Electrical and computer engineering internships

Supervisor: Dr Ian Abraham

Eligibility: 

    -    Candidates should have a strong background in planning, controls, software engineering,            systems integration, with experience in robotic perception.

    -    Experience with mobile robotic platforms and legged robotics is a plus as well as            experience with reinforcement learning.

Project Description:

Existing legged robotic systems are optimized on fixed frequency perception-action control loops. Consequently, computational resources are often wasted on simple locomotion strategies across regular and smooth surfaces while reacting too slowly in more complex, demanding environments. Rather than enforcing a uniform control frequency across all operating conditions, this project seeks to develops methods that allow robots to allocate computation where and when it matters most. This adaptive approach is expected to improve agility, stability, and energy efficiency while reducing unnecessary processing overhead. The research will combine learning and control strategies to compute locomotion policies and adaptive perception strategies to enable robust movement across changing terrains. Outcomes will include new algorithms for variable-rate decision making, validation in simulation and hardware experiments, and design principles for more responsive, resource-aware legged robots in search and rescue, inspection, and field robotics.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Dr Ian Abraham

Eligibility: 

    -    Candidates should have a strong background in software engineering, systems              integration, with experience in robotic perception, planning, and controls.

    -    Experience with mobile robotic platforms and legged robotics is a plus as well as            experience with reinforcement learning.

Project Description:

The advancement of robotic hardware has enabled vast opportunities in supporting ventures in remote and desolate environments. However, the built-in automation system (known as the autonomy stack) that orchestrates perception, decision-making, planning, and control, still requires a persistent human presence to adjust and tune to the specific environment. As a result, routine tasks like inspection, maintenance, and scientific exploration in extreme environments becomes challenging to execute reliably. To address the reliability gap, this project seeks to build a custom plug-and-play autonomy stack equipped with the latest tools in robust perception, planning, and control to orchestrate general robotic hardware in extreme environments, reducing human exposure to hazardous settings, and increasing operational efficiency. The primary use-cases will be in 1) maritime uncrewed underwater systems (UxS) mapping ocean floor near littoral regions and 2) legged robotic system inspection in remote field sites. Key performance will be measured through coverage, accuracy, and repeatability.

Requirement to be on campus: Yes *dependent on government’s health advice.

Supervisors: Prof Yonghui Li, Haonan Zhou, Gaoyang Pang

Eligibility: 

    -    Open to students in Mechatronics, Electrical, Computer or Software Engineering with           strong programming ability. Experience in robotics, control, ROS/ROS 2, embedded            systems, reinforcement learning, or state-machine design is desirable.

    -    Students must be comfortable with testing, debugging, and safe laboratory work            practices.

Project Description:

Humanoid robots require hierarchical control architectures to translate high-level goals into safe, interpretable, and stable whole-body behaviours. This project will investigate an Intent-Task-Skill-Action (ITSA) framework for humanoid control, with a simulation-first workflow and, when available, optional supervised hardware validation. The student will assist in implementing interfaces between ITSA layers, prototyping task decomposition and skill sequencing logic, and evaluating timing budgets, watchdog mechanisms, and recovery behaviour under perception dropouts. Applications include assistive humanoids, collaborative robots, and embodied AI systems operating in semi-structured environments. The internship will provide hands-on experience in robot software architecture, control integration, safety-aware experimentation, and quantitative evaluation, with mentoring for students interested in robotics research careers.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Prof Yonghui Li, Haonan Zhou, Gaoyang Pang

Eligibility: 

    -    Open to students in Mechatronics, Electrical, Computer or Software Engineering (or            related disciplines). Prior Python experience is required.

    -    Experience with ROS/ROS 2, SLAM, navigation, mobile robots, or sensor fusion is highly            desirable.

    -    Students should be willing to conduct structured testing in the lab and document            quantitative results.

Project Description:

Reliable robot navigation is increasingly important not as a standalone problem, but as part of task execution in dynamic human environments. This project will explore task-oriented navigation and dynamic obstacle avoidance for robots operating indoors, where a planned task trajectory must be updated online to safely avoid moving obstacles such as nearby pedestrians while maintaining progress toward the original goal. The student will assist in simulation-to-lab testing, ROS 2 integration, parameter tuning, route- and task-based evaluation, and quantitative analysis of failures such as delayed replanning, unsafe clearance, localisation drift, and perception degradation. Applications include service robotics, warehouse automation, assistive mobility, and humanoid robots performing real-world tasks in shared spaces. The internship will provide hands-on experience in autonomous systems, robot navigation software, structured experimentation, and reproducible evaluation methods for embodied robotics research.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Prof Yonghui Li, Haonan Zhou, Gaoyang Pang

Eligibility: 

    -    Open to students in Mechatronics, Electrical, Computer or Software Engineering (or            related fields) with an interest in robotics and AI.

    -    Python programming is required.

    -    Experience with ROS/ROS 2, computer vision, HRI, or motion planning is desirable but            not mandatory.

    -    Students with strong prototyping and evaluation skills are encouraged to apply.

Project Description:

Effective robot communication depends not only on spoken responses but also on body language that is timely, meaningful, and easy for users to interpret. This project will explore vision-language grounded body language generation, where high-level intent and scene context are mapped to expressive motion primitives such as greeting, acknowledging, polite interaction cues, emotional expression, and directional pointing. The student will assist in building a body-language library, implementing a motion selection and timing policy, and designing evaluation scenarios to measure clarity, appropriateness, and user perception. Applications include assistive robotics, social robots, and multimodal interfaces in shared human environments. The internship will provide hands-on experience in human-robot interaction, multimodal learning, prototype development, and user-centred evaluation, while building strong foundations for future research in embodied AI.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Prof Yonghui Li, Haonan Zhou, Gaoyang Pang

Eligibility:

Open to students in Computer, Software, Electrical or Mechatronic Engineering with strong Python programming skills. Experience with distributed systems, ROS/ROS 2, software architecture, backend services, or integration testing is desirable. Students should be comfortable working across multiple modules and producing clear technical documentation.

Project Description:

Multi-robot systems require coordination frameworks that allocate tasks, manage resource conflicts, and maintain traceable execution across heterogeneous platforms. This project will investigate a service-level orchestration layer that links perception, dialogue, gesture, humanoid control, navigation, and teleoperation fallback through a unified task representation and state-based workflow. The student will assist in implementing task queues, timeout/retry handling, interface logging, and integration tests in simulation, and will analyse task success, intervention rate, and recovery time from experiment logs. Applications include collaborative robotics, automated service workflows, and distributed embodied AI systems. The internship will provide hands-on experience in systems integration, robotics software architecture, observability, and quantitative evaluation, with strong preparation for honours or postgraduate research.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisors: Prof Branka Vucetic, Dr Chentao Yue, Gaoyang Pang

Eligibility: 

    -    Open to students in Electrical Engineering, Computer Engineering, Computer Science,           Mathematics, or related disciplines with an interest in information theory and machine            learning.

-        Python programming experience is desirable;

-        Familiarity with probability, linear algebra, optimisation, or basic deep learning will be            helpful.

Project Description:

Error control coding is a cornerstone of reliable wireless communication, where redundancy is introduced to detect and correct errors caused by noise and channel impairments. An analogous problem arises in AI systems, where model imperfections, uncertainty and distribution shift can lead to erroneous predictions or hallucinated outputs. This project will investigate whether ideas from classical channel coding can be adapted to improve the reliability of AI inference. The student will review relevant literature in coding theory and robust machine learning, formulate coding-inspired mechanisms for representing or verifying model outputs, and evaluate their effect on prediction accuracy and robustness in selected AI tasks. The project is expected to contribute to a new research direction at the interface of communications and artificial intelligence, with potential applications in trustworthy AI, large language models, and safety-critical decision-making systems. Students will gain experience in coding theory, AI modelling and experimental research.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Prof Glenn Platt

Eligibility: WAM>75 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application.

Project Description:

Automated sound classification is a growing function that tries to identify the source of a particular sound or noise. These techniques are used across a range of applications including automatic identification of animals in an ecosystem, noise compliance in industry, and public emergency recognition systems. The automated classifiers used in such systems are often based on open-source AI techniques.

Many AI based classifiers are developed using supervised training techniques that process high quality uncompressed audio recordings.

This project will investigate the impact on classification performance if the audio recording is compressed (such as mp3 based recording). This investigation will help inform whether existing classifiers may be reliably used with lower quality data types, or whether re-training on lower quality data is required.

The project will be conducted with an industry partner. The project is most suitable for software engineers or similar.

Requirement to be on campus: No

Supervisor: Prof Glenn Platt

Eligibility: WAM>75 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application.

Project Description:

Tracking of personnel entering and leaving commercial and industrial facilities is a critical safety and security function. Many facilities use manual or human-intensive processes for this function, but the latest image classification and edge-computing techniques should allow such functions to be automated. One challenge to the development of such systems using the latest in AI techniques is how to respect the privacy of humans walking past.

This project aims to design and build a privacy-respectful automated personnel tracking system. It will develop a hardware and software system that tracks personnel in and out of a room by using a camera and computerised image recognition techniques. The project is likely to take advantage of open source image classification/recognition libraries and will operate on low cost consumer grade computing and camera equipment.

The project is predominantly a software development project and so best suited to software or computer engineering student.

Requirement to be on campus: No

Supervisor: Prof Yonghui Li, Dr Chentao Yue

 Eligibility:

        -      Strong background in telecommunications engineering or computer science

        -       Proficiency in Python and machine learning frameworks.

Project Description:

Semantic communication is a next-generation wireless paradigm that prioritizes transmitting the fundamental "meaning" or importance of messages rather than raw bit sequences. A critical challenge in such systems is efficiently managing limited network resources. This project explores the use of Reinforcement Learning (RL) to dynamically allocate wireless resources, such as data rates, based on the specific semantic importance of the transmitted information. Operating under a fixed total resource constraint, the RL agent will learn optimal allocation policies to maximize overall system utility and communication efficiency. The intern will assist in formulating the RL environment, developing the resource allocation algorithms, and evaluating system performance through simulations. This project provides hands-on experience in telecommunications and computer science, allowing the student to work on a challenging research project with experienced researchers.

Requirement to be on campus: Yes*as per government’s health advice.

Supervisor: Dr Cuo (Charlie) Zhang

Eligibility: 

        -        WAM over 80.

        -        Solid skills of Matlab/Python programming and data analysis.

Project Description:

Large language models, such as generative pre-trained transformer (GPT), have been widely used in our daily life, and it is a great opportunity to apply a large language model agent for decision making on power system operation and control. This project will develop a model of linguistic stipulations, which contains context, question and answer, as an agent. Then, this project will design a method of presenting the power system operating conditions via this agent and interacting with large language models to improve the performance of decisions. This large language model assisted method is designed for targeted power system problems including volt/var control in an unbalanced distribution network and optimal power flow in a microgrid. 

Requirement to be on campus: No

Supervisors: Dr Ruigang Wang; Prof Daniel Quevedo

Eligibility: Control theory; Deep Learning; PyTorch

Project Description:

Training large language models (LLMs) requires enormous computational resources. This project aims to develop more efficient pretraining methods by leveraging ideas from optimal control theory.

We will first interpret the transformer architecture as a nonlinear dynamical system and reformulate pretraining as an open-loop optimal control problem. This perspective enables the use of powerful tools from control theory to analyze and improve the training dynamics of LLMs.

Building on this formulation, we will explore several approaches to improve training efficiency, including controllability and stability analysis of LLMs, moving horizon control strategies, and other control-inspired optimization techniques.

Finally, the proposed methods will be evaluated through large-scale experiments using a state-of-the-art NVIDIA GPU cluster. This project offers an opportunity to work at the intersection of control theory and modern AI, contributing to the development of more efficient and scalable training methods for LLMs.

Requirement to be on campus: Yes *as per government’s health advice.

Supervisors: A/Prof Steve Shu

Eligibility:

    -    Familiarity with machine learning models and optimization methods.

    -    Skills in mathematical modelling image processing, signal fusion strategies            and computational imaging. 

Project Description:

Modern ptychography enables high-resolution imaging by scanning coherent beams over samples and recording diffraction patterns. How- ever, the limited dynamic range of existing detectors leads to information loss—either through saturation at high intensities or insufficient signal at low ones. This bottleneck becomes critical in capturing fine structural details, especially under challenging illumination conditions.

This project investigates a novel approach to enhance ptychographic imaging by integrating multi-exposure image fusion (MEF) techniques. By collecting diffraction patterns at multiple exposure levels and combining them into high dynamic range (HDR) measurements, the method seeks to preserve rich intensity details across all regions of the diffraction space. 

Requirement to be on campus: Yes *as per government’s health advice.

Supervisors: A/Prof Zihuai Lin

Eligibility:

    -    Up to 3 students are required for this project.

    -    The students participating in this project should have good knowledge on wireless                    cellular networks, communication theory and signal processing. Matlab,  C++ and Python            programming skills are essential.

    -    Students with average marks above 75 are preferred.

Project Description:

The ambient backscatter communications (AmBC) technology used in IoT systems is a practical application of energy harvesting from ambient radio frequency (RF) signals. In a typical AmBC-empowered IoT system, a passive or semi-passive IoT device harvests RF energy from the ambient RF source (e.g., a cellular network base station (BS) or a Wi-Fi router) to power its circuits, and then the IoT device modulates its information bits onto a sinusoidal signal by intentionally changing its amplitude and/or phase. In this project, we will build an AmBC-empowered battery-free IoT device prototype.

Requirement to be on campus: Yes *as per government’s health advice.

Supervisors: : A/Prof Zihuai Lin, Prof Branka Vucetic

Eligibility: Up to 3 students are required for this project. The students participating in this project should have good knowledge on smart phone APP development. Programming skills are essential.

Project Description:

This project will invent, train, and deploy AI models that produce accurate, comprehensive matching feedback of a patient’s clinical trial outcomes—which we call Patient recruitment system for global medicine companies

    1.      Computer program development based on server system 

            Develop AI based computer software which can automatically output clinical trial             matching score. The java or .C# programming is necessary, priority is given to whom             familiar with server programming.

    2.      PC & APP UI development based on HTML5 

            Similar to the first project, now the platform is multi-platform crossed, specially  co-            development for user interface.

    3.      Development & deployment of LLM algorithm 

            Developing and deploying LLM generative AI to clinical trial recruitment system, priority             is given to whom familiar with NLP or LLM programming experience.

Requirement to be on campus: Yes *as per government’s health advice.

Supervisors: Prof Xiaoke Yi, Dr Liwei Li

Eligibility: Year 3/4/5 or Master students.  Students from electrical engineering, computer engineering, mechatronics, or telecommunications are encouraged to apply. No prior background in photonics is required.

Project Description:

Join us in exploring emerging applications of radio frequency (RF) photonics for future 6G and satellite communication systems. RF photonics combines optical and microwave technologies to process high-frequency signals with extremely large bandwidth and low loss, enabling new capabilities beyond conventional electronic systems. In this project, students will investigate the potential of RF photonics to overcome traditional system limitations, analyse and optimize system performance, and contribute to novel solutions to solve industrial challenges. Students will gain hands-on experience in RF systems, photonic technologies, and communication engineering while working alongside PhD researchers and industry collaborators. The project provides valuable exposure to next-generation wireless technologies that are shaping future communication networks.

Requirement to be on campus: Yes *dependent on government’s health advice.

Supervisor: Prof Xiaoke Yi and A/Prof Liwei Li

Eligibility: Year 3/4/5 or Master students.  Students from electrical engineering, computer engineering, mechatronics, or telecommunications are encouraged to apply. No prior background in photonics is required.

Project Description:

Artificial intelligence (AI) is rapidly transforming science, technology, and industry. However, conventional electronic processors are facing increasing challenges in speed and energy efficiency. We are now exploring new computing technologies that use light instead of electricity to perform key AI operations.

This project will investigate emerging light-based computing technologies that enable ultra-fast AI processing with low power consumption.

Students will explore simulation, data processing, or experimental testing of the new photonic AI chips developed at the University of Sydney (https://www.sydney.edu.au/news-opinion/news/2026/03/10/sydney-researchers-build-ultra-compact-ai-chip-operating-at-spee.html ), which demonstrate the potential of optical hardware for next-generation AI systems.

Students from electrical engineering, computer engineering, mechatronics, or software engineering are encouraged to apply. No prior background in photonics is required.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Prof Xiaoke Yi, A/Prof Luping Zhou, A/Prof Liwei Li

Eligibility: Year 3/4/5 or Master students. Electrical engineering, Computer science, Mechatronics, Computer engineering, Telecommunication, Software engineering

Project Description:

The state-of-the-art sensing technology is rapidly growing and will play a critical role in the near future. For instance, smart phones, which play a significant role in our daily life, have a fingerprint identity sensor that makes it easy for us to access the device, and they also use an ambient brightness sensor to adjust the display brightness, etc.

The project is to deliver the superior, advanced sensing platforms assisted by machine/deep learning to address the important challenges across a diverse range of applications in various fields, particularly in lab-on-chips, Internet of Things, broadband communications and biomedical applications. The internship project focuses on electrical circuits design and data processing as well as machine learning and software programming. The aim is to realize ultra-sensitive, high resolution and extreme-range sensing.

The intern will closely work with a research team including PhD students and postdoctoral research associates. Innovative signal processing and design in both hardware and software will be carried out during the project.

Requirement to be on campus: Yes *dependent on government’s health advice.

Supervisor: Dr Sid Assawaworrarit

Eligibility:WAM>75 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application.

Project Description:

In this project, the student will explore how silicon–silica core–shell nanoparticles generate colour through wavelength‑selective light scattering with minimal light absorption. These nanoparticles can be engineered so that their size determines which visible wavelengths they scatter strongly, producing vivid natural colours without absorbing heat. The student will use Mie‑theory calculations to model how changing particle radius and shell thickness shifts the scattering spectrum. They will then convert the simulated spectra into predicted colours using CIE colour‑space tools. The project will culminate in a “colour map” showing how different nanoparticle sizes produce different apparent colours and calculated heat absorption from solar radiation. This project is primarily computational and is ideal for students interested in photonics, optics, or materials design

Requirement to be on campus: No

Supervisors: Dr Sid Assawaworrarit

Eligibility: WAM>75 and Undergraduate candidates must have already completed at least 96 credit points towards their undergraduate degree at the time of application.

Project Description:

This project aims to design silicon-based devices that can control the absorption and emission of mid-infrared thermal radiation. Potential applications for this device include energy generation from waste heat, radiative cooling, and sensing.

In this project, you will use photonic computational tools to analyse the spectral absorption and emission characteristics of common silicon devices, such as p-n diodes and MOS capacitors, in the mid-infrared region and to analyse how the absorption and emission properties of these devices can be adjusted by an externally applied voltage. You will then incorporate photonic designs, such as gratings or photonic crystals, to shape the spectral profile of the radiation absorption/emission. Finally, you will incorporate manufacturability to your designs to ensure they are realistic for manufacturing.

Useful knowledge: Semiconductor devices, Light absorption calculation

Requirement to be on campus: No

Supervisors: Dr Wibowo Hardjawana 

Eligibility: 

    -    WAM >= 80.

    -    A strong background in wireless communication and a deep learning background equivalent to that covered in ELEC5508 Wireless Engineering are required.

`    -    Programming knowledge C/Python are required.

Project Description:

5G delivers high data rates, low latency, and massive connectivity. Open RAN (O-RAN) introduces open interfaces and software-driven functions for flexibility, with the Radio Intelligent Controller (RIC) enabling applications to optimise radio resources. This project explores large language models (LLMs) for intelligent resource allocation in O-RAN systems, applying natural-language network instructions. Students will first deploy and test a 5G system, then develop and evaluate LLM-based RIC applications to improve spectral or energy efficiency. Using tools such as FlexRIC, ns-O-RAN, 5G-LENA, Sionna RT, NS-3, students will gain hands-on experience in 5G deployment, RIC app development, and AI integration.

Requirement to be on campus: Yes *dependent on government’s health advice.

Supervisors: Dr Wibowo Hardjawana

Eligibility: 

    -    WAM >= 80.

    -    A strong background in wireless communication and a deep learning background            equivalent to that covered in ELEC5508 Wireless Engineering are required.

    -    Programming knowledge C/Python are required.

Project Description:

The project aims to develop a basic real-time 6G air‑interface reference design using Software-Defined Radio (SDR). The targeted air‑interface technologies are single-input-single-output (SISO) and orthogonal frequency-division multiplexing (OFDM). The student will implement the signal processing of the targeted technologies in hardware descriptive language (HDL) and import them into an field programmable gate array (FPGA) chip by utilising the following Matlab workflows,

- SISO-OFDM HDL: https://au.mathworks.com/help/wireless-hdl/ug/hdlofdmmatlabreferences.html)

- SDR-based SISO OFDM implementation: https://au.mathworks.com/help/wireless-testbench/ug/introduction-to-custom-ofdm-on-ni-usrp-radio.html

- SISO-OFDM transceiver example: https://au.mathworks.com/help/comm/ug/ofdm-transmitter-and-receiver.html

This is a challenging project suited for students interested in advanced wireless systems, FPGA prototyping, and AI‑based signal processing and moving towards in-depth study in wireless communications.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: Dr Shahadat Uddin

Eligibility Criteria: Students should have knowledge about testing, and ideally come from software engineering background.

Project Description:

Millions of people rely on disease‑risk models to guide early diagnosis and treatment; yet even the most accurate models can unintentionally disadvantage certain groups. This project invites you to help change that.

You will develop fairness‑aware ML models for disease risk prediction using real-world and publicly available health datasets. The project includes: (1) reviewing common fairness metrics (e.g., demographic parity, equalised odds) and bias sources in health data; (2) implementing baseline models such as logistic regression, random forest, and XGBoost; (3) evaluating discrimination, calibration, and subgroup fairness; (4) applying one bias‑mitigation approach and analysing its trade‑offs; and (5) adding explainability via SHAP along with a concise model card documenting transparency and risks.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisors: Prof Yonghui Li, Haonan Zhou, Gaoyang Pang    

Eligibility:

    -    Open to students in Electrical Engineering, Computer Engineering, Software            Engineering, Computer Science or related disciplines. 

    -    Strong Python skills are essential. Experience with speech processing, LLMs/VLMs, NLP,            or human-computer interaction is desirable.

Project Description:

Human-robot interaction in indoor environments requires robust speech understanding, grounded dialogue, and reliable intent inference under noise and ambiguity. This project will investigate a multimodal interaction pipeline that combines speech input, live captioning, grounded question answering, and intent classification using map/signage context and a curated knowledge base. The student will assist in data preparation, prompt/retrieval design, latency and accuracy evaluation, and error analysis under realistic acoustic conditions. Applications include interactive robots for education, healthcare, and public service environments where consistency and accessibility are important. The internship will provide hands-on experience in speech and language processing, multimodal AI systems, experimental evaluation, and research-style reporting, with mentoring for students considering honours or postgraduate study.

Requirement to be on campus: Human-robot interaction in indoor environments requires robust speech understanding, grounded dialogue, and reliable intent inference under noise and ambiguity. This project will investigate a multimodal interaction pipeline that combines speech input, live captioning, grounded question answering, and intent classification using map/signage context and a curated knowledge base. The student will assist in data preparation, prompt/retrieval design, latency and accuracy evaluation, and error analysis under realistic acoustic conditions. Applications include interactive robots for education, healthcare, and public service environments where consistency and accessibility are important. The internship will provide hands-on experience in speech and language processing, multimodal AI systems, experimental evaluation, and research-style reporting, with mentoring for students considering honours or postgraduate study.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: A/Prof. Steve Shu

Eligibility:

    -     Familiarity with deep learning frameworks (PyTorch orTensor- Flow).

    -      Basic foundation in computational imaging, signal processing, or microscopy

Project Description:

Electron ptychography has emerged as a powerful technique for achieving atomic-scale resolution by capturing coherent diffraction patterns during probe scanning. However, reconstructing high-quality phase images from these patterns is computationally intensive and

sensitive to noise, especially when dealing with large-scale datasets from advanced electron microscopes.  This project explores the application of deep learning to accelerate and improve phase retrieval in electron ptychography. Traditional iterative solvers, while accurate, are often too slow for real-time imaging. Deep neural networks offer a promising alternative by learning to map diffraction data directly to phase reconstructions, drastically reducing inference time.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: A/Prof. Steve Shu

Eligibility:

    -    Background in engineering, physics, computer science, or a related discipline

    -    Basic programming experience (Python preferred)

    -    Interest in computational imaging and optimisation algorithms

    -    Basic understanding of automatic differentiation or gradient-based optimisation           methods.

Project Description:

Ptychography is a computational imaging technique that reconstructs high-resolution object images from multiple diffraction measurements. Recently, automatic differentiation (AD) has been introduced into ptychographic reconstruction to enable flexible optimisation and model-based reconstruction. This project aims to investigate and improve the optimisation strategies used in AD-based ptychography algorithms. The student will explore learning-rate scheduling, convergence behaviour, and robustness under noise using simulated datasets. The project will involve implementing and modifying existing reconstruction algorithms, running numerical experiments, and evaluating reconstruction quality using standard metrics. 

Through this project, the student will gain hands-on experience in computational imaging, optimisation methods, and scientific programming using Python and  modern machine-learning frameworks. The outcomes will contribute to improved efficiency and stability of ADbased ptychographic reconstruction methods.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor: A/Prof Steve Shu

Eligibility:

    -    Background in physics, electrical engineering, mathematics, or related disciplines

    -    Basic programming experience (Python preferred

    -    Interest in computational imaging, microscopy, or scientific computing

Project Description:

Electron ptychography is a computational imaging technique that reconstructs specimen structure from diffraction patterns collected in scanning transmission electron microscopy (STEM). For thick specimens, multislice models are required to describe electron propagation through depth, but these models significantly increase computational complexity.  Differential Phase Contrast (DPC) imaging and center-of-mass (CoM) analysis of diffraction patterns provide rapid estimates of local phase gradients from 4D-STEM data. This project will

investigate whether DPC/CoM signals can provide useful structural priors for thick-sample ptychography.  The student will generate or use simulated 4D-STEM datasets, compute 1CoM/DPC maps, and compare these signals with projected phase gradients and structural features of the specimen. The project will evaluate whether such signals correlate with regions of strong scattering and could inform future adaptive modelling strategies in multi slice ptychographic reconstruction.

Requirement to be on campus: Yes *dependent on government’s health advice

Supervisor- Dr Ian Abraham

Eligibility:

Candidates should have a strong background in controls, optimization, software engineering, systems integration, with experience in robotic perception and planning. Experimental experience with mobile robotic platforms and legged robotics is a plus as well as experience with reinforcement learning.

Project Description:

This project explores how humanoid robots can learn ultra-responsive motor skills that enable them to act with speed, precision, and fluidity in the physical world. The focus of the project will be on three demanding capabilities: dexterous catching, precise throwing, and sustained cyclical object transport. These tasks are crucial for enabling robots to operate in general logistics and manufacturing settings. Robots must perceive and anticipate the motion of incoming objects, react within milliseconds, and coordinate their full bodies to catch, transfer, and reposition objects while preparing for the next interaction. The goal is to learn control policies that continuously perceive, predict, and act in a tightly coupled sensorimotor loop. The outcome will be a new class of agile robotic behaviours that are dynamic, expressive, and highly capable in real-world environments. Students are expected to contribute to robot learning, experimental evaluation, and system design for fast, adaptive manipulation.

Requirement to be on campus: Yes *dependent on government’s health advice