Electrical, Computer, and Software Engineering

Applications for 2023-2024 are now closed.

Using Board Games to Teach Children Computational Thinking

Supervisor

Craig Sutherland

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG060

Project

This project will look at how we can use board games to teach computational thinking to primary school-aged children. It will explore elements of board games to design a board game for primary school-aged children (i.e. five to ten years of age.) The summer research project will evaluate existing board games to identify components that involve computational thinking and look at combining them together. It will also look at programmes like CS Unplugged and tangible programming environments to form a theoretical foundation for a board game. Due to the timeframe the project will focus mainly on preparing a game that can be evaluated with children in 2024.
This project does not require any previous skills. Instead, an interest in board games and a desire to teach young children is needed. This will be a joint project with the school of Fine Arts - Design in the faculty of Creative Arts and Industries.

Smart farming: plant disease detection from early stages

Supervisor

Mahla Nejati

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG061

Project

With the increasing demand for smart farming, a plant's effective growth and productivity are essential. All over the world, farmers struggle to prevent various harm from bacteria or pathogens such as viruses, fungi, worms and insects. Finding the presence of disease earlier in a greenhouse is key to preventing future damage. In New Zealand, tomatoes have nearly doubled (99%) in price in 12 months from December 2020. The reason for this cost growth was less availability. From early this year, a disease called Pepino mosaic virus (PepMV) has been impacting New Zealand growers; this affects how quickly tomatoes grow and the yield in which the plants produce.

This project aims to design a system to detect plant diseases from earlier stages and inform farmers about them. This system will be integrated into a plant monitoring system and can be tested on in the greenhouses.

The student will develop a deep learning-based computer vision model to detect plant disease within a multispectral image dataset. This project will primarily be done using Python and/or C++.

Tracking System for Fish Tank Drones

Supervisor

Trevor Gee

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG062

Project

The Centre for Automation and Robotic Engineering Science (CARES) has purchased a fish tank to prototype underwater robotic systems. One of the tasks in converting this fish tank into a research environment is to develop an external tracking system. The goal is to set up a multi-camera rig capable of fusing input data to perform real-time position and orientation tracking of small underwater mobile robots.

In this project, you will dabble in machine learning for underwater object detection, underwater optics, C++ and ROS (Robot Operating System), data fusion, data visualization and experimentation with optimization algorithms.

Algorithms developed in the fish tank are envisaged to be ported across to larger underwater drones such as the Boxfish ROV (https://www.boxfish.nz/products/boxfish-rov/features/), with our end goal being to create systems that can support aquaculture automation. Aquaculture is a promising strategy to manage better our marine resources, which is strongly connected to making our environment more sustainable.

Three-Phase Grid Connected Inverter

Supervisor

Duleepa J Thrimawithana

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG063

Project

Grid connected inverters are used in applications such as bi-directional sources or renewwable energy systems to provide a bi-directional link between the grid and a DC-link. This project involves developing a 100kW three-phase grid connected inverter that is capable of bi-directional power transfer.
The student will work closely with a postdoctoral research fellow to develop firmware to implement a stationary reference frame controller for a three-phase grid connected converter. We will use the C2000 family dual-core MCU to implement the controller. The controller should be able to maintain a programmable DC-link voltage by transferring power to or drawing power from the utility grid. High power electronics is an interesting field and the student will learn aspects of power systems, electronics, instrumentation, embedded systems and plumbing!

Sign-language understanding using Reinforcement learning and DNN

Supervisor

Ho Seok AHN

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG064

Project

This project will understand English sign-language from hand motion, gesture, and facial emotion. There are a lot of research on recognizing alphabets and numbers, but not word and sentence itself, which is more realistic and useful in real-world. Do literature survey on conventional approach, and select the best one and develop it as the first version. Then improve the method as the second version and/or integrate them to understand nuance of sign language. You may use reinforcement learning and/or Deep Neural Network (DNN). This project may be combined with other related projects, such as social human robot interaction projects and healthcare robot projects. All necessary devices will be provided.

This project includes to:
• do literature survey on conventional approach
• find the best one for English,
• develop and test, and
• improve the method.

Project scope will be decided after the meeting with supervisors. You will bring to the role a passion for research and engineering, excellent computing skills (including a high level of programming ability), and a strong sense of responsibility. It will be better if you have an experience on vision processing and/or using DNN.

Guide and logistics robot system

Supervisor

Ho Seok AHN

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG065

Project

This project will develop a multi-robot system using a guide robot and a logistics robot, which are already developed by part 4 students. We have three robots (a receptionist robot EveR, a guide robot Silbot, and a delivery robot GoCart), which can communicate with building system to control doors and elevators. This project will improve the integrated system focusing on navigation and/or interaction. You will work with other researchers who have done this project.

This project includes to:
• develop a web based graphical user interface (GUI) SW,
• interface three robots,
• improve social human-robot interaction skills on the existing guide robot, such as distinguishing the person who is passing and following,
• improve navigation,
• integrate with Multi-human detection system
• run regular operation for confirming its robustness, and
• evaluate its performance.

Project scope will be decided after the meeting with supervisors. You will bring to the role a passion for research and engineering, excellent computing skills (including a high level of programming ability), and a strong sense of responsibility.

Plant monitoring system

Supervisor

Ho Seok AHN

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG066

Project

This project will develop a system to monitor the change of plant growing. 3D reconstructed model may be needed to measure the number and size of fruit and leaves. May need to train Deep Neural Network (DNN) models to detect different fruits, such as kiwi, apple, grape, strawberry, tomato, paprica. We have a well working detection model for kiwi and apple fruits, and can easily make different models for other fruits using DNN. All necessary devices will be provided.

This project includes to:
• do 3D reconstruction of plant,
• make different models to detect different fruit and leaves,
• find easy way to train models, and
• develop a dashboard for user.

Project scope will be decided after the meeting with supervisors. You will bring to the role a passion for research and engineering, excellent computing skills (including a high level of programming ability), and a strong sense of responsibility. It will be better if you have an experience on vision processing and/or using DNN.

Interactive chatbot system for social robot

Supervisor

Ho Seok AHN

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG067

Project

This project will develop an interactive chatbot system for social robot that talk with visitors at the university reception. We have a working version using DialogFlow, and will develop a better chatbot. You may use reinforcement learning and/or Deep Neural Network (DNN) if needed. So when the text is given, chatbot generates its reactive speech by considering history of conversation. This project is related to ongoing 5 years research project, SHRI, and will work with PhD and professional staff. All necessary devices will be provided.

This project includes to:
• do literature survey on interactive conversation of human,
• find the way to consider relation and history of conversation,
• develop a chatbot, and
• apply it to one of robot platform if possible.

Project scope will be decided after the meeting with supervisors. You will bring to the role a passion for research and engineering, excellent computing skills (including a high level of programming ability), and a strong sense of responsibility.

Reinforcement Learning: Learning by Novelty and Surprise

Supervisor

Henry Williams

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG068

Project

Under the guidance of the CARES Reinforcement Learning (RL) team, this project will seek to experiment with novel ideas for the development of improved RL approaches. Reinforcement Learning is a machine learning approach that seeks to enable an agent (e.g. a robot) to autonomously learn to operate based on its interaction with an environment without human input or control. 
In this project, we will explore how we can incorporate the ideas of novelty, surprise, curiosity, and dopamine into the learning process to encourage exploration and improve the learned behaviours of our agents. These approaches will be evaluated against the OpenAI gymnasium environments (https://gymnasium.farama.org/). Successful approaches will then be applied to real-world robotic platforms. 

This project will require strong programming skills, specifically in Python, and will require the students to frequently come into the robotics lab to work on the project. Prior experience with Pytorch would be beneficial but is not required, as we will teach you those as part of the project. The project forms a part of the SFTI Rangatahi Robotics project - (https://www.sftichallenge.govt.nz/news/rangatahi-mission-lab/) and will work closely with the existing team in the CARES (https://cares.blogs.auckland.ac.nz/) group working in this space. We will seek to publish this project's results at an international conference.

F1tenth: Learning to race autonomously like an F1 Driver!

Supervisor

Henry Williams

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG069

Project

The Formula SAE competition is introducing an autonomous vehicle category to the control of the Formula SAE cars. To meet this challenge, we are aiming to develop a fully autonomous control system that enables the car to learn to drive itself through Reinforcement Learning. Reinforcement Learning is a machine learning approach that seeks to enable a robot to learn to operate based on its interaction with the environment without human input or control. This project will seek to develop an RL-based control system for an F1Tenth race car, capable of learning to race both in simulation and on a real-world race track - F1Tenth.  

This project will require strong programming skills, specifically in Python, and will require the students to frequently come into the robotics lab to work on the project. This project is not suitable for remote development due to the requirement of working with and testing on the physical vehicle. Prior experience with Pytorch or ROS (1 or 2) would be beneficial but not required, as we will teach you those tools as part of the project.The project forms a part of the SFTI Rangatahi Robotics project, and will work closely with the existing team in the CARES group working in this space. We will seek to publish this project's results at an international conference.

Get hands-on with Reinforcement learning for dexterous robotic manipulation

Supervisor

Henry Williams

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG070

Project

This project will develop a Reinforcement Learning (RL) based control system for dexterous robotic manipulation. RL is a part of Artificial Intelligence (AI) which seeks to enable robots to learn to perform tasks without human interaction or guidance autonomously. Through RL, a robot learns to perform actions based on observations of its environment through its sensors and feedback in the form of “rewards” for taking good actions, aiming to maximise the overall reward it receives while carrying out a given task.

This work will extend work developed to control a low-cost robotic gripper by increasing the complexity of the tasks the robot is required to perform.

This project will require strong programming skills, specifically in python, and will require the students to frequently come into the robotics lab to work on the project. This project is unsuitable for remote development due to the requirement to work with and test the physical grippers. Prior experience with Pytorch or ROS (1 or 2) would be beneficial but not required, as we will teach you those tools as part of the project.

The project forms a part of the SFTI Rangatahi Robotics project and will work closely with the existing team in the CARES group working in this space. We will seek to publish this project's results at an international conference.

Give us a hand developing a dexterous robotic manipulator

Supervisor

Henry Williams

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG071

Project

This project will develop the third iteration of a low-cost robotic gripper to develop and explore Reinforcement Learning-based control systems. The current designs of the low-cost grippers will be extended to be capable of in-hand manipulation of objects and incorporate improved touch-based sensors throughout the gripper.

The project will require strong design skills, specifically in Fusion360, and a good understanding of mechatronic design. This project is unsuitable for remote development due to the requirement to work with and test the physical grippers. 

The project forms a part of the SFTI Rangatahi Robotics project and will work closely with the existing team in the CARES group working in this space. We will seek to publish this project's results at an international conference.

Under the Sea: Navigating the Depths with Robotic Learning!

Supervisor

Henry Williams

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG072

Project

The Centre for Automation and Robotic Engineering Science (CARES) has purchased a fish tank to prototype underwater robotic systems. This project will develop and explore using Reinforcement Learning (RL) to control a UAV under turbulent conditions in the fish tank. RL is a part of Artificial Intelligence (AI) which seeks to enable robots to learn to perform tasks without human interaction or guidance autonomously. Through RL, a robot learns to perform actions based on observations of its environment through its sensors and feedback in the form of “rewards” for taking good actions, aiming to maximise the overall reward it receives while carrying out a given task. This project will require assisting in implementing and constructing the UAV platform for autonomous control in the fish tank in addition to the algorithmic development. 

This project will require strong programming skills, specifically in python, and will require the students to frequently come into the robotics lab to work on the project. This project is unsuitable for remote development due to the requirement to work with and test the physical UAV platforms. Prior experience with Pytorch or ROS (1 or 2) would be beneficial but not required, as we will teach you those tools as part of the project.

The project forms a part of the SFTI Rangatahi Robotics project and will work closely with the existing team in the CARES group working in this space. We will seek to publish this project's results at an international conference.

High current buck converter using parallel crygenically cooled GaN transistors

Supervisor

Matthew Pearce

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG073

Project

Cryogenic power electronics and GaN HEMTs Transistors can help us improve electrical motor power density to electrify large scale transport, such as aircraft. This research work focuses on evaluating the performance of parallel GaN HEMTs, specifically understanding how the load current is shared between them when operated at room temperature and in liquid nitrogen (-200degC). The inverters in this work will transfer a power level of around 5 kW. Heavy current electronics is an interesting field and the student will learn aspects of power conversion, PCB design, embedded systems and making things realllly cold.

Superconducting Inductor for Electric Aviation

Supervisor

Duleepa J Thrimawithana

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG074

Project

Electric aviation requires light weight electronics. Magnetic components that are in power converters used to drive electric aircrafts are one of the heaviest components. I significant portion of this weight is due to the copper windings.

Superconductors may be used to replace the copper windings to reduce weigth of a traditional magnetic component. This project thus aims to experimentally investigate advantages and disadvantages of using a superconducting inductor in comparison to a tradition inductor with a copper winding. Student will be enaged in developing power electronics, FEM simulation, building inductors and finally conducting laboratory tests using state of the art equipment.  

Digital educational engineering

Supervisor

Nasser Giacaman

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG075

Project

Digital Educational Engineering (DEE) is about using an engineering approach to design, build, and evaluate a software-based solution that will address some education-based problem. This project will involve developing a software application or software tool, which can utilise a range of digital technologies. The particular project selected will be determined closer to the time, after meeting the allocated student, in order to carry out a project that they are technically confident in. To be successful in this project, you should be a strong programmer confident in using technologies such as HTML, CSS, JavaScript, React, Python, etc.  

Modern Visualisation of Parallel Computing Schedules

Supervisor

Oliver Sinnen

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG076

Project

Parallel computing is notoriously difficult as many additional challenges have to be overcome in comparison to sequential programming. To efficiently use a multiprocessor system, the program must be divided into subtasks, their dependences must be analysed and they must be mapped and scheduled to the processors. The scheduling of the tasks is a particularly crucial step for the efficient execution of a parallel program. The better the schedule, the faster the execution and hence the obtained speedup. Many algorithms exist for this difficult optimisation problem. However, to gain insights and to improve the schedules and in turn algorithms, a meaningful visualisation of the schedules are essential. With large numbers of tasks and processors it is difficult to gain insights and to find weakness without a powerful tool.

In this project you will design and implement a modern visualisation of schedules created by existing algorithms. The goal is that the user can interact, inspect and modify these schedules for various different models. The tool needs to be fast, elegant and have a modern appeal. The work will be done in Java and can leverage existing scheduling libraries developed in the Parallel and Reconfigurable Computing Lab (PARC).

Requirement: Very good knowledge of Java

Development of an Experimental Testbed for Studies on Protection and Fault Location of Hybrid AC/DC Grids: Student 1

Supervisors

Nirmal Nair

Abhisek Ukil

Tran The Hoang

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG077

Project

The rapid growth in new energy resources and diverse load types has led to increased utilization of DC power in electrical grids. While a complete transition to DC is a future goal, the use of a hybrid AC/DC grid is a more practical solution, applicable across various voltage levels from transmission to distribution and household networks. Ensuring the reliable operation of hybrid AC/DC grids requires a focus on protection and fault location, which are critical areas of study.

The objective of this project is to develop a hybrid DC/AC Experimental Testbed specifically designed for characterizing, verifying, and validating new technologies related to protection and fault location in hybrid AC/DC grids. Existing studies in this field have predominantly relied on simulations or PHIL/CHIL technologies for validation. Therefore, this project aims to establish an Experimental Testbed that allows for practical characterization and validation of protection and fault location technologies in hybrid AC/DC grids. The Experimental Testbed may include a Multiterminal DC system interconnected within an AC grid, or a Point-to-Point DC system interfaced with two AC grids. It is flexible in terms of voltage levels, accommodating studies at medium voltage (MV) or low voltage (LV) levels.

This project requires the involvement of at least two students, each with specific responsibilities:
Student 1:
• Formulate MTDC (Multiterminal DC) or P2P (Point-to-Point) DC links using the VSC converters as part of the existing Experimental Testbed.
• Conduct a comprehensive study on the configuration and implementation of these VSC (Voltage Source Converter) converters and their DSP controllers in the high-risk lab at the University of Auckland.
Experimental Testbed Formation and Validation:
Upon completion of their individual tasks, both students will collaborate to form the Experimental Testbed. The Experimental Testbed will then undergo validation through several test cases and verified with results obtained from real-time simulations in OPAL RT OP5700.
Specific requirements:
• BE(Hons) - Electrical and Electronic Engineering (EEE) student- Third (second Pro) or Fourth (Third Pro) year.
• Good knowledge of power system grids and power electronics
• Enthusiastic applicants (any nationality) that want to make a positive impact in the world and can work in a collaborative environment

Resources:
High power equipment in high-risk lab at the University of Auckland
PhD thesis: https://researchspace.auckland.ac.nz/handle/2292/61901
Journal paper: Fault Location Estimation in Voltage-Source-Converter-Based DC System: The - Location | IEEE Journals & Magazine | IEEE Xplore

Development of an Experimental Testbed for Studies on Protection and Fault Location of Hybrid AC/DC Grids: Student 2

Supervisors

Nirmal Nair

Abhisek Ukil

Tran The Hoang

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG078

Project

The rapid growth in new energy resources and diverse load types has led to increased utilization of DC power in electrical grids. While a complete transition to DC is a future goal, the use of a hybrid AC/DC grid is a more practical solution, applicable across various voltage levels from transmission to distribution and household networks. Ensuring the reliable operation of hybrid AC/DC grids requires a focus on protection and fault location, which are critical areas of study.

The objective of this project is to develop a hybrid DC/AC Experimental Testbed specifically designed for characterizing, verifying, and validating new technologies related to protection and fault location in hybrid AC/DC grids. Existing studies in this field have predominantly relied on simulations or PHIL/CHIL technologies for validation. Therefore, this project aims to establish an Experimental Testbed that allows for practical characterization and validation of protection and fault location technologies in hybrid AC/DC grids. The Experimental Testbed may include a Multiterminal DC system interconnected within an AC grid, or a Point-to-Point DC system interfaced with two AC grids. It is flexible in terms of voltage levels, accommodating studies at medium voltage (MV) or low voltage (LV) levels.

This project requires the involvement of at least two students, each with specific responsibilities:
Student 2
• Collaborate with Student 1 to work on VSC control, as well as other components of the Experimental Testbed such as AC sources, transformers, conductors, filters, and load banks.
• Document the specifications and configurations of the Experimental Testbed, ensuring comprehensive documentation of all aspects. Experimental Testbed Formation and Validation:
Upon completion of their individual tasks, both students will collaborate to form the Experimental Testbed. The Experimental Testbed will then undergo validation through several test cases and verified with results obtained from real-time simulations in OPAL RT OP5700.
Specific requirements:
• BE(Hons) - Electrical and Electronic Engineering (EEE) student- Third (second Pro) or Fourth (Third Pro) year.
• Good knowledge of power system grids and power electronics
• Enthusiastic applicants (any nationality) that want to make a positive impact in the world and can work in a collaborative environment

Resources:
High power equipment in high-risk lab at the University of Auckland
PhD thesis: https://researchspace.auckland.ac.nz/handle/2292/61901
Journal paper: Fault Location Estimation in Voltage-Source-Converter-Based DC System: The - Location | IEEE Journals & Magazine | IEEE Xplore

Development of experimental setup dedicated to investigations on Arc Faults in LVAC and LVDC grids- Student 1

Supervisors

Nirmal Nair

Abhisek Ukil

Tran The Hoang

Michael Gibson

Andre Cuppen

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG079

Project

The continuous expansion of new energy resources and the increasing variety of load types have resulted in the growing utilization of DC power in electrical grids. While a complete transition to DC is a long-term objective, the implementation of hybrid AC/DC grids presents a practical solution applicable across various voltage levels, from transmission to distribution and household networks. Ensuring the reliable operation of hybrid AC/DC grids necessitates a focus on protection and fault location, which are crucial areas of study. Among the various concerns, DC and AC arc faults at LV levels have garnered significant attention due to their potential to cause equipment failures, endanger residents' safety, and increase the risk of fire.
The objective of this project is to construct an experimental setup dedicated to studying LVAC/LVDC arc faults. This setup will serve as an effective tool for researching the characteristics of AC and DC arc faults at LV levels, thereby facilitating the improvement of fault detection and isolation methods. Ultimately, this research aims to prevent arc faults from causing more severe consequences for individuals, personnel, and equipment.

This project requires the involvement of two students, each assigned specific responsibilities:
Student 1:
• Develop an Arc generator following the UL-1699B standard to control the distance between electrodes and create arc faults. The arc generator will consist of step motor controllers, two electrodes (one fixed and one moveable).Both collborative tasks for student 1 and Student 2 are expected to be completed within the first four weeks of the project duration.

Benchmark Formation and Validation:
Once the individual tasks are finished, both students will collaborate to form the setup, which will then undergo validation through several test cases. The students will also apply recently published methodologies to analyse the features of the obtained voltage and current waveforms and compare them against the published results.

Specific requirements:
• Interests in power system grids and power electronics
• Basic skills in Data Processing.
• Enthusiastic applicants (any nationality) that want to make a positive impact in the world and can work in a collaborative environment Specific requirements:
• Interests in power system grids and power electronics
• Basic skills in Data Processing.
• Enthusiastic applicants (any nationality) that want to make a positive impact in the world and can work in a collaborative environment

Resources:
High power equipment in high-risk lab at the University of Auckland
W. Miao, Z. Wang, F. Wang, K. H. Lam and P. W. T. Pong, "Multicharacteristics Arc Model and Autocorrelation-Algorithm Based Arc Fault Detector for DC Microgrid," in IEEE Transactions on Industrial Electronics, vol. 70, no. 5, pp. 4875-4886, May 2023, doi: 10.1109/TIE.2022.3186351.
Q. Lu, Z. Ye, M. Su, Y. Li, Y. Sun and H. Huang, "A DC Series Arc Fault Detection Method Using Line Current and Supply Voltage," in IEEE Access, vol. 8, pp. 10134-10146, 2020, doi: 10.1109/ACCESS.2019.2963500.

Development of experimental setup dedicated to investigations on Arc Faults in LVAC and LVDC grids- Student 2

Supervisors

Nirmal Nair

Abhisek Ukil

Tran The Hoang

Michael Gibson

Andre Cuppen

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG080

Project

The continuous expansion of new energy resources and the increasing variety of load types have resulted in the growing utilization of DC power in electrical grids. While a complete transition to DC is a long-term objective, the implementation of hybrid AC/DC grids presents a practical solution applicable across various voltage levels, from transmission to distribution and household networks. Ensuring the reliable operation of hybrid AC/DC grids necessitates a focus on protection and fault location, which are crucial areas of study. Among the various concerns, DC and AC arc faults at LV levels have garnered significant attention due to their potential to cause equipment failures, endanger residents' safety, and increase the risk of fire.
The objective of this project is to construct an experimental setup dedicated to studying LVAC/LVDC arc faults. This setup will serve as an effective tool for researching the characteristics of AC and DC arc faults at LV levels, thereby facilitating the improvement of fault detection and isolation methods. Ultimately, this research aims to prevent arc faults from causing more severe consequences for individuals, personnel, and equipment.

This project requires the involvement of two students, each assigned specific responsibilities:

Student 2:
• Investigate other equipment necessary for the setup, including DC/AC power sources, various load banks, and measuring and recording units … Both collborative tasks for student 1 and Student 2 are expected to be completed within the first four weeks of the project duration.
Benchmark Formation and Validation:
Once the individual tasks are finished, both students will collaborate to form the setup, which will then undergo validation through several test cases. The students will also apply recently published methodologies to analyse the features of the obtained voltage and current waveforms and compare them against the published results.

Specific requirements:
• Interests in power system grids and power electronics
• Basic skills in Data Processing.
• Enthusiastic applicants (any nationality) that want to make a positive impact in the world and can work in a collaborative environment

Resources:
High power equipment in high-risk lab at the University of Auckland
W. Miao, Z. Wang, F. Wang, K. H. Lam and P. W. T. Pong, "Multicharacteristics Arc Model and Autocorrelation-Algorithm Based Arc Fault Detector for DC Microgrid," in IEEE Transactions on Industrial Electronics, vol. 70, no. 5, pp. 4875-4886, May 2023, doi: 10.1109/TIE.2022.3186351.
Q. Lu, Z. Ye, M. Su, Y. Li, Y. Sun and H. Huang, "A DC Series Arc Fault Detection Method Using Line Current and Supply Voltage," in IEEE Access, vol. 8, pp. 10134-10146, 2020, doi: 10.1109/ACCESS.2019.2963500.

Studying the Effectiveness and Limitations of ChatGPT in Automatically Generating and Improving Program Assertions

Supervisor

Valerio Terragni

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG081

Project

We are currently witnessing significant advancements in AI, particularly with the emergence of Large Language Models (LLMs) such as ChatGPT. The capabilities demonstrated by ChatGPT in tackling complex tasks have been remarkable. LLMs are trained on diverse textual data and predict subsequent text based on the input given, thus enabling it to generate human-like text. Most LLMs (including ChatGPT) are also trained with source code files, enabling the generation of source code.

ChatGPT is able to generate Java program assertions: executable boolean expressions that predicate on variable values at specific program points. Program assertions should pass (return true) for all correct executions and fail (return false) for all incorrect executions, assisting in bug detection.

This summer research project aims to investigate the effectiveness and limitations of ChatGPT in generating or improving program assertions. In particular, your task will be to design and conduct an empirical study to compare the performance of ChatGPT with one of my techniques for enhancing program assertions. This project not only offers an exciting study but is also topically relevant and important for studying ChatGPT's potential in automating software testing, something that the software engineering research community is currently very eager to understand. To be successful in this project you need to be proficient in Java.

Real-time Execution of Neural Network Models on the Edge

Supervisors

Nathan Allen

Partha Roop

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG082

Project

As the prevalence of neural networks, and AI in general, is constantly increasing so too is their use in performance-constrained edge devices. Applications running on these kinds of devices typically need to be efficient in their execution in order to minimise their power consumption while maintaining real-time behaviours. However, most neural network progress focusses on the desktop, making use of high-performance GPUs and increasingly complex models to achieve their goals.

This project will investigate tackling this problem in two ways. Firstly, by creating a mechanism to convert pre-trained networks into more-efficient implementations in languages such as C. And secondly, creating an approach to automatically reduce the complexity of a pre-trained model with minimal impact to its accuracy. The student should have some existing knowledge of neural networks and associated tools such as PyTorch.

Resources:
- Pearce, H., Yang, X., Roop, P. S., Katzef, M., & Strøm, T. B. (2020). Designing neural networks for real-time systems. IEEE Embedded Systems Letters, 13(3), 94-97.

Mobile OCR Engine for Accurate Reading and Verification of VIZ and MRZ on Travel Documents

Supervisors

Waleed Abdulla

Felix Marattukalam

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG083

Project

Introduction: VIZ and MRZ scanners play a crucial role in reading and extracting data from passports, ID cards, and other travel documents that comply with ICAO standards. The VIZ contains human-readable information, while the MRZ is machine-readable using OCR technology. This project aims to develop a machine learning-based OCR engine that can run on smartphones to accurately read and verify VIZ and MRZ data. Additionally, the project will focus on implementing a checksum algorithm for data validation and cross-checking with the VIZ.

Objectives:
1. Develop a machine learning-based OCR engine optimized for smartphone platforms.
2. Implement algorithms to accurately read and extract data from VIZ and MRZ on travel documents.
3. Incorporate a checksum algorithm to validate the integrity of MRZ data and cross-check it with VIZ.
4. Design a mobile application that utilizes the OCR engine for efficient VIZ and MRZ scanning and verification.

Methodology:
1. Research and gather a dataset of sample Machine Readable Travel Documents (MRTD) in accordance with ICAO Doc 9303.
2. Train a machine learning model using the dataset to recognize and extract relevant information from VIZ and MRZ zones.
3. Develop and optimize an OCR algorithm specifically tailored for smartphones to ensure accurate reading of VIZ and MRZ data.
4. Implement a checksum algorithm based on specific weights assigned to each data field within the MRZ to validate data integrity.
5. Design and create a user-friendly mobile application that utilizes the OCR engine and checksum algorithm for VIZ and MRZ scanning and verification.
6. Conduct extensive testing and refinement of the OCR engine and mobile application using a diverse range of MRTDs.

Expected Results:
1. Achieve the depicted objects following the proposed methodology.
2. You have the liberty to come up with a better methodology if you are keen to do so.
Conclusion: This project aims to develop a mobile OCR engine capable of accurately reading and verifying VIZ and MRZ data on travel documents. The proposed solution will provide a convenient and reliable method for individuals and organizations to process travel documents efficiently by incorporating machine learning, a checksum algorithm, and optimizing the OCR algorithm for smartphones.

Note: This project requires keen, proactive students working under minimal supervision

Leveraging ChatGPT for Automatically Rectifying Breaking Changes in Java Projects

Supervisor

Kelly Blincoe

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG084

Project

Software isn’t developed in isolation; it’s built upon libraries that simplify the development process. These libraries, much like any software project, are dynamic and evolve over time. Breaking changes occur when an update or modification to a library modify its behavior, potentially impacting any client projects that depend on it. The possibility of breaking changes often discourages developers from updating their libraries due to potential compatibility issues. This is because adapting the client project to a new library version can be a challenging and time-consuming task.

The goal of this project is to investigate the use of ChatGPT to automatically adapt client code to accommodate these breaking changes.

To be successful in this project you should be proficient in Java programming and experienced with the build automation tool Maven. The primary deliverables for this project will be (i) a research prototype capable of automatically fixing breaking changes in a Maven-managed Java project, utilising the power of ChatGPT, and (ii) a research report that discuss the effectiveness and limitation of ChsatGPT (and Large Language Models in general) to automatically adapt the client code to library breaking changes. Studying the capability of ChatGPT to perform automated code fixes is something that the software engineering research community is very eager to understand.

Investigating how gender stereotypes impact the perceptions of who should become an engineer

Supervisor

James Tizard

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG085

Project

The gender gap in engineering continues to be a significant issue, with both academia and industry being dominated by men. This topic continues to receive a growing amount of research attention, looking at both why women are less likely to go into engineering, and why women engineers may not be retained in industry.

This work aims to help understand how gender stereotypes impact the perceptions of who should become an engineer. We plan to follow the recent work from Cutrupi et al. [1], in which they asked school aged children (Primary and Intermediate) to draw a software engineer, in order to collect their perceptions and investigate whether gender stereotypes still persist. This work will include, helping to conduct the “draw an engineer” sessions, analysing the output of these sessions, and preparing a research paper.

[1] Cutrupi et al., “Draw a Software Engineer Test - An Investigation into Children’s Perceptions of Software Engineering Profession” at ICSE 2023

Exploring Deep Neural Networks for Wildfire Detection

Supervisors

Waleed Abdulla

Felix Marattukalam

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG086

Project

Introduction: In recent times, the world has been plagued by devastating wildfires, causing significant damage to landscapes, homes, and lives. The rapid spread of wildfires calls for innovative solutions to detect and respond to these incidents promptly. This project will pilot a deep neural network study to develop an efficient wildfire detection system.

Problem Statement: Wildfires' escalating scale and frequency pose a considerable threat to communities and ecosystems globally. The release of dense smoke affects air quality, impacting the health and safety of people residing both near and far from the affected areas. The project focuses on leveraging deep learning techniques to detect wildfire outbreaks swiftly, enabling timely response and mitigation measures.

Project Execution:
1. Data Collection and Analysis:
• Utilize datasets from NASA satellites to extract vegetation indices data and identify areas prone to wildfires.
• Visualize and analyze the data to gain insights into wildfire patterns and potential risk factors.
2. Dataset Preparation:
• Access an initial dataset comprising 1126 labeled images, classified into wildfire and non-wildfire categories.
• Split the dataset into training (60%) and validation (40%) subsets to facilitate model development and evaluation.
3. Deep Neural Network Development:
• Train a Convolutional Neural Network (CNN) model using the labeled image dataset.
• Explore and experiment with various CNN architectures to identify the most effective model for wildfire detection.
• Optimize the model's hyperparameters to enhance its accuracy and reliability.
4. Model Evaluation:
• Assess the performance of the trained CNN model using the validation dataset.
• Measure key metrics such as precision, recall, and F1 score to evaluate the model's effectiveness in detecting wildfires.
• Analyze and interpret the results to identify areas for improvement and further refinement.

Expected Outcomes:
1. A trained CNN model capable of detecting wildfire outbreaks in images.
2. Evaluation metrics and insights into the model's performance for further enhancement.
3. Documentation of the project's methodologies, findings, and recommendations for future research and development.

Conclusion: By harnessing the power of deep neural networks, this pilot study aims to contribute to the early detection and mitigation of wildfires. The project provides an opportunity for students to explore cutting-edge technologies in machine learning and make a significant impact in addressing the growing challenge of wildfires.

Qualifications for Students: We seek highly motivated and proactive students proficient in Python programming. The ideal candidate should possess strong problem-solving skills, be able to work with minimal supervision and demonstrate a keen interest in machine learning and image classification.

Developing a speech and language demo for high school students

Supervisor

Jesin James

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG087

Project

The University of Auckland organizes various outreach activities, particularly through initiatives like the women in engineering network, to engage high school students. These activities aim to provide students with opportunities to explore the practical applications of STEM in areas that interest them. One emerging field of interest is speech and language technology, driven by advancements in machine learning and large language models like ChatGPT.

However, the current development of speech and language technology predominantly focuses on well-resourced languages such as American English, leaving many other languages behind. This is where the linguistic knowledge of high school students can contribute to technology development. To foster this mindset among students, a project is being undertaken to create an outreach activity that can be developed in the native languages of students in Aotearoa New Zealand schools. By doing so, this project aims to highlight the importance of students' languages as a valuable asset in shaping technology development. Furthermore, it encourages students to see engineering as an exciting discipline that allows them to combine their language skills with technological advancements.

The project involves conducting a literature review to gather insights on existing activities for high school students, collaborating with the University of Auckland's outreach activities team, devising a detailed plan, developing the outreach activity itself, and creating a training manual for volunteers. Proficiency in coding, preferably in Python, is essential for successful participation in this project.

Seeing emotions: Developing and testing an emotion annotation tool

Supervisors

Jesin James

Felix Marattukalam

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG088

Project

Emotions are felt. But can they be detected and visually represented as well? Yes, there are two or three dimensional models that help us visualise our emotions. An example of such a model is the Russel's circumplex model of emotions that depicts emotions in a two-dimensional space of valence and arousal dimensions. Here, valence indicates how positive or negative an emotions is. And arousal indicates the if the emotion was expressed using high energy or low energy. Previous research at UoA (by Speech Research Group @ UoA https://speechresearch.auckland.ac.nz/) has developed speech emotion recognition algorithms and also some approaches to visualise emotions. Isn't that cool to be able to "see" the emotions we feel? Lets take that a bit further then!

We are all used to technology that speaks to us and recognises what we say, such as Siri, Alexa, Google Assistant, and similar human-computer interaction technology. With the demand for contactless interaction increasing globally, systems that can interact with users via speech are becoming more relevant than ever. Also, speech technology is trying to recognise the emotions in human speech to respond to human users effectively. This is where Speech Emotion Recognition (SER) in speech technology becomes used. But this technology development is often concentrated on a few well-studied languages (99 out of 4097 world languages) like American English and Mandarin Chinese. The languages spoken by users of speech technologies are not restricted to these well-studied languages. As a result, speech technology may not correctly recognise the emotions of speakers of less-studied languages. One example of a less-studied language is New Zealand English, where all the technology development for Aotearoa New Zealand must come from Aotearoa itself.

The aim of this research project is to integrate speech emotion recognition to approaches that visualise emotions. This means we will be able to not just recognise people’s emotions but also be able to display them in an intuitive manner. We will also aim to develop this specifically for New Zealnd English accent. The pre-requisites of this project are Python/Java, developing GUIs using JavaFX/PyQT, developing. It is also essential that a thourough reporting of the project is done as a research paper as this work has progressed quite a lot in the past few years. The project will involve
1. Conducting a systematic literature review of speech emotion recognition approaches and emotion annotation, visualisation techniques.
2. Using existing emotional speech databases to train a Speech Emotion Recognition model and fine-tune it to New Zealand English.
3. Design a GUI/Web interface to visualise and annotate emotions and evaluate its design.
4. Test the system with speech samples unseen speech recognition model.
5. Prepare a summary paper regarding the review and development

Enhancing IoT Systems through Ambient Energy Harvesting and Supercapacitor Energy Storage: Empowering Sustainable and Autonomous Operation

Supervisor

Dulsha Kularatna-Abeywardana

Discipline

Electrical, Computer, and Software Engineering

Project code: ENG089

Project

This project focuses on improving an existing prototype that harnesses and stores ambient energy, specifically radio frequency (RF) and thermal energy, using supercapacitors. The ambient environment offers a vast array of energy sources, yet only a fraction of it is effectively utilized for useful work. By utilizing supercapacitors, which offer simplified charging and discharging processes, we aim to overcome some of the challenges associated with energy storage. However, supercapacitors also present their own hurdles, such as low voltage ratings and linear discharge voltage characteristics. The project builds upon a base prototype that already incorporates RF and thermal energy harvesting, providing an opportunity to enhance its capabilities and expand its functionality.

Music & Artificial Intelligence

Supervisor

Dr Fabio Morreale

Discipline

Music

Project code: CAI020

Project

This project explores the use of AI for music composition and production. Possible topics include critical examination of existing commercial tools for music generation; development of AI-tool to co-create music; consultation with local communities to gauge perception around the benefits and risks of AI in music.

Please get in touch with Dr Morreale to discuss your idea before applying.

This project is suited to a Music, Design, Arts, Engineering, or Computer Science student.

Apply for this project in the Faculty of Creative Arts and Industries form.