry 4.0?” The project will build upon the foundations of classical statistical process and quality control. However, it will focus on modern, data-oriented machine learning approaches ranging from classical ML to deep learning. The rich set of problems and datasets provided by multiple industry partners help scoping this project. Therefore, research aims are aligned with practical goals of industry partners such as defect detection, root-cause analysis, data mining for manufacturing, predictive maintenance, and zero-defect manufacturing, which have immediate and significant practical impact.

The interested students should have basic knowledge in probability/statistics and machine learning, and good Python programming skills. The project will emphasise engineering over theory and will heavily use machine/deep learning and data science methods. In addition, students should be willing to learn basics of manufacturing and industrial engineering.

Contact supervisor