Over the past few years, there has been a general increase in people’s stress and anxiety levels. Events like the COVID-19 lockdowns have caused an increase in people with chronic stress, while the shortage of medical workers has compounded the impact due as there is insufficient support for people. Inner PEACE is a mobile app that has been developed to help people self-manage anxiety or stress. It has been developed to provide support for marginalized communities that might be unable to justify the cost of other therapies and solutions.

In this project, you will enhance the Inner PEACE application by connecting it to a smartwatch or similar device to collect physiological measures of stress (like heart rate), and adding additional functionality based on the smartwatch data. You will then compare the enhanced version of the application to evaluate whether the benefits justify the additional costs (e.g., the smartwatch.)

This project is new for 2024. It is a collaboration with the School of Design in CAI.

Location: Radio Systems (405.736, Lab)

Software Engineering has moved from the domain of single programmers working on a codebase to teams developing together. The recent lockdowns and the rise of modern web-based tools have also changed how developers work together. Developers can now collaborate in real-time on a codebase, even on individual files simultaneously. But what effect does this have on their productivity and the final code they produce?

In this project, you will explore how people use online, collaborative code editors. You will run evaluation studies where you observe how people work together and make recommendations on how to improve collaborative code editors.

Location: Radio Systems (405.736, Lab)

Ethics is becoming increasingly important in society. But despite its importance, it can be a complex subject to teach. Students often have preconceived ideas and are not aware of the consequences of these views. Board games offer a fun, low-risk way to explore the consequences of our actions, they allow the players explore other viewpoints and see the results of their choices.

This project will look at extending and evaluating a board game for teaching ethics in a robo-ethics course at the university. We currently have two versions of a boardgame to encourage ethical discussions about robotics. You can choose the aspects of ethics you are interested in and extend the games to include them. Alternatively, the previous projects have suggested some new directions for future versions of the game that you could implement.

This project builds on a previous iteration of this from last year. It is a collaboration with the School of Design in CAI.

Location: Radio Systems (405.736, Lab)

Trying to understand a large, unknown computer codebase is a challenging task. Currently, developers use Integrated Development Environments (IDEs) to navigate through a codebase while trying to connect different parts of the system. While the tools in an IDE can help, they still rely on people being able to build a mental map of the code. Another approach is to use people's spatial abilities to help with this process. Potentially, codes editor could allow developers to position code in ways that help them see paths between different components.

In this project, you will extend a spatial code editor to evaluate new directions to help programmers understand code. As this project extends an existing project, you can either suggest your own ideas or use some of the suggestions from the previous iteration. You will then evaluate the system with programmers to test the effectiveness of the system and whether it helps programmers understand code better than a traditional IDE.

This project builds on a previous iteration of this from last year.

Location: Radio Systems (405.736, Lab)

The inverse problem of electrocardiography essentially requires one to predict what is happening on the heart surface of a patient (called EGMs), given their body surface potentials (called ECGs). If a good prediction can be made, then cardiologists can use these predictions to easily and efficiently plan heart surgeries.

We have developed mathematical models for predicting EGMs from ECGs. Vison transformer AI models, in particular Dino v2 (https://github.com/facebookresearch/dinov2) from Meta AI has been used to learn EGMs represented as heat maps from ECGs also represented as heat maps. We have got very good preliminary results.

The aim of the project is to extend the AI model to handle complex arrhythias (heart problems).

Location: Embedded Systems (405.760, Lab)

It is well known that sentiments can move the stock markets. Discovering the price point for a given stock (or index) using financial text sentiment is a fundamental problem in finance.

The current state of the art sentiment analysis in finance is based on the standard frequency of words in the text technique. However, this technique misses the semantic meaning of the sentence. For example, a tweet such as: apple gains over samsung! has a positive sentiment, because of bi-gram (gains over). Simlarly, samsung gains over apple! also has a positive sentiment. However, one is unable to distinguish, which company has a positive sentiment over other using the current technique. This is because the meaning/semantics of the sentence is unknown.

Generative AI techniques (e.g., ChatGPT) encode the sentiment of the sentence. The AIM of the project is to use generative AI engines to perform sentiment analysis of textual data (structured from news) and unstructured from tweets to build a semantic preserving sentiment stream.

Once successful, we would like to understand a number of fundamental factors about sentiment and their impact on market price. Such as: (1) how much positive sentiment affects the market price, (2) Is the affect transient of permanent, etc.

Location: Embedded Systems (405.760, Lab)

In the rapidly advancing field of satellite technology, the effective utilisation of High-Level Synthesis

(HLS) is critical. Our research seeks to investigate the feasibility and efficiency of employing HLS

techniques to target and optimise several small devices intended for use in satellites. This comes as

the demand for miniaturised, energy-efficient satellites rises, and the need to develop streamlined

and effective programming methodologies becomes paramount. Some key objectives include

evaluating the performance gains, resource utilisation, and power efficiency achieved through HLS

and exploring potential challenges and limitations in the process.

This is a pre-allocated project.

Students pre-allocated:

Franklin O'Sullivan: fosu562@aucklanduni.ac.nz

Dylan Chamberlain: dcha504@aucklanduni.ac.nz

Location: Embedded Systems (405.760, Lab)

The integration of Photovoltaic (PV) systems with Hybrid Energy Storage Systems (HESS) in Microgrid Systems is a critical area of research and development, particularly in the context of enhancing efficiency and reliability. Two significant control strategies used in this domain are Maximum Power Point Tracking (MPPT) for PV systems and Model Predictive Control (MPC) for HESS. MPPT algorithms are designed to maximize the energy harvested from PV panels. They continuously adjust the operating point of the PV system to ensure it operates at its maximum power point, despite changes in environmental conditions like irradiance and temperature. Various MPPT techniques exist, including Perturb and Observe (P&O), Incremental Conductance (Inc Cond), and more sophisticated methods like those using artificial intelligence. MPC in the context of HESS is used to optimize the charging and discharging of the storage system, which often includes a combination of batteries and supercapacitors. MPC predicts future power requirements and manages the energy flow within the HESS to meet these demands efficiently. It considers various factors like state of charge, load demands, and generation forecasts. The use of MPC in HESS helps in prolonging battery life, reducing operational costs, and improving the overall stability and efficiency of the microgrid.

In this project, the PV system will be managed using MPPT algorithms, while the HESS will be regulated through an MPC-based control strategy. The test-bed at Power Systems Lab at Dept of ECSE, UoA would be used to experimentally validate the concept, and extend it further.

Location: Power Systems (405.628, Lab)

The Formula SAE competition is introducing an autonomous vehicle category to the control of the Formula SAE cars. To meet this challenge,, we seek 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.

Last year, work started to develop a means of enabling a Formula SAE car to autonomously navigate through reinforcement learning-based control on a simulated F1Tenth race car. This project will extend this work to operate on a real-world F1Tenth (https://f1tenth.org/) race car to navigate a "race track" in the lab using additional sensing and control modalities.

This project will require strong programming skills, specifically in Python, and will require the students to come frequently 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 -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.

Location: Robotics (405.652, Lab)

This project is seeking to develop a Reinforcement Learning (RL) based control system for the grasping of objects. 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 maximize the overall reward it receives while carrying out a given task.

This project extends upon work completed last year that enabled a two-fingered and three-finger robotic gripper to manipulate objects through the state-of-the-art in RL - this work can be seen here:

1)https://www.youtube.com/watch?v=2QBPtrrYIrY

2)https://www.youtube.com/watch?v=xkack0nEHlM.

3)https://youtu.be/0kii1EJjOzw

This project will extend this work to a full Robotic hand (https://github.com/UoA-CARES/Gripper-Code) which will increase the complexity of the manipulation tasks the system can learn. Furthermore, we will apply novel RL techniques to all robotic grippers to explore further improvements to the learning system itself.

The physical gripper and experimental environment necessary for conducting these experiments have been constructed already - the teams will focus on developing and evaluating the novel learning algorithms.

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 -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 be seeking to publish the results of this project at an international conference.

Location: Robotics (405.652, Lab)

Pokémon Blue (and its counterpart, Pokémon Red) are excellent games to explore Reinforcement Learning. Reinforcement learning is a type of machine learning where an agent learns to make decisions by interacting with an environment. The agent receives feedback through rewards or penalties based on the actions it takes. The goal of reinforcement learning is for the agent to learn a policy, a set of rules or strategies, that maximizes the cumulative reward over time. The learning process involves the agent exploring different actions and exploiting its knowledge to make better decisions. Reinforcement learning is commonly used in scenarios where explicit programming is challenging, and the optimal strategy needs to be discovered through trial and error. It has applications in various fields, including robotics, game playing, finance, and natural language processing.

Modern RL agents can easily play Atari Games and even learn to carry out simple physical tasks on robotic platforms - however, these examples are highly constrained compared to learning to conduct complex tasks in real-world environments. Pokemon presents an interesting complex task to explore advanced RL techniques:

State Space Complexity: While the game is not overly complex compared to some modern video games, it still offers a reasonably rich state space. The Pokémon world includes various locations, Pokémon species, moves, and strategic interactions. This complexity allows for interesting challenges in learning optimal strategies in large and diverse environments.

Exploration-Exploitation Tradeoff: In Pokémon, there is a need for the player to balance exploration (trying new moves or strategies) and exploitation (using known effective moves). Reinforcement learning algorithms often deal with the exploration-exploitation tradeoff, making Pokémon a good environment for experimenting with such concepts. How do you enable an agent to quickly and efficiently explore a complex state space safely?

Reward Design: Reinforcement learning requires a well-defined reward structure to guide the learning process. In Pokémon, winning a battle is an obvious positive reward, while losing is a negative one. Crafting a reward system that encourages the learning of effective strategies can be an interesting aspect of reinforcement learning experimentation. But what is a reward system that enables an AI to catch them all or become the Pokemon master? How do you balance short-term goals (winning battles) vs. long-term goals of progressing the game?

Long Horizons: The player in Pokemon has to complete a long series of puzzles and quests to progress the story and complete the game - there isn't a clear indication of progress and in some cases the player has to back-track through the environment. How do you train the agent to learn how to solve the puzzles? how does the agent learn to teach their Pokemon HM moves to progress the story? Is there a means of doing this that doesn't rely on expert knowledge and can allow the agent to learn the context of these requirements without human input?

Action Space: How does an agent learn complex actions (e.g. training pokemon to level up, catching pokemon) from lower-level action sequences (button presses)? How can an RL agent learn a hierarchy of higher-order actions from simple actions?

This research will explore using Intrinsic Motivation theory to teach an AI agent to play Pokemon. Intrinsic motivation theory suggests that individuals are driven to engage in activities for the inherent satisfaction and enjoyment they derive from the activities themselves, rather than relying solely on external rewards or punishments. This theory posits that people have an innate desire to explore, learn, and master new skills, driven by their internal psychological needs for autonomy, competence, and relatedness. Intrinsic motivation is characterized by a genuine interest and personal enjoyment in the task at hand, fostering a sense of curiosity and self-determination. Researchers often highlight the importance of fostering intrinsic motivation as it is associated with sustained effort, creativity, and improved overall well-being.

Location: Control Systems (405.722, Lab)

With the rise of electrification, power electronics is a growing industry. Inverters, a class of power converters, are used in applications such as power conditioning and renewable generation to create an AC output from a DC source. An ideal inverter will create a purely AC output, but in practical systems a DC offset can exist. This DC component can lead to a circulating DC current, decrementing the system performance. As such, the measurement of the circulating DC current is important to help improve system performance.

Measuring the circulating DC current can be challenging, as the AC current can be orders of magnitude larger than the DC current. Consequently, methods proposed to-date are complex and costly to implement.

This project aims to research and then develop a low-cost method to accurately measure the DC circulating currents generated by a high-power inverter, similar to a one that may be employed in a renewable generation or power conditioning system. After conducting a detailed literature research, the project would investigate both electronic and electro-mechanical measurement techniques using multi-physics simulations. Finally, a proof of concept for a current probe that meets the requirements listed below, will be designed, built and characterized to confirm the viability of our proposed solution.

• Maximum AC Current Sense: 150 Arms
• Maximum DC Current Sense: 10 A
• DC Current Accuracy: ±1 A
• Isolation Rating: 22 kV

This project is in collaboration with ABB Napier.

Location: Green Electronics (405.712, Lab)

Creating and maintaining strong magnetic fields using high-temperature superconductors (HTS) is an important step in improving many technologies, such as HTS motors, NMR/MRI and nuclear fusion. Conventional power supplies are ill-suited to providing the high DC current required to magnetise high­temperature superconducting coils due to their lossy nature and a variety of other factors.

Flux pumps offer a promising method of providing and maintaining the magnetising current through wireless current injection. They can be divided into two categories as dynamo flux pumps and transformer-rectifier flux pumps. Transformer-rectified flux pumps offer the advantage of not having moving mechanical parts, but driving such a system is complicated, as it requires an efficient and controllable current source on the primary side. The rectification process can also be quite complex to ensure minimal losses are introduced.

The primary goal of this research project is to develop an efficient driver circuit that is able to source a controllable current to drive primary side of a transformer-rectifier flux pump. After conducting a thorough literature survey, a driver circuit will be developed and designed to meet the requirements listed below. A prototype will be constructed to confirm the viability of the proposed design experimentally. To simplify experimental validation, a solid-state rectifier will be used on the secondary.

Specifications:

Current - 3 A

Voltage - 40 V

Slew Rate - 150 A/s

Frequency - 5 Hz

Location: Green Electronics (405.712, Lab)

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

Location: HASEL (405.662, Lab)

This project aims to investigate the use of an LLM to help grade student assessments, specially open ended assessments, similar to the once used in design courses. The students will first investigate the use of ChatGPT, and compare this against other LLMs. Using the ChatGPT API, students will look at the best way to implement an automated assessment grader in Python.

Using the ELECTENG/COMPSYS 209 laboratory assessments, the viability of the automated grading tool developed will be investigated and improved.

Location: Radio Systems (405.736, Lab)

HVDC is an alternative technology to AC for transmission of power over long distances, and also commonly used for interconnections. It is a relatively new technology, first implemented in the 1950s. Protection technology for DC uses is not as mature as for AC technology, which negatively impacts its reliability and safety and increases costs.

This final year project aims to improve the reliability and security of HVDC transmission systems by developing and improving existing protection methods, through experimentation, simulations and research. To simulate a HVDC system, we plan to utilize the test bed at UoA's power system's lab and use a RT-LAB platform in OPAL-RT to connect a Simulink model to a physical DC system.

Location: Power Systems (405.628, Lab)

Previous research has indicated a gender bias in how contributors act, but is the reliability of the surrounding documentation also affected? We want to investigate whether there are biases in Open Source documentation that impede accuracy and compare it to AI summarised and generated documentation, to see whether AI could be utilised to overcome entry barriers to new open source contributors.

This also has the potential to be useful in the open source security space, to detect malicious code or packages.

Location: HASEL (405.662, Lab)

We aim to address the persistent challenges faced by students in effectively communicating their work experience and projects in crucial sections of CVs. Acknowledging the essential role CVs play in the initial screening rounds conducted by companies, we recognize that subpar work experience and project sections can serve as a substantial barrier for students, potentially resulting in missed opportunities.

To gain an understanding of what makes an effective CV, we will examine existing literature, academic papers, and online articles to grasp what contributes to a strong work experience and project sections. Additionally, we will gather insights from industry professionals and recruiters to identify trends, preferences, and key elements that stand out. This information will be used to create a matrix with key attributes such as measurable numbers, impact, and skills.

By leveraging this research, we aim to develop a platform that can offer tailored advice and tools aligning with current industry expectations, ensuring students develop CVs that resonate with employers. To educate students on effective CV crafting, we will utilize interactive tools and educational resources that we will develop through research in this area. The platform will also prompt users to write their work and project experiences on the platform before and after our provided resources, to achieve two purposes:

1. Assess the effectiveness of our resources.

2. Gather feedback on the student's written experiences.

Both of the above could be measured/achieved using an LLM trained for this purpose.

Ultimately, the project aims to empower students with the knowledge and tools to bridge the gap between academic achievements and professional success and secure the coveted interview invitations that can pave the way for their professional success.

Location: MDLS (405-569)

Different methods of energy delivery are become of interest as wireless power transfer technology applications grow. Laser Power Transfer (LPT) is a method of deliverying energy to a load with high directionality and range. One such application for LPT is to deliver energy to a drone in flight, reducing the battery size required.

This project aims to integrate the receiver module of a LPT system into a drone. The receiving photovoltaic cell must be carefully chosen and a power electronic converter will need to be integrated in a candidate drone to enable the system to receive power without significantly increase its weight.

The key research questions for this project includes:

1. What are the challenges in the integration of the charging module into a drone
2. Can LPT be used to deliver power to a drone

Location: Power Electronics (405.614, Lab)

This research initiative focuses on leveraging artificial intelligence to change the way students prepare for technical interviews in the software industry. Recognizing the need of algorithmic and data structure questions in hiring processes, our project aims to bridge the gap between theory taught in university courses and the practical skills demanded by technical interviews.

Adaptive Learning Paths:

The Al-driven system adapts to students' performance, ensuring that the learning paths are dynamically adjusted based on their strengths and weaknesses. This approach goes beyond traditional static content, providing an interactive and personalized learning experience.

Continuous Refinement:

The system is designed for continuous refinement, learning from each user interaction and updating its recommendations accordingly. This iterative process aims to enhance the system's accuracy in providing tailored challenges and explanations over time.

Defining relationships between topics:

The system will define relationships between certain topics that are intertwined and create recommendations based on that. Such as if a student is weak at the topic of 'Two Pointer' then he probably cannot do 'Binary Search', but if he is good at 'Binary Search' and 'Linked List' then he is probably good at 'Two Pointer'

Location: MDLS (405-569)

Typical graphics processing units (GPU’s) are optimized for massively parallelised algorithms implemented as shaders, such as the rendering pipeline, which act on sequential memory access, and leverage SIMD capabilities. Path Tracing is a computationally expensive rendering technique that renders scenes in a photorealistic and physically accurate way, by projecting light rays into the scene and using phenomena such as reflection and refraction to determine the color of each pixel. Due to the unpredictable divergence of the rays after the first cast, there is an inherent amount of divergence in both memory access and instruction in determining what each ray collides with which manifests as a being a very expensive problem to solve for. Recent advancements in modern hardware have made certain types of realtime raytracing possible in games, namely the RT cores offered by Nvidia. While this is a major advancement in the field of realtime raytracing, we believe that the traversal of certain scenes and collections of entities in offline path tracers where such tree reconstruction is seldom, could be made more efficient through further development of certain acceleration structures, whilst leveraging specialized hardware. This project will investigate different methods for eliminating potential intersection candidates, both in their construction and traversal, to determine the feasibility of using specialized hardware to further boost performance of offline path tracers.

Location: Embedded Systems (405.760, Lab)

With the introduction of wireless power into the consumer electronics in the past 10 years, smaller form factors for lower-power WPT systems are niche cases targeting alternative markets in the robotics and manufacturing industries. Potential use cases include products where wiring and electrical connections are difficult to maintain and where size and portability play a big role. The introduction of WPT into products like cardiac pacemakers/implants in the medical industry, lightweight drones, or mobile robots could enhance the functionality and performance of these products in their specific use case significantly.

WPT systems require large coil arrangements in order to facilitate resonance at low frequencies. This increases the weight and size of the transmitter and receiver required to resonate properly and transmit power efficiently with minimal leakage fields.

Increasing the frequency to the lower MHz range is a method of reducing size and weight because it reduces the wavelength, and thus, the length of copper required for a resonant topology. Higher frequency WPT allows the usage of smaller coupling inductors with equivalent power transferred, therefore decreasing the size of the charging pad respectively.

In addition, using a three-phase approach can reduce electromagnetic field emission, due to more directed fields.

With reduced size, weight, and field emission, new applications for WPT can be conceived.

An application of low EM field emission and small form factor wireless charging is in manufacturing or for use in small robots. Since these products often require very clean areas or places where wiring can be particularly difficult, effective WPT for recharging can facilitate manufacturing processes or make recharging of robots autonomous. In spite of the high path attenuation associated with HF fields, it is applicable where low transmission distance is required, such as mobile robots or the charging of drones.

Location: Power Electronics (405.614, Lab)

While electric vehicles have grown in popularity, their charging solutions take longer than conventional petrol vehicles. Wireless charging aims to reduce these barriers and encourage adoption of EVs among the general public by increasing the convenience of charging. A low-loss system compatible with the majority of vehicles would lead to this goal being met. New WPT systems are now being proposed that use multi-coil magnetics on the vehicle side to enable higher power, and power class interoperability. Sensing coils are also integrated into the vehicle magnetics for alignment.The inclusion of multiple coils provides an opportunity for new control approaches which can be considered in this project. Current bipolar pad designs can transfer power through one or both coils. As the desired power transfer level changes, it can be advantageous to operate only one of the coils. Doing so requires a rectifier with active elements on the vehicle side. A method of wirelessly synchronizing the secondary side is necessary to ensure that the power transfer can be controlled and maximise the efficiency of the system.This project aims to begin with a review of active bidirectional controllers for IPT charging. The focus of the research is to look and compare various means of deriving and controlling the synchronisation signal. These include transmitting a synchronisation signal from the grid-side to the vehicle-side of the system or using secondary side sensing coils (such as one of the power coils of the bipolar pad or other sense coils) to derive the synchronisation signal. Having made a decision on the best approach for the chosen vehicle system, the project will simulate, and build a prototype system that integrates the synchronisation signals into an existing controller to synchronise the frequency.

Location: Power Electronics (405.614, Lab)

This is a software engineering project that will be centred around the healthcare technology domain. It will involve the research of the effects and challenges faced by patients that are suffering from cancer as well as possibly other terminal illnesses, why those challenges occur, and what solutions can be used to address these challenges. In particular the challenges of transparent and easily accessible information will be researched. These challenges and solutions will be investigated from multiple perspectives. For example, one perspective can be the clinical point of view, i.e. what kinds of information clinicians and medical professionals can provide, and the means for communicating this information. Another perspective can be the patient's perspective, i.e. how the patients can access this information in an accessible manner, and how there can be some kind of two-way communication.

The solutions may be a combination of already proposed solutions in other research papers, as well as our own proposals based on our own research. We will then research how these solutions can be implemented via a technical approach, in particular using a web and/or mobile based software application. We will create a comprehensive requirements write up that will be elicited from our research. Subsequently we will produce a technical design for the application which will include User Interface (UI) designs, architectural diagrams, use case diagrams, and documentation explaining design principles.

Once the design is complete, we will build a prototype application on the basis of the design we have produced that is functionally usable and demonstrable to the extent that an external user or client would be able to gauge its intended workings and its potential value. We will then critically evaluate the prototype against our original requirements, how well it meets said requirements, and potential future improvements. We will outline how elements from our original research have been considered and implemented in our prototype, and a conclusion as to how well it satisfies the intentions behind the project as well as what further research may be required for future progress.

Location: MDLS (405-569)

Note: This is a cross-department project with Civil Engineering

Understanding the mobility patterns and transportation needs of rural Māori communities is critical for providing equitable and sustainable transportation solutions across Aotearoa New Zealand. However, these communities are often underrepresented in transportation research and planning. This cross-departmental Part IV project aims to fill this knowledge gap by analyzing the travel behavior and electric vehicle (EV) adoption potential of a rural Māori community. The project will involve two components - a mobile application for collecting travel data, and analysis of the data to model mobility patterns and EV feasibility. Two Part IV students from Electrical and Computer Systems Engineering will develop a mobile app to capture GPS traces, trip purpose, fuel usage, vehicle type, and other relevant data from volunteer participants in the community. Meanwhile, the student leading this project will collaborate with them to design the data collection protocol and variables. Once travel data has been gathered over a 2-3 month period, the student will use Python or another programming language to analyze the dataset. They will generate descriptive statistics on trip distances, purposes, timing, and fuel consumption overall and segmented by demographic factors like age, gender, employment status, and vehicle ownership. These insights will feed into modeling daily and weekly mobility patterns and estimating the share of trips feasible to be fulfilled by EVs based on trip distance and purpose. Advanced students may also explore time series forecasting of EV adoption based on mobility trends. Visualizations and interactive dashboards will be created to summarize key findings and allow further exploration of the travel data. Results will be framed to highlight mobility needs, gaps, and opportunities for EVs to improve transportation equity in the community. The student will write a report discussing the implications of the travel patterns analysis for rural Māori EV adoption and make concrete recommendations for transportation planning and policy aimed at this demographic.

Location: HASEL (405.662, Lab)

Children are often fascinated with robots and want to control (program) them. Currently, most programming environments are screen-based, requiring the use of either a mouse and keyboard or a touch screen. This project will investigate the use of physical coding blocks (tangibles) as an alternative approach to programming robots, and writing software programs.

The project will primarily focus on the creation of an embedded system to detect and interpret a chain of physical programming blocks, generate the intended software program, and display the associated output.

Location: Control Systems (405.722, Lab)

Introduction:

Over the past year, we have witnessed long and challenging battles against record-breaking wildfires worldwide. The wildfire spread over vast areas around the world. Some of these fires burned thousands of acres of land and destroyed hundreds of homes and buildings. Poor air quality developed due to the dense smoke, which raised health concerns for people near the fire and at distances hundreds of kilometers away. The proposed project develops a deep-learning neural network system to detect fire outbreaks quickly.

Objectives

- As a starting point, a dataset of satellite images is available from NASA and classified as wildfire and non-wildfire. The dataset includes 1126 labeled images. The dataset is divided into two subsets: training 60%, validation 20% and 20% testing

- Train a Convolutional Neural Network (CNN) model to predict whether an input picture indicates a wildfire. Different NN structures will be tested to find the best model.

- You may use the transfer learning technique to minimize the computational cost model complexity.

- The ultimate goal is to develop a system and scenario that can be used to detect, predict, and alert the authorities of wildfires.

Conclusion:

Through the utilization of deep learning techniques, this initial study aims to contribute to the early identification and mitigation of wildfires. The project presents an opportunity to explore cutting-edge machine-learning technologies while tackling the escalating challenge posed by wildfires.

Location: Control Systems (405.722, Lab)

There is immense interest in developing and deploying deep neural networks to solve complex problems in many challenging areas, such as speech recognition and image semantics extraction. Neural networks (NN) generally act as a black box that can classify signals end-to-end without knowing the essence of the processed signal. That is contrary to the statistical approaches; they need a profound understanding of the analysed signal to extract suitable features besides mastering the classical machine learning techniques to classify the desired outcomes. The NN requires training to discover the underlying structure of the data and is dependent on heuristics and intuitions to choose the needed hyperparameters. The recent insurgence of NN, after many years of recession, is due to the introduction of the Deep Convolutional Neural Network (DCNN) paradigm that involves many layers using new training techniques. Current models use a backpropagation process, which has limitations in expressing the embedded knowledge and slow processing and training speed. We will develop a new mathematical model verified experimentally for the DCNN, which does not use the backpropagation process. We are looking for Artificial Intelligence, Neural Networks, and Machine Learning in a successful applicant. Mastering Python Language. Has an excellent background in maths. Objectives This project is about introducing a new theoretically derived model for deep neural networks for processing big data. The main difference from the current deep neural networks is it does not have the time-consuming backpropagation stage. The semantics of the embedded knowledge within it can also be more approachable to interpret. The mathematical model should be supported by experiments on big data from different datasets to validate the theoretical results.

Location: Control Systems (405.722, Lab)

This project aims to investigate and design a sensor-less control technnique for 3-phase brushless DC  (BLDC) motors that operate under frequent mechanical impulse loading and near-stall conditions. In contrast to common operating conditions of most BLDC motor drives, these special conditions are unavaoidable in compact robotic applications, where high-energy mobile robots driven by BLDC motors operate in an unpredictable environment with impulse loading from sudden impacts and energy transfer between robots and near stalling consitions.  This project aims to develop a sensorless control technique that allows for 3-phase BLDC motor drive to respond quickly under these special impulse loading and stall/near-stall operating conditions.The project invloves proposing a sensorless control mechanism, and subsequent theoretical analysis, modelling, simulations and bulding a prototype to demonstare the applicablility of the proposed control technique.

Location: Power Electronics (405.614, Lab)

Music is an auditory experience that enriches experiences and the cultures we encounter, frequently making us feel certain emotions and, in some cases, physical sensations. There are existing products designed for individuals to feel the rhythm of a musical piece or haptic suits that mimic what a song would 'sound' like.However, there is the issue of timbre discrimination, which hinders the ability to differentiate between different instruments in a musical piece - an example being unable to determine the piano and violin as they play the same melody.We would therefore want to tackle this issue by applying signal processing techniques and look into how haptic feedback could give individuals the physical sensation of different instruments in order to add depth to understanding a musical piece multi-modally and possibly further enrich the musical experience. This would be an exploration of a proof of concept and has the scope to be applied to a variety of ways if successful, such as allowing individuals to understand sound better while making music to having a more immersive experience whilst listening.

Location: Control Systems (405.722, Lab)

Capsule networks (CapsNets) are a class of neural networks that address one of the main issues and deficiencies of CNNs. In CNNs, extracted features in initial layers pass through pooling layers where critical information, including spatial relationships, could be missed, resulting in incorrect classifications. CapsNets can better model hierarchical relationships and have shown great performance in many fields. CapsNets have a different architecture in their layers and hence, different computation complexity. This project investigates hardware accelerations for different layers of this class of neural networks for real-time applications.

Location: Embedded Systems (405.760, Lab)

Recent advances in Machine Learning (ML) has resulted in widespread deployment of ML solutions in several areas. ML solutions are usually characterised by their high computational complexity. Heterogeneous computing platforms are introduced to optimise the performance and energy efficiency. They consist of a range of processing units, including Central Processing Unit (CPU), Graphics Processing Unit, Tensor Processing Unit (TPU), and Field Programmable Gate Array (FPGA). Each of the processing units on heterogenous platforms are sourced by its own specific voltage rail. This research will investigate the impact of undervolting on the accuracy of ML solutions implemented on AMD Xilinx heterogeneous platforms (Zynq MPSoC, Versal ACAP).

Location: Embedded Systems (405.760, Lab)

At TRV Trading, we know that the financial market offers different opportunities to profit from trading according to the way it's behaving. For instance, when the market is very active (a lot of trading is going on), there are more opportunities to profit, when it's inactive, there are less. TRV uses trading models to execute trades in the financial market for profit, these trading models have parameters which determine how the trading is performed. The trading model parameters that cause the model to generate optimal profit in an active versus an inactive market are known to be different. Currently the setting of trading model parameters is performed manually by human traders. We want to have an automated system that autonomously and continuously adjusts trading model parameters to maximise profits in the current market conditions.

Location: Embedded Systems (405.760, Lab)

The interenet is now crucial to all our lives, and whilst the importance of digital inclusion is well recognised in New Zealand, the level of service people have access to is not equitable. The difference in service between rural and urban New Zealand is well documented, with the rural services proving lacking. But even within an urban environment is it equitable across the whole city? In a preliminary analysis done by SOFTENG 364 students in 2023 it was found that whilst most public libraries in Auckland had the ability to access to broadband, four did not (i.e. approximately 8% of Auckland public libaries do not have the ability to even access broadband). But the ability to access a service does not incidcate whether it has been acctually connected, nor does it indicate the quality of that service.

This project is going investigate the quality of internet services across Auckland via a series of experiments which will be designed to measure the service quality. The project will require going to sites all over Auckland to make the measurements.

Location: Control Systems (405.722, Lab)

Do you find it hard understanding your lecturer speak in class? Background noise and the acoustics of the room can make this task very difficult, especially if you have hearing loss or your first language is not the same as the one the lecturer is using.

Acoustics of rooms such as background noise and reverberation affect our hearing detrimentally by degrading the intelligibility of speech, making it harder to understand what speakers are saying. When the listener has hearing impairment or the language of the speech is not the listener’s first language (i.e. non-native listener), the detrimental effect will be much more severe, causing inequity in accessing crucial information between listeners. This inequity is particularly serious in education sector in the modern society where students with various backgrounds are learning together.

Studies have shown that speech intelligibility varies due to not only to the room, but also due to where the speaker and listener are placed relative to each other. Poor speech intelligibility can really impact on the quality of the education experience, and as the person is always struggling to process the speech it can also have a negative impact to a persons sense of belonging. This final year project will investigate the impact room acoustic has on speech intelligibility. The project aims to identify whether there is an inequity in learning environments for non-native listerners in the university through the collection of science evidence. The project will also investigate what recommendations should be made to the university to rectify the issue.

This project will be done in collaboration with Dr Yusuke Hioka, an expect on room acoustics.

Location: Control Systems (405.722, Lab)

Electromagnetic anechoic chambers are frequently used to experimentally determine the radiation characteristics of antennas.  In order to perform measurements of an antenna under test (AUT), a second antenna is required which must have a sufficiently broad beamwidth to evenly illuminate the AUT, and must operate over as wide a range of frequencies as possible.  One such example is the omnidirectional bicone antenna (which can be engineered to have a wide bandwidth) although an antenna with a somewhat narrower beamwidth would be more practical.  The best antenna for this application is likely to involve some sort of horn structure.

The aims of this project are to (i) identify appropriate structures for a wideband horn antenna (ii) design and model one or more candidate examples of these (using CST Microwave Studio); and (iii) fabricate and measure the perfomance of these.

Location: Radio Systems (405.736, Lab)

Waveguides for high-frequency signals have traditionally been realised using bulky three dimensional structures (e.g. coaxial cable, rectangular/circular waveguides supporting non-TEM modes, etc).  However, recently there has been some interest in developing planar waveguiding structures using metasurfaces.  Metasurfaces are two-dimensional equivalents of metamaterials, which are structures that contain a repeating microstructure - the elements of which are much smaller than a wavelength. Although the basic design of metasurfaces is well understood, high-performance designs (which have low loss and low dispersion) can be challenging to realise.

In 2017 a working prototype of a metasurface waveguide was sucessfully produced, and an appropriate waveguide launching structure was designed in 2021.  This project will be extended this year to consider additional bend geometries, improved launching structures and designs which minimise the overall waveguide dimensions by (i) identifying appropriate implementation strategies; (ii) designing, modelling and optimising these structures using CST Studio Suite/Microwave Studio; and (iii) fabricating the design(s) and measuring their performance.

Location: Radio Systems (405.736, Lab)

Rapid increase in using artificial intelligence (AI) in different applications requires considering high-level of accuracy and performance in addition to energy efficiency especially for edge devices as the main design constraints.

Machine learning models for AI are trained based on application-specific data and then deployed in target execution platform to perform useful predictions. Machine learning algorithms are computationally intensive so, deployment of these algorithms for embedded systems is very challenging as not only achieving the required system performance and prediction accuracy are essential, reducing energy consumption is also critical. The aim of this research is to investigate the computational complexity of machine learning algorithms, especially, deep learning algorithms for embedded applications (such as object recognition) and develop a reconfigurable hardware accelerator, which is implemented on FPGA to satisfy the required performance and accuracy while minimizing the energy consumption.

Location: Embedded Systems (405.760, Lab)

Computer vision systems are used for many applications such as transportation, healthcare, and manufacturing. Many computer vision applications use machine learning algorithms for object detection and tracking to achieve a higher level of accuracy required for the target applications. However, high computational complexities of these algorithms require optimisations at both hardware and software levels to trade-off the performance and energy efficiency, while satisfying the required design constraints especially for embedded systems (or the so-called edge devices). While using hardware may increase performance and energy efficiency, software provides more flexibility. General purpose processors are usually not efficient enough for such applications therefore, an application-specific instruction processor (ASIP) is used. ASIPs provide more flexibilities as they can be programmed based on the target application requirements. However, their instruction set architecture should be designed considering the processing requirements of target applications. The aim of this research is to investigate the basic computational complexity features of typical machine vision algorithms and extend a general-purpose instruction set architecture (such as open source RISC-V implementations or Nios V) with custom instructions to improve the performance and energy efficiency and develop the required application programming interface (API) for software development.

Location: Embedded Systems (405.760, Lab)

To satisfy different design constraints and employ the available processing power, computing systems for embedded applications are typically designed as Multi-Processor Systems in Chip (MPSoC). This provides more opportunities to investigate the design trade-offs and optimize the target execution platform based on the application needs.

The aim of this research is to investigate the efficiency and performance overhead for inter-processor communication for MPSoC implemented on an FPGA using soft-core processors and optimize the scheduling of different tasks of a multi-tasking application properly to improve the performance of the systems. Issues such as cache coherency and synchronization overhead will be investigated.

Location: Embedded Systems (405.760, Lab)

This project addresses the growing need for precision agriculture techniques that can provide detailed, real-time insights into crop conditions. The use of such multispectral cameras, capable of capturing a broad spectrum of light including wavelengths beyond human vision, allows for the detection of subtle changes in plant health that are often invisible to the naked eye. The potential to employ these cameras on UAV Drones for autonomous data collection in agricultural fields further underscores its potential in transforming traditional farming practices, in the pursuit of more efficient, sustainable, and productive agriculture.

This Project includes to:

- Data Collection: Drones capture high-resolution spectral images.

- Point Cloud Formation: Images are processed into a point cloud.

- 3D reconstruction: Experiment with current SOA(state of the art) techniques for 3D reconstruction and explore new techniques such as Gaussian Splattering.

- Machine Learning Analysis: ML analysis on data point clouds/3D reconstruction.

- Pattern Recognition: Identifying health indicators like disease, moisture, nutrient level, and yield estimation among others of interest.

- Actionable Insights: Providing farmers with guidance for informed crop management.

Location: Control Systems (405.722, Lab)

To reliably engage pallets without human intervention, an autonomous forklift should be able to detect the pallet pose and determine its state/load. Such modern computer vision problems are often approached using machine learning techniques, however collecting and labelling large data sets for model training is often difficult and costly. This Crown-sponsored multi-year project seeks to understand the benefits and limitations of synthetic data for training machine learning models for warehouse applications.

In previous Part IV projects, we have proven the feasibility of using synthetic data to train a model to detect pallets from photos and video, demonstrating the model running from a live camera feed.

Prior PTIV projects provide:

• Existing code that can generate a dataset of labelled synthetic pallet images from a simulated warehouse
• An automated benchmarking tool for training and tuning models
• An existing selection of approximately 350 photos of pallets in Crown’s test warehouse – partially manually labelled.

This year’s project seeks to build upon prior work to provide a pallet pose directly usable by an autonomous forklift, while also reporting information about the pallet’s state.

Location: Robotics (405.652, Lab)

This research initiative endeavors to construct a speaker diarization framework tailored for the medical field, specifically targeting patient-doctor interactions. Presently, doctors dedicate a significant portion of their time to transcribing medical notes onto Electronic Health Records (EHR), leading to exhaustion. While professional medical scribes are often employed, they pose challenges in terms of training and retention. Moreover, the delay in receiving finalized notes, spanning between 30 minutes to an hour, disrupts workflow efficiency. The objective of this project is to develop an end-to-end deep learning driven speaker diarization system to improve the overall efficiency and workflow in medical setting. This system will seamlessly process raw audio sourced from either a microphone of computer/laptop or any mobile device and produce formatted medical notes.

Location: HASEL (405.662, Lab)

The increase penetration of renewable energy and use of electronic load into the electric grid poses significant challenge regarding the power quality issues and energy management. These penetrations will increase harmonic distortion, frequent switching of load will result in flicker etc. This project intends to assess the power quality using machine learning techniques.

Location: Control Systems (405.722, Lab)

The proposed project will design and implement intelligent controllers for renewable energy systems with special emphasis on wind enderguy systems. The controller design will progeressively developed from sinmple PI controller to  machine learning based controllers. These controllers will be implemented in OPAL-RT. This impolementation part is crucial for success of the projec.

Location: Power Systems (405.628, Lab)

This project aims to address the unmet mental health needs of New Zealanders, worsened by the COVID-19 outbreak in 2020. A long-standing nominal recovery approach, despite its strong principles, has limited access to operationalized tools informed by evidence-based practice and lived experience of recovery. Sustaining the use of these tools has proven difficult. There is an urgent need for effective, accessible tools to improve mental health and well-being, especially for different ethnic groups. 'The Gift Box,' a tool available only in physical form in limited numbers, addresses this need (https://talkthatheals.org/the-gift-box.php). With a profound theoretical foundation, this toolbox could greatly benefit from modern IT solutions. Digitalization would make it widely and affordably accessible to all patients and health professionals. The recent advancement of AI technologies, such as the Large Language Model (LLM) and prompt-engineering techniques, allows for the extension of this digitalized solution to include an AI-based mental health and well-being expert system for collaborative assessment and recovery. This project's impact and significance will extend beyond the immediate beneficiaries to the entire population of New Zealand, making a meaningful difference in the country's mental health landscape.

Location: MDLS (405-569)

Daylight control, a practice that has been in existence for several decades, plays a crucial role in mitigating energy consumption and finds widespread application in the construction of eco-friendly buildings. Despite its inherent benefits, it is essential to acknowledge that implementing daylight control incurs certain expenditures and may potentially have repercussions on the overall comfort experienced by occupants within a given space.

In the ongoing project, our primary objective is to comprehensively evaluate the merits associated with the utilization of various lighting control systems. Within this context, there exists a notable contention suggesting that the potential cost savings derived from daylight utilization might not be significant enough to justify the expenses incurred in installing and maintaining daylight sensors.

A significant aspect of our research focuses on the nuanced evaluation of comfort. The intricate relationship between daylight control and its impact on the comfort of occupants remains enigmatic, necessitating a meticulous and thorough assessment to unravel these complexities. The ultimate goal is to provide well-founded recommendations based on the insights gained from this in-depth evaluation.

It is worth noting that the predominant tool employed for conducting the extensive research outlined above is the advanced lighting design software Dialux. By leveraging the capabilities of this sophisticated software, we aim to enhance the precision and comprehensiveness of our investigations into the multifaceted aspects of daylight control and its implications on energy consumption and occupants' well-being.

Location: Radio Systems (405.736, Lab)

In this project, our primary objective is to conduct an in-depth study of emergency lighting systems tailored for installation in low-visibility spaces. The successful execution of this project will necessitate extensive literature review, DIALux simulations, and rigorous engineering evaluation efforts. These critical components are essential for comprehending existing knowledge, leveraging advanced simulations, and conducting thorough engineering analyses to enhance our understanding of emergency lighting in environments with limited visibility.

As part of our investigative approach, there will be instances where the creation of photometric files becomes imperative. These files will play a crucial role in validating and confirming the viability of novel concepts and ideas within the realm of emergency lighting design. Through the utilization of photometric files, we aim to ensure the practicality and effectiveness of innovative approaches in low-visibility scenarios.

Additionally, the project acknowledges the potential need for prototyping in certain cases. Prototyping offers a hands-on methodology for testing and refining emergent ideas, contributing to the practical implementation and improvement of emergency lighting systems. This consideration reflects our commitment to exploring diverse avenues and methodologies to enhance the overall effectiveness and reliability of emergency lighting solutions.

Location: Radio Systems (405.736, Lab)

Communication via light has been established for a considerable duration, primarily utilized in conjunction with sensors, remote control systems, headphones, and similar devices. In this project, our focus will be on exploring methodologies for transmitting audio stereo signals using light as the medium of communication. It is important to note that this project is hands-on in nature, involving activities such as designing, simulating with LTSpice, and prototyping.

The successful completion of this project will necessitate a comprehensive performance evaluation report. This report will serve as a critical component, detailing the efficacy and efficiency of the developed system for transmitting audio stereo signals through light. The hands-on approach, involving design, simulation, and prototyping, ensures a practical and tangible outcome, contributing to the advancement of communication technologies in an innovative manner.

Location: Radio Systems (405.736, Lab)

This project addresses New Zealand's escalating mental health needs. The current approach to mental health support, while principled, lacks practical, evidence-based tools that accommodate the diverse recovery experiences of individuals. There's a pressing need for accessible, effective solutions, especially among various ethnic groups. We propose the development of an AI-enabled chatbot for mental and wellbeing care. This system will facilitate real-time conversations between users and the AI-based expert system. It will comprehend user queries, provide pertinent information, and offer the necessary support. The system will employ current ChatGPT-based technologies and a practical mental health conversational standard. Criteria will be defined and evaluations conducted to measure the effectiveness and reliability of the tool.

Location: MDLS (405-569)

Aged care has become a popular domain for computerized applications. Similar to hospitality systems, AI-enabled assistants have been deployed in rest homes and elderly care facilities to provide a smart environment of personalized care. These smart systems not only dramatically reduce the physical workload of caregivers but also offer artificial intelligence-driven precision in healthcare. This project extends an existing implementation of a senior citizen companion application into a smart AI assistant. The aim is to develop a software solution with AI support for smart home care. Activity-related information will be collected and analyzed to determine patient risk factors. The study has the potential to provide clinicians and researchers with insights into rest home patients and their care management, thereby improving overall healthcare practices.

Location: MDLS (405-569)

Dementia is a complex and progressive neurodegenerative condition. People with dementia often like to walk around their home or leave the house to walk around the neighborhood, sometimes referred to as wandering behavior. Wandering behavior can result in the person with dementia getting lost or encountering dangers such as falls. This concern drives our project to develop an innovative indoor tracking system. It aims to reduce the risks associated with wandering behavior and to enhance safety, freedom, and dignity among people living with dementia. This project extends an existing implementation that utilizes Ultra-Wideband (UWB) technology and is aimed at enhancing the risk management functionalities. Evaluation will be conducted to measure the effectiveness of the solution.

Location: MDLS (405-569)

Setting up the digital platforms, including the software architecture and databases is a manualprocess, which is both time consuming and error prone. Also, ensuring that the ethics approvedprotocol is adhered to has to be ensured using testing. Can this be mitigated?We take our inspiration from two different fields. First, using advances in natural language processing (NLP),especially through large language models (LLMs), we seek to convert the description of the clinical trialprotocol into a machine readable format. During this step, we will try to group clinical trials into specific categories, such as randomised controlled trials (RCTs) and observational studies etc.During the next phase, we will explore automation of the software architecture using the synchronousapproach. Specifically, we will explore the recently developed HipHop.js language, which capturesthe reactive aspects, including concurrency, preemption effectively. This approach, will thus enable effective automation of clinical trials. We will explore a few clinical trials such as the "chemotherapy related cognitive decline" trial at Tata Memorial Hospital.

Location: Control Systems (405.722, Lab)

We have received funding from Google to collaborate on a new protocol for data centres called bittide. Bittide is a novel architecture for distributed systems, proposed by Google. Bitide is a physical layer ptotocol, which achieves "logical synchrony", without the need for actual clock synchronisation protocols. Logical synchrony ensures that buffers between communicating nodes are always bounded and also the communication delay is logically bounded.

In this project, we will develop an emulation platform of a network of bitide nodes. We will use a set of FPGA boards, which will extend the RISC-V architecture for prototyping a bittide-like system of distributed FPGAs that ensure determinitic execution of applications. Some sample applications for consideration include synchronous neural networks, microsimulation of vehicle plattons etc.

References:

Lall S, Caşcaval C, Izzard M, Spalink T. Modeling and Control of bittide Synchronization. In2022 American Control Conference (ACC) 2022 Jun 8 (pp. 5185-5192). IEEE.

Roop PS, Pearce H, Monadjem K. Synchronous neural networks for cyber-physical systems. In2018 16th ACM/IEEE International Conference on Formal Methods and Models for System Design (MEMOCODE) 2018 Oct 15 (pp. 1-10). IEEE.

Location: Control Systems (405.722, Lab)

Wearable smart watches are increasingly being used for the detection of different health conditions by using machine-learning models to make inferences. We are, in particular, interested in the UoA developed MoodAI platform, which is developed for real-time mood state classification, based on the data obtained from Fitbit smart watches. The current platform, while being able to perform mood state classification using deep learning models, has some limitations. First, it only deals with data from Fitbit devices. Second, the ML models are black-box in nature and hence not explainable.

Wearable devices from many different vendors have their APIs and hence accessing wearable data from a range of wearable smart watches remains challenging. In this project, we seek to develop an unified interface that will enable data capture and processing from many different smart watches.  More importantly, we will develop explainable deep ML models, which not only detect the onset of depression but also identify the key reasons for such a diagnosis, on an individualised basis. This project will be jointly supervised by Dr. Fred Sundram, who is a clinical psychiatrist and is based at FMHS.

Location: Control Systems (405.722, Lab)

The software industry lacks diversity. Recent research has theorised that software development tools can have inbuilt biases, making them easier to use for people who think in a certain way. For example, GitHub, a commonly used software tool, has been found to have missing information, unclear terminology, and unexplained processes. This can make GitHub non-intuitive for people with process-oriented learning styles as they prefer to have complete information before starting a task.

This project will select a common software development tool or practice and investigate potential inbuilt biases in-depth by designing a lab experiment.

Location: HASEL (405.662, Lab)

The project is an exploration into the integration of blockchain technology with the service

delivery of Mobile Network Operators (MNOs) and Network providers struggling with the data roaming problem. It

focuses on the development of a consensus engine and a semantic engine to manage roaming contracts

between MNOs efficiently where we will compare the scalability and performance of different

consensus engines. The consensus engine is a part of a distributed ledger system that ensures

mutual agreement on roaming contracts without the need for central authority. Common consensus

engines include Proof of Stake (PoS) and Proof of Work (PoW). The semantic engine interprets and

enforces the logic and rules encapsulated within these contracts during their lifecycle without

modifying them.

Location: Radio Systems (405.736, Lab)

This project will develop a system to monitor the changes of plant growing. We have a working version that detects target, which is tomato, and measures its size and volume. May need to train Deep Neural Network (DNN) models to improve the accuracy of existing methods, detect disease and measure its ripeness. We will finally estimate their harvesting timing and let farmers know. All necessary devices will be provided.

This project includes to:

• start from the current working version including detection and size measurement,
• improve the performance if needed,
• detect disease,
• measure the ripeness, and
• estimate the harvesting timing.

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.

Location: Robotics (405.652, Lab)

As humans increasingly relocate to cities and urban environments become denser, an opportunity emerges to grow plants in areas that receive a good amount of natural light but are not easily accessible by humans. The sides of large buildings, rooftops, or other hard to reach places could be planted with decorative or horticultural plants and be maintained by a fleet of autonomous drones. In this way, cities can be greenified, food can be produced locally, and air pollution can be mitigated.

This project involves exploration and validation of the underlying technologies to evaluate the feasibility of the proposed concept. The technologies at play are Inductive Power Transfer {IPT), Computer Vision (CV), and a UAV control system.

Why these technologies?

Inductive Power Transfer (IPT) - Provides power to the UAV, reduces reliance on batteries, reduces the weight of the UAV and therefore reduces power consumption, allows for non-stop flight, and reduces demand for rare earth materials.

Computer vision (CV) - Enables the UAV to avoid obstacles, informs the UAV positioning system, monitors and evaluates plant health.

UAV control system - Fuses data from the CV, IPT systems and other sensors and data streams and produces adaptable real world positioning data

Location: Power Electronics (405.614, Lab)

This project aims to develop a sophisticated web-based application for the management of various devices, particularly focusing on robotics and IoT (Internet of Things) devices. The core functionality of this application encompasses streaming video data, sensor data management, and deploying software or AI models to the connected devices.Key Features

1. Device Management: The application allows for the management of individual devices which can operate standalone or be interconnected with other robots or devices.
2. Data Streaming: Users can stream video data from the devices. This is particularly useful for real-time monitoring of robotic operations or for IoT scenarios where visual data is crucial.
3. Sensor Data Handling: The application can handle data from various sensors attached to the devices. This feature is essential for monitoring environmental conditions, device performance, and other critical metrics in real-time.
4. Software Deployment: Users have the capability to deploy software remotely to the devices. This includes updates, patches, or entirely new software packages, ensuring the devices are always up to date with the latest technology.
5. AI Model Integration: The application supports deploying AI models to devices. This is particularly important in scenarios where devices need to make autonomous decisions based on machine learning algorithms.
6. User-Friendly Interface: The web interface is designed to be intuitive and user-friendly, catering to both technical and non-technical users. It allows for easy navigation and control of the connected devices.
7. Security and Authentication: Given the sensitivity of controlling physical devices, especially robots, the application incorporates robust security measures and authentication protocols to prevent unauthorized access.

Scalability and Compatibility: The application is built to scale, accommodating a growing number of devices and varying types of robotics and IoT technologies.

Location: Robotics (405.652, Lab)

Generating dynamic facial expressions for head robots represents a cutting-edge intersection of robotics, artificial intelligence, and human-computer interaction. Traditional methods in facial expression synthesis for robots are often limited by predefined expression ranges or lack the ability to accurately mirror complex human emotions. This project invites students to explore and develop more sophisticated, adaptive techniques for creating facial expressions on head robots, aiming to bridge the gap between robotic capabilities and the nuanced expression of human emotions.

This student project is centered around the development of innovative approaches to enable head robots to generate facial expressions that accurately convey a broad spectrum of emotions with adaptive style control. Leveraging the principles of large-scale contrastive language-image pre-training models, students will investigate how to effectively interpret diverse inputs such as textual emotions descriptions, emotional cues from images, or direct human interactions to dynamically generate facial expressions on robots that are realistic, expressive, and emotionally resonant.

Key Areas of Focus

● Interpreting Emotional and Stylistic Cues: Create algorithms capable of understanding complex emotional states and stylistic preferences from various inputs to guide the expression generation in robots.

● Robotic Expression Synthesis: Design and implement models that enable head robots to produce facial expressions that are not only mechanically feasible but also closely match the emotional and stylistic nuances conveyed by the input.

● Adaptive Mechanisms for Expression Generation: Employ adaptive strategies to adjust the robotic facial expressions in real-time, ensuring a high degree of accuracy and flexibility in reflecting intended emotions and styles.

● Alignment of Emotion and Style in Expressions: Develop methods to ensure that the robot's facial expressions are both emotionally authentic and stylistically appropriate, enhancing the robot's ability to engage with humans.

● Fine-Tuned Control of Robotic Expressions: Enable precise manipulation of the robot's facial features, allowing for detailed customization of expressions to better match human interactions and preferences.

Location: Robotics (405.652, Lab)

The challenge of generating human-like co-speech gestures represents a fascinating intersection of technology and creativity, offering ample opportunities for student research and innovation. Traditional methods in this domain have often been constrained by rigid style control mechanisms, utilizing predefined text labels or specific motion examples. Such limitations underscore the necessity for more adaptable solutions that can accurately capture and express a broader range of user intents.

This student project aims to explore advanced techniques for synthesizing realistic and stylized co-speech gestures, emphasizing the flexibility of style control beyond conventional boundaries. Leveraging the potential of large-scale contrastive language-image pre-training models, students are encouraged to develop methodologies that effectively extract and apply style representations from varied input sources, including text, motion clips, and videos. This initiative seeks to address and bridge the gap between user expectations and technological capabilities in gesture generation.

Key Areas of Focus

● Efficient Style Representation: Investigate methods for extracting detailed style representations from diverse inputs to capture nuanced user intents.

● Gesture Generation Techniques: Explore and develop models capable of producing high-quality, realistic gestures that align closely with intended styles.

● Dynamic Style Application: Implement adaptive techniques for infusing extracted style elements into the gesture generation process, ensuring accuracy and adaptability.

● Semantic Alignment: Devise strategies for aligning generated gestures with speech content semantically, using contrastive learning to ensure relevance and naturalness.

● Granular Style Control: Enable precise control over the styling of individual body parts, offering users the ability to fine-tune gestures to their specific preferences.

Location: Robotics (405.652, Lab)

Machine learning models are increasingly becoming larger as we try to capture complex behaviour. However, with this increasing complexity it can become difficult, or even impossible, to run such networks on typical embedded hardware. Instead, what methods can we apply to “optimise” these networks to make them simpler and more efficient?

This project will investigate approaches to simplify the structure of pre-trained MLP models to reduce their computational complexity while maintaining high accuracy. This may be through the use of a range of approaches such as pruning (removing the least-relevant components) or subdividing (splitting the network up into several smaller networks).

Through this approach, engineers would be able to train complex models offline on powerful hardware, and then create simplified versions that can be distributed for execution.

Location: Embedded Systems (405.760, Lab)

Music can be beautiful, messy, and emotive, but it almost always tells a story. This story is created by a composer who writes in the language of music. Popular AI models have begun to aid people in writing their own stories by taking in a few words and predicting the next set of words. Can we use AI models to help composers write music?

This project will build on a previous year’s project with the aim of integrating potential new techniques (such as audio diffusion, etc.) to generate music that fits within a given style. The idea would be to integrate this into a tool such as Antescofo to allow for generated accompaniments to be played in real-time.

Location: Embedded Systems (405.760, Lab)

Electric vehicles are becoming a large section of the automotive and motorsport industries. Following this trend, lots of student led Formula SAE racing teams design, build, and compete at international competitions have switched to electric vehicles. There are significant differences between EV and IC (internal combustion) and a Battery Management System (BMS) and Regenerative Braking is one of them.

BMS have balancing technology to ensure majority of the cells are at even voltages. The purpose is to increase the effective energy that can be used from the battery pack and ensure the cells are within their working voltage range. There are lots of balancing topologies including passive method, simply dissipate excess energy into heat, and active, which distribute unbalanced energy to each other's cells. Regenerative braking is to convert and recover vehicle's kinetic energy to increase the overall efficiency of the battery pack.

However, these balancing technologies have progressed mainly for commercial vehicles, which have different conditions to electric 'race' cars. Electric race cars have more aggressive driving, with much shorter brakes and high-power output compared to commercial EVs. This project proposes to find how well these commercial balancing technologies are adopted to the race car industry and further investigate new balancing technologies cooperating with regenerative braking to target specific, lower-capacity cells to raise their voltage for the purpose of increasing the usable energy within the pack. This project will then quantify how this technology will improve the battery pack's performance using the case study of a Formula SAE race car.

Location: Green Electronics (405.712, Lab)

With the recent weather events in Auckland, and many other recent disasters around the world, there has lacked a central online place where those affected can go to find out specific, up-to-date information (for example, a map of flooding-affected areas and road closures). This information exists online but is often in a variety of formats and in multiple locations.

In this project, you will investigate the use of modern data gathering and analysis techniques to automatically (as much as possible) curate and summarize important information related to disasters. The ideal end goal will be a prototype app or website where users can go to be presented with the most relevant, up-to-date information based on their location.

We already have a prototype of such an app from a previous part-4 project. This project will focus on applying more techniques, such as machine learning and data mining, to provide the prototype app with more accurate / useful news information.

Location: MDLS (405-569)

With the recent weather events in Auckland, and many other recent disasters around the world, there has lacked a central online place where those affected can go to find out specific, up-to-date information (for example, a map of flooding-affected areas and road closures). This information exists online but is often in a variety of formats and in multiple locations.

In this project, you will investigate the use of modern data gathering and analysis techniques to automatically (as much as possible) curate and summarize important information related to disasters. The ideal end goal will be a prototype app or website where users can go to be presented with the most relevant, up-to-date information based on their location.

An existing prototype app exists from a previous part-4 project. In this project, you will focus on undertaking a through analysis of the existing user interface, taking into account HCI - and more importantly, accessibility - principles. You will then refine and evaluate the prototype with a goal to be able to present the disaster information in a useful manner to users with a variety of needs - for example, "regular" users, elderly, visually impaired, etc.

Location: MDLS (405-569)

The Centre for Automation and Robotic Engineering Science (CARES) is investigating robot-driven underwater farm (aquaculture) automation. The motivation for this is that aquaculture has the potential to give us the ability to manage our marine resources better. A bottleneck for the uptake of aquaculture is the challenges of working underwater, and we believe robotics can help.

This project proposal is related to the CARES aquaculture investigation. However, it targets a fundamental problem associated with the camera-based vision systems for underwater robots, namely scene motion. Scene motion is typical since much of what we capture tends to be a non-rigid matter that flutters in the constant ebb and flow of underwater currents. The problem is that image sequence processing is complex since the motion makes it difficult to reconcile commonality across images.

A potential solution to this problem is Artificial Neural Network (ANN) based image synthesis. Such systems take image sequences as an input and use ANN technology to "learn" relationships between light ray directions and projected colours, allowing novel perspectives to be synthesised through inference by interpolation. Several of these approaches also incorporate a time element into their models [1,2], allowing the capture of dynamic scenes (scenes with movement). The claim is that these systems can synthesise images (generated from a sequence captured of a moving scene) at a "fixed time" (thus producing a synthetic image set from the perspective of a moving camera, in which the scene appears static[see here]).

The key deliverable of the proposed project is an assessment of some of the "state-of-the-art" versions of these algorithms from the perspective of using synthetic outputs for 3D spatial measurement and 3D/2.5D scene reconstructions.

References

[1] Z. Li, S. Niklaus, N. Snavely, and O. Wang. Neural scene flow fields for space-

time view synthesis of dynamic scenes. In Proceedings of the IEEE/CVF

Conference on Computer Vision and Pattern Recognition, pages 6498–6508,

2021.

[2] A. Pumarola, E. Corona, G. Pons-Moll, and F. Moreno-Noguer. D-nerf:

Neural radiance fields for dynamic scenes. In Proceedings of the IEEE/CVF

Conference on Computer Vision and Pattern Recognition, pages 10318–10327, 2021

Location: Radio Systems (405.736, Lab)

There is an immediate need to develop technologies that can precisely capture the emergence and trajectory of peoples’ resilience in high-risk industries. Despite government efforts, high risk industries such as construction continue to lead work-related harm and fatalities. The integration of psychophysiological monitoring with virtual reality (VR) makes it possible to document an individual’s experiences as they interact in simulations of real-world contexts. However, conventional psychophysiological-VR systems are constrained to measuring limited features of behaviour in fixed VR environments. This project will develop a VR simulation of health and safety situations that can be experienced and engaged with by professionals working in the construction industry. A range of psychophysiological measures will be captured from the professional within the VR environment to help identify the most effective set of sensors and measures to identify a person’s risk resilience. Ethics approval is already in place for this project for 2024.

Location: Green Electronics (405.712, Lab)

A problem in carparks around Auckland is scooter riders dashing in and out of carparks usually to commit an offence. Surveillance cameras cover the entrances to the carparks, but they don’t have automated detection of what is entering or exiting the carpark. This project will look to develop an accurate binary classifier to determine whether an image contains a scooter rider. A range of AI/vision techniques and various parameters could be used to perform the classification. Part of the project will be assessing which approach gives the most accurate identification within a short timeframe and with limited compute resources.

Data: A labelled set of images from carparks is available from the company Kauricone who are looking at ways their classification can be improved.

Location: MDLS (405-569)

Despite the successes and fast improvements in artificial intelligence, the interpretation of legal documents is still a highly manual process. Because of ethical considerations and the missing interpretability of modern deep learning models, to date, the preferred automation methodology is to translate the normative documents into a logic representation, which allows execution by an automated theorem prover. Because regulations are drafted in natural language and frequently amended, an automated or semi-automated translation process is required.

An end-to-end deep learning translation model was trained that allows experts to translate regulation clauses into LegalRuleML, the representation format selected for this study. A user interface to interact with this transformer-based deep learning model was developed to provide the user with full translations and auto-completion options. This project will investigate more user-friendly visual languages to represent the regulation clauses.

In this context, the following two aspects need consideration:

1.            Can the textual representation be mapped into a visual language?

2.            How can the auto-completion functionality be integrated into a visual editor?

In addition to developing a modern user interface, this project lets you experience direct interaction with state-of-the-art deep learning models and modify them where required to provide the best possible user experience.

Location: MDLS (405-569)

To this day, legal documents are primarily distributed as PDF documents, making it complex for machines to access the content. To automatically check for compliance with legal requirements, a common approach is translating the legal documents into a logical representation, which allows execution by an automated theorem prover. While textual content can be successfully translated into such a logical representation, there is limited research investigating how to translate tabular content into a logical representation.

The objective is to utilise modern deep learning models for natural language processing and computer vision to parse the tabular content and translate them into logic rules. The existing deep learning model or training data for translating textual content can be reused or modified.

This project lets you experiment with cutting-edge deep learning models, utilising them for a domain application.

Location: MDLS (405-569)

The Formula SAE competition is introducing an autonomous vehicle category. To meet this challenge, we aim 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 allow a robot to learn to operate based on its interaction with the environment without human input or control.

While RL has demonstrable capability at "learning" to extract relevant information from images directly through the auto-encoder paradigm, this approach adds complexity to the problem due to the non-trivial encoding of information within images. To assess how much of an impact this has, it is possible to compare the performance of RL systems that use images directly with systems that use a custom image data extraction layer using computer vision techniques.

This project proposes doing precisely this. Your mission is to create a computer vision layer for detecting and localising a slower-moving opponent vehicle on the track to support overtaking manoeuvres based on control logic learned through RL. In its simplest form, you will use images from an RGB-D active depth camera to (1) determine the position of the target vehicle in the image and (2) the vehicle's location (and orientation) in 3D space. Repeated measurements should give you speed and acceleration estimates (relative to your vehicle) and a set of historical waypoints that would allow you to interpolate a trajectory of the opponent vehicle indexed by time (also from your vehicle's frame of reference). Ambitious students may even try their hand at future trajectory extrapolation. Once the data extraction problems have been solved, the next challenge will be to design a logical structure for presenting this information to an RL algorithm. You will also need to design and develop real-world experiments (not in simulation) to verify the accuracy of your solution.

Time permitting, the project's final goal will be to compare your computer vision approach's RL training times and performance against a purely image-based RL solution (most likely in a simulated environment). This will be done with the collaboration and support of other teams and members within CARES (https://cares.blogs.auckland.ac.nz/). Note that much of the hardware and RL infrastructure for this project was developed within CARES by Henry and his team, and you will work with them on the RL side of the project while Trevor will provide support on the computer vision side.

Location: Radio Systems (405.736, Lab)

Seasonal infectious diseases such as influenza are a reoccurring challenge for healthcare systems worldwide. Besides influenza, several respiratory diseases cause symptoms similar to those caused by influenza, broadly called influenza-like illness or ILI. While most such ILI infections result in mild symptoms that are sometimes not even recognized by the infected person, there are still numerous severe ILI cases every year. The hospitalization burden caused by such severe cases is extremely difficult to foresee as many factors, such as weather, international travel, or immunization in the populations, have an impact on the transmission of respiratory diseases. Models that provide accurate forecasting of ILI cases could be extremely beneficial as they allow for proactive hospital management to be better prepared for higher demands during the winter season. Moreover, policymakers can use forecasting as guidelines to make better decisions about using intervention strategies that could be applied to decrease transmission rates before the health system is pushed beyond its limits. Machine learning can learn patterns from data to provide accurate models in forecasting time series data. Such time series can be derived from infectious disease surveillance systems becoming increasingly available in different countries. One of these systems was established in 2010 in the USA to collect ILI cases within different states providing a dataset covering ILI cases from more than one decade. Furthermore, given that ILI cases are separated by state, the data includes a geospatial component that multivariate algorithms can leverage. However, this data comes at weekly resolution, resulting in only 52 time points per year and a small data set. Small data is challenging for machine learning, and models trained on data with small sample sizes are often unreliable, especially for long-term prediction, as reported in a study exploring the USA-ILI dataset. Transfer learning uses pre-trained models to overcome the small data learning challenge. This concept is implemented within foundation models such as generative pre-trained transformers (GPTs) that got a lot of attention during the last few years as they, trained on an enormous amount of data, have been shown to provide highly accurate results on new tasks. Recently, TimeGPT-1, a GPT for time series forecasting, has been released, which could revolutionize forecasting based on small but highly informative disease surveillance datasets. This project aims to investigate the potential of a GPT model to forecast ILI cases compared to other baseline forecasting algorithms and more sophisticated machine learning concepts covering tree-based and artificial neural network models. This project aims to do this benchmarking while implementing a generic pipeline that can be applied to other datasets, potentially including an ILI surveillance dataset from New Zealand.

Location: MDLS (405-569)

This project involves iteratively refining a participative Unity-based installation where users experience immersive sound and visual feedback with a motion capture suit and VR headset. The immersive application directs gallery-visitors’ attention in various ways with sound and particle effects (for example: https://www.youtube.com/watch?v=wb_jM2kbqTA).  The practical goal would be to refine the development of a complete gallery installation experience. The research goal would be to understand how the application affects sensory perception, embodiment, and subjective experiences. Over the year, we would customise the application based on stakeholder feedback.

This project is connected to three others that focus on: 1) an embodied AI agent, 2) mixed reality responsive arduino controls, and 3) a Māori creative story (pūrākau) scene.

Location: MDLS (405-559)

This project involves iteratively refining an intelligent body tracking to a Unity-based installation where users experience immersive sound and visual feedback with motion capture and VR headset. The immersive application will display an intelligent agent that mimics variations of the visitor's movement.  The practical goal would be to integrate an intelligent movement avatar into a gallery installation experience. The research goal would be to understand how the agents affects movement, sensory perception, embodiment, and subjective experiences. Over the year, we would customise the agent based on stakeholder feedback.

This project is connected to three others that focus on: 1) the foundational visual and audio effects within the installation, 2) mixed reality responsive arduino controls, and 3) a Māori creative story (pūrākau) scene.

Location: MDLS (405-559)

This project involves iteratively refining a participative Unity-based installation where users experience immersive sound and visual feedback with a motion capture suit and VR headset. In previous versions, the installation was completly in virtual reality, and this year, we would like to Adruino prototyping for responsive theatre lighting. The real-world lights and immersive application directs gallery-visitors’ attention in various ways (for example: https://www.youtube.com/watch?v=wb_jM2kbqTA).  The practical goal would be to refine the development of a complete gallery installation experience and performance. The research goal would be to understand how the mixed reality and arduino controlled lights affects sensory perception, embodiment, and subjective experiences of dancers and audience members. Over the year, we would customise the application based on stakeholder feedback.

This project is connected to three others that focus on: 1) the foundational visual and audio effects within the installation, 2) an embodied AI agent, and 3) a Māori creative story (pūrākau) scene.

Location: MDLS (405-559)

The primary objective of this project is to improve the accuracy of results when testing using non­invasive methods to calculate glucose levels. The project aims to design and implement a signal processing and filtering system for an existing non-invasive blood glucose sensor.

Around 415 million individuals today are afflicted with diabetes. These individuals need to monitor their insulin levels, necessitating self-testing through the use of existing medical devices that involve pricking, which can be a painful and tiring process. While some non-invasive blood-glucose detection devices are available in the market, they lack the accuracy of traditional devices. In recent years, there has been extensive research on non-invasive blood-glucose detection using techniques such as NIR light, with universities around New Zealand such as the University of Canterbury exploring such methods in this field .

Many previous studies have highlighted similar challenges, such as the difficulty in accurately detecting, filtering and clearing the photo-diode signals to get an accurate result of glucose levels.

The aim of this project is to design a filter for the existing blood-glucose sensor with the purpose of enhancing the identification and effective filtration of the blood-glucose signal; aiming to increase accuracy of non-invasive methods when compared to invasive methods currently publicly available. In developing this design, we will explore options such as FIR filter, IIR filter, and look through parameters such as filter output stability, among others to identify relationships between the signals detected and glucose levels. The outcomes will be communicated through a dedicated application, providing users with a user-friendly tool for monitoring their blood-glucose levels.

Location: Control Systems (405.722, Lab)

Modern circuit board designs incorporate surface-mounted electronic components for their smaller sizes and better manufacturability on production lines with industrial pick-and-place machines. Unfortunately for a design engineer tasked with prototyping a board, these machines are too slow to programme and require too much overhead to be used when only a handful of boards are needed. The engineer's only real alternative is to prototype by hand, usually with a pair of tweezers, and either a soldering iron or a stencil with a reflow oven to place the components. All of these options are painstaking, time consuming, and prone to causing ergonomic strain.

Our project aims to investigate the usage of a lightweight gantry and vacuum pickup system (which is already an option to engineers) but with the addition of light machine vision to assist with the tedious task of component alignment. Unlike all other existing computerised systems, the user will drive the machine in real time. Such a device would be of considerable benefit to a design engineer's productivity and health, and especially for those that may have limited dexterity or impaired vision.

To this end, the project will involve the assembly of a mechanical pick-and-place machine with fly-by-wire motion control as an initial benchmark; the conception of a mathematical and algorithmic method of interpreting the operator's input; the implementation of a cooperative machine vision algorithm to provide positioning assistance that is responsive to this input; and the integration of these together to create a pick-and-place machine under shared control.

The end goal is to demonstrate a machine vision algorithm that is sympathetic to the intentions of a human, materialised as part of a pick-and-place machine where control over the placement head is intuitively and effectively shared.

Location: Control Systems (405.722, Lab)

This project will examine the impact of criticism during software code review, particularly destructive v constructive critism, on software code quality. The project may involve collecting data from open source software projects to analyse real code review comments and their impact on code quality. It may also involve designing and deploying a questionnaire to examine perceptions of software practitioners on these topics.

This is an extension of a previous Part IV project which was published (see https://kblincoe.github.io/publications/2022_CSCW_Destructive.pdf).

Location: HASEL (405.662, Lab)

The project explores fusion of a commercial software for building fixed industrial automation systems (IAS) and novel software paradigm that makes those systems dynamically reconfigurable. Dynamically reconfigurable systems support advanced features of IAS such as dynamic change of the products on the same production line, fault-tolerance and changes in physical configuration (adding, removing parts of the system). As the starting point, it uses Beckhoff commercial hardware and software platform for industrial control systems based on a network of industrial PCs and interfacing nodes connected through EtherCAT, an industrial proof networking system, and TwinCAT software framework for development of control systems.

The new software paradigm uses SystemJ concurrent programming language for the development of control that supports (dynamic) reconfigurability to tackle specific scenarios such as failure of specific parts of the system and fail-safe continuation of its operation, change of production process and products, etc.  The new paradigm will enable modelling the system operation and simulation before the deployment of software on the real platform with practically no changes of the developed controller model. As a testbed, prototype manufacturing system in Industrial Informatics lab will be used.

Location: Embedded Systems (405.760, Lab)

We are exploring the use of digital twin (DT) ideas in industrial automation systems (IAS). The project has a big level of novelty where the whole IAS will be based on the use of three types of DTs: (1) DTs of workpiece which after completion of the “recipe” becomes a final product, DTs of workstations that make physical changes on the workpieces as required by recipe, and DTs of transportation system (such as conveyors, turntables, arm manipulators, temporary storage bays, and robots). By using DT technology we are proposing a completely new production paradigm and formalisation of IAS that produces final products by applying discrete operations as specified in the recipe.

Location: Embedded Systems (405.760, Lab)

A new concurrent programming language called SystemGALS proposed within Embedded Systems Research Group, which can be considered an extension of C programming language, but with formal mathematical foundation, requires new computer platforms to efficiently execute concurrent programs that mix reactive behaviours with more elaborate data processing/computations.

The project will focus on the key building block, a tandem processor that a in-house developed reactive processor and Intel’s Nios 2  processor (or its new RISC-V based Intel version). The new processor is referred to as ReactiveNiOS. In its first incarnation of ReactiveNiOS, NiOS will be considered the main processor and reactive processor will be considered its co-processor. Programming the ReactiveNiOS tandem will be done through NiOS programming tools, and the ReCOP instructions will be made available for use as the extended NiOS instruction set.

Location: Embedded Systems (405.760, Lab)

Multi-modal learning uses multiple representations and interactions for supporting learning. For example, mathematics can be taught by using number representations, symbols (Latin numbers), or objects (both physical and virtual). It has been shown that many students learn better by using multiple representations and interactions (seeing, talking, doing).In this project we investigate whether tangible Augmented Reality (AR) can be used for learning data structures. Tangible AR combines real world objects with virtual augmentations. For example, students might use wooden blocks or coloured pieces as list elements and the program augments them with virtual information, e.g., links between elements.The project continues a SEP4P from last year, which developed a prototype for teaching the linked list data structure. We will extend the current prototype to add more functionalities and learning tasks, investigate adding additional data structures, and perform an extensive user study with stage 1 students investigating the tool’s usability and effectiveness.

Location: MDLS (405-559)

The performance of machine learning models depends on the quality of the underlying data. Malicious actors can attack the model by poisoning the training data. Current detectors are tied to either specific data types, models, or attacks, and therefore have limited applicability in real-world scenarios. We have developed a novel fully-agnostic framework, DIVA (Detecting InVisible Attacks), that detects attacks solely relying on analyzing the potentially poisoned data set. DIVA is based on the idea that poisoning attacks can be detected by comparing the classifier's accuracy on poisoned and clean data and pre-trains a meta-learner using Complexity Measures to estimate the otherwise unknown accuracy on a hypothetical clean dataset. The framework applies to generic poisoning attacks. For evaluation purposes, we focus on label-flipping attacks.

In this project, we will use DIVA’s existing implementation and set up, run, improve, visualize, and interpret a comprehensive set of experiments gaining insights into DIVA’s behavior, strengths and weaknesses. The ideal outcome of this project is a conference publication and a detailed report on DIVA's performence.

Location: MDLS (405-559)

Resilience and, in more specific terms, robustness is more relevant now than ever, with the increase of extreme weather events in frequency and intensity due to the effects of climate change. This is especially relevant here in New Zealand, where we are very susceptible to such events as can be seen last year with, for example, Cyclone Gabrielle.It is important for the grid to be robust, a part of resilience which means the grid's ability to resist these extreme weather events. For example, the least amount of customers on the grid possibly losing power during a severe weather event.

Thiss project will focus on whether areas with more renewable energy experience fewer power outages during extreme weather events compared to areas that mostly rely on fossil fuels to generate their electricity. We will also investigate if Microgrids with small-scale renewable energy generation, like Solar and small-scale battery storage experience fewer outages during extreme weather events.

The project team will analyse Hazard Data like wind and precipitation, potentially from government entries like metservice, niwa and geonet. And compare it to outage data which we can get from Transpower for transmission outage data and the local distribution companies like Vector for more localised outage information. We will perform a Geospatial analysis on this data, and with the help of machine learning, we can determine which areas were the most prone to these outages and determine if the increased presence of renewable energy allowed certain regions to fare better during extreme weather events. From this, we could potentially recommend if further investment in renewable energy in the form of either large hydro generation or small-scale microgrids involving solar and battery storage would improve grid robustness.

Location: Power Systems (405.628, Lab)

The primary goal is to create an intuitive and engaging dashboard web/app that presents users with a visual summary of their emotional states, detected through our ongoing emotion recognition research. This app will serve as a tool for individuals to monitor their emotional health, identify trends, and potentially inform their mental health care strategies.

We are looking to get students with skills in software development, data visualization, and web/mobile app development. The students will be responsible for:

1.     Translating UX design concepts into functional app features.

2.     Implementing data visualization techniques to represent emotional states in an understandable and aesthetically pleasing manner.

3.     Integrating the app with existing databases to fetch and display emotional data.

4.     Testing and refining the app based on user feedback to ensure usability and effectiveness.

Location: MDLS (405-559)

The chemical molecule space is gigantic, and machine learning (ML) models predicting molecule properties are typically developed for small and specific applications within that space. One example is the field of drug discovery where models help to focus experimental efforts on only the most promising drug candidates by predicting their properties (such as toxicity and reaction with specific bacteria) beforehand. A common concept to navigate the gigantic molecule space is to define an applicability domain for each model – the domain within the chemical space for which the model can make predictions of a given reliability.In machine learning research, the question of whether a model can make reliable predictions also is a hot topic. Research has shown that many ML algorithms are vulnerable to adversarial attacks. In an adversarial attack, an attacker meticulously crafts adversarial examples by exploiting the ML model's weakness. By adding small perturbations to the benign instances, the adversary forces the model to produce erroneous predictions with high confidence. Investigating successful adversarial attacks contributes to understanding the model’s weaknesses, and adversarial defenses have been developed to detect these attacks.Both applicability domain and adversarial defenses aim to detect weaknesses of a model by searching for model inputs resulting in false or unreliable outputs. However, they approach the problem from different perspectives testing for different vulnerabilities. We hypothesize that connecting both research areas will result in meaningful insights that ultimately enhance research in both areas.In this project, we will build on preliminary implementations and experimental setups for both applicability domain and adversarial attacks, and design, implement, run, visualize and analyze a comprehensive set of experiments. These experiments will transfer applicability domain approaches to adversarial defenses and vice versa.

Location: MDLS (405-559)

Emotions significantly influence human behaviour, affecting cognition, decision-making, attention, perception, and learning. These emotions are often shaped by social interactions, past experiences, and autobiographical memory (AM)—a construct of self-referential and emotional processes. With the advent of immersive virtual reality (VR) technologies and wearable physiological sensing, new avenues have opened for exploring how VR can enhance the ecological validity of psychophysiological research. This project aims to investigate the retrieval of autobiographical memories and the induction of emotions within VR environments. By leveraging VR's immersive capabilities alongside physiological measures such as EEG, we seek to understand how VR experiences can influence AM retrieval and emotional responses, comparing these effects to those observed in traditional laboratory settings.

Part 4 student’s Role:

·       Design and implement VR experiments that simulate real-life experiences to study AM retrieval and emotion induction.

·       Collect and analyse physiological data (e.g., EEG) to assess memory retrieval mechanisms and emotional responses within VR settings.

·       Compare VR-induced AM retrieval and emotional responses to those elicited in conventional laboratory environments.

·       Contribute to the understanding of the underlying neural mechanisms associated with AM retrieval and emotion induction in VR.

Location: Radio Systems (405.736, Lab)

Grid-interconnection of BESS (Battery Energy Storage Systems) to MV/LV distribution network would impact the distribution network voltage profile, thermal loading, power flow, safety, and protection. Grid-interconnections standards and guidelines mandate the interconnection requirements to regulate general technical requirements, power quality requirements and safety and protection requirements. There are no interconnection standard or guideline exclusively for BESS. A 2020 part-IV project foucssed on developing some best practices of guidelines that could be followed by NZ Distribution Utilities in this space. This could be used for grid-following system

In the last couple of years we have supported SolarZero towards developing solutions towards grid-supporting(Link)

Recently there has been development from IEEE on Grid-supporting and Grid-forming Microgrid architectures in IEEE publications.  The implementation is primarily based on inverter-interfaced PV, storage and other micro-generation resources.  There is a need to research into developing interconnection guidelines if these microgrid architecture become operational in NZ in the coming years which is the focus of this project.

This project won theNZ Energy Innovation Award in 2023

We recently organized aninternational eventwhere in one of the panels the development, standardization and uptake was discussed.    If required, we can connect with those experts from international national labs like NREL and Sandia.

Location: Power Systems (405.628, Lab)

Running complex sets of machine learning experiments is challenging and time-consuming due to the lack of a unified framework. This leaves researchers forced to spend time implementing necessary features such as parallelization, caching, and checkpointing themselves instead of focussing on their project. To simplify the process, we have developed MEMENTO, a Python package that is designed to aid researchers and data scientists in the efficient management and execution of computationally intensive experiments. MEMENTO has the capacity to streamline any experimental pipeline by providing a straightforward configuration matrix and the ability to concurrently run experiments across multiple threads.Research has shown that many ML algorithms are vulnerable to adversarial attacks. In an adversarial attack, an attacker meticulously crafts adversarial examples by exploiting the ML model's weakness. By adding small perturbations to the benign instances, the adversary forces the model to produce erroneous predictions with high confidence. Investigating successful adversarial attacks contributes to understanding the model’s weaknesses, and adversarial defenses have been developed to detect these attacks.New adversarial attacks and defenses are published frequently, making the task of benchmarking them increasingly tedious. To thoroughly evaluate an attack, a researcher would need to test the new attack against existing attacks on multiple models on multiple datasets under different conditions and while facing different existing defense strategies. This process can be drastically simplified using MEMENTO, particularly if most benchmark results are already stored and only the experiments regarding the new attack need to be run.In this project, we will build upon a previous project developing a consistent framework integrating multiple attacks into MEMENTO. We will extend this framework, run all set up experiments, provide output reports and visualizations for the researcher to use, and showcase contributions in a research paper.

Location: MDLS (405-559)

Ensuring consistent quality in manufacturing lines can be a challenge. This is where vision systems with AI come in, offering a potential solution to streamline quality control and improve overall consistency.

Fisher & Paykel Healthcare is looking to conduct research into the viability of using development of a vision system that can be used in line with a manufacturing process in the OSA department.

The desired output is a self-contained system that will output a pass/fail result and categorize the failure condition as components are presented to the system within a 10-second cycle time.

You will be provided with predefined pass and fail component samples and a document that defines the failure categories.

You will have access to an engineering manager and process development engineer for guidance on a regular basis. You will also have supervised access to Fisher & Paykel Healthcare’s manufacturing plant, workshops and labs as required.

Location: Robotics (405.652, Lab)

Your mission in this project is to construct an indoor 3D scanner using a robot arm (link). You will experiment with two camera systems: a calibrated stereo pair and a structured lighting Zivid (link).

Note that coming into the project, you will be provided with all the software you need to control the arm and the cameras. Concerning the stereo camera system, you will be given a means to calibrate the cameras and generate 2.5D images of the scanned object. The robot arm software you are provided with should also give you accurate (sub-millimetre) pose estimates, allowing you to determine the relative positions of each 2.5D image captured to assemble a full 3D reconstruction.

Your primary focus will be on combining individual frames into a cohesive 3D model, which can be tricky. Typically, a significant overlap exists between frames, which leads to considerable data redundancy. Measurement noise (along with minor changes in lighting and shadows) is another issue that can produce noticeable data inconsistencies. Your challenge in this project is to find ways to combine individual point clouds into a single cohesive model with an appropriately sized data footprint, appropriate noise filtering, and sufficient interpolation between frames to minimize "seam" artefacts as optimally as possible.

Concerning theory, older techniques such as marching cubes, Poisson surface reconstruction and ball pivoting lack robustness and require high-quality raw point clouds. Volumetric approaches are promising, and it is worth looking at Kinfu libraries in OpenCV's "contrib" modules - although this approach is limited in resolution. Modern machine learning-based approaches are currently leading the way [1,2].

Note that you will have access to a 3D printer that might be a helpful tool for assessing models.

Your research will be done in the context of the CARES (link) robotics faculty centre, and your work could contribute to ground-truthing robotic vision systems. Note that in this context, there will likely be a focus on 3D reconstruction for agricultural applications.

[1] Wang, Fangjinhua, et al. "Patchmatchnet: Learned multi-view patchmatch stereo." Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2021.

[2] Azinović, Dejan, et al. "Neural rgb-d surface reconstruction." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022.

Location: Radio Systems (405.736, Lab)

This project builds upon the previous part 4 project focused on disease detection using machine learning. A notable gap in the prior work was its sensitivity to lighting conditions. In the previous proposal, a multispectral camera was employed to identify spider mites on beans, with handheld Halogen lights. This project aims to develop a housing unit where the halogen/LED light is seamlessly integrated with the camera and a prove of concept that the current system can improve the result of disease detection.

The student involved will be responsible for gathering multispectral/NIR data from plants and annotating them. Subsequently, they will delve into the creation of a deep learning-based computer vision model geared towards disease detection on both leaves and stems. This initiative will provide hands-on experience in integrating lights with multispectral cameras. The students will find the suitable combination of lighting to have a consistent disease detection results.

Execution of this project will predominantly involve Python and/or C++ within a Robotic Operating System (ROS) framework. Additionally, there may be opportunities to leverage existing Python projects from the CARES and MaaraTech repositories, potentially streamlining aspects of the research.

Location: Robotics (405.652, Lab)

Fault Induced Delayed Voltage Recovery (FIDVR)is a well-understood issue with various solution strategies known for transmission systems.

Due to  well established IEC 61850 implementation methods, smarter solutions for FIDVR are now deployed in transmission networks across the world.However, for distribution networks, FIDVR impacts on protection has been primarily around delayed over-current operation.  Active distribution network operation is proliferating across the globe with large-scale penetrations of PV, Battery, EV and other distributed energy technologies which primarily has an inverter-interface at the Point of Common Coupling (PCC) to the grid.  This project will look towards assessing if FIDVR phenomenon has become pronounced due to "Low Voltage Ride Through" (LVRT) and "High Voltage Ride Through (HVRT) grid codes being proposed for Smart Inverter operation during distribution grid disturbances.Solutions to these using IEC 61850 based solutions like R-GOOSE and novel protection algorithms/implementation will also need to be investigated.

With large-scale solar farms to be grid connectedwithin the next couple of years in NZ, this aspect is likely to emerge as a potential trigger for large-scale blackout in Norht Island.UoA Power Systems Group (PSG) have carried out several IEC 61850 based projects with New Zealand distribution  utilities in the past and these entities will be supporting the development of the prototype solutions. In particular Transpower and Vector protection engineer's will be providing their inputs along with active support from PSG PhD's and postdoc's actively researching in Smart Grid projects. (e.g.  Automatic Bus Transfer scheme). Two-decades of IEC 61850 has been captured in thisCIGRE GREEN Book.This e-book is available from UoA library and Chapter 8 identifies how functional schemes could be developed for various schemes.

Location: Power Systems (405.628, Lab)

Music is a powerful motivator and is extensively used in gyms to make exercises more enjoyable and motivational. Music rhythm can help to increase physical activity by encouraging users to perform exercises, such as running, with the rhythm of the music. Several rhythm games, such as Dance Dance Revolution and Beat Sabre have become commercial successes.

Most existing rhythm games have been designed for home used. To encourage outdoor exercises mobile AR games have been developed. In this research we will investigate whether concepts from rhythm games can be combined with concepts of mobile AR games to increases users’ motivation, enjoyment, and performance when doing outdoor exercises.

In a previous SEP4P we developed simple mobile AR rhythm game which measure users’ step frequency and the music rhythm. The game play encourages increased synchronisation between step frequency and music rhythm using a scoring / reward mechanism and appropriate visual and auditory feedback.In this project we will investigate and implement functionalities to make the game more effective and useful. Ideally, this means implementing a streaming functionality and recommended system, allowing different types of exercises (slow/fast walking, running), and different types of auditory feedback and reward systems. The improved game should be evaluated with a substantial user study.

Location: MDLS (405-559)

Perceptual abilities can be affected by various conditions and also naturally decline with age. However, it has been shown that perceptual training can be used to mitigate decline and for rehabilitation.

AR and VR are increasingly being used to diagnose, treat and manage a wide range of health issues. However, a significant portion of these virtual diagnostic and treatment tools have been designed and developed for specific use cases. There exists little prior work that has incorporated a wide range of tools and techniques into a single platform to deliver rehabilitation services.

In previous work we developed in collaboration with researchers from Audiology bi-modal (audition and vision) AR and/or VR platform to deliver rehabilitation services to individuals diagnosed with tinnitus and balance disorders. In this research we will improve the prototype to make it more suitable for clinical research. This includes an improve interface for clinical users and in particular mechanisms to enable/disable parts of the environment (unisensory vs. multisensory stimulation), a suitable patient interface, improved tools for assessing/recording patient’s perceptions and reactions, investigating the use of different HMDs with eye tracking, adding new virtual environments and extending existing ones to allow patients longer and more stimulating explorations.

Location: MDLS (405-559)

When an algorithm is processing some data, e.g. numbers are sorted, an image is filters or a neural network is trained, the foremost objective has been to do as this fast as possible. For decades new processors have become faster, by executing more operations per second, and new algorithms have been more efficient with shorter runtimes. Today, however, computers, from small devices to large data centres and supercomputers, consume a significant and growing amount of electrical energy. In order to achieve more sustainable computing, we need to reduce the energy consumption of computation. While this is being addressed on a technological level, with smaller processors, memories and networks, the algorithmic side receives little attention.In this project you will investigate and compare computing tasks, for example processing a batch of images or training a neural network, in terms of the energy that is needed to complete the computation. For many problems there exist alternative approaches, only think of the many ways numbers can be sorted. We want to find out which of these alternatives for a given problem needs less energy to obtain the result. This will be done by studying the literature on the topic, identifying computational problems and algorithms with alternative approaches and then measuring  and comparing their energy consumption of a given lab computer of the PARC (Parallel and Reconfigurable Computing) lab. Energy consumption can vary not only by how long it takes to compute, but by how many processors are involved, how much data is transferred and how often processors are idle. We aim to characterise patterns and approaches that can be used to identify the best techniques to be used from areas like data science and machine learning which have strong saving potential and a growing appetite for computing power.

Location: Embedded Systems (405.760, Lab)

Today, virtually all computers are parallel systems with multiple processors. To efficiently use such a system it is crucial to carefully map and schedule the tasks of a program onto the processors. This is a very hard optimisation problem and can be described as a Mixed Integer Linear Program (MILP). Despite extensive research, only small instances can be computed optimally by modern solvers today.Delayed Column Generation is an algorithm that has been successfully deployed for large MILP in other domains. The advantage is that not all possibilities need to be enumerated. One starts with a smaller problem and only brings in new variables as needed. In practise this can be significantly faster than directly solving the original MILP. In this project you will be investigating the modelling and implementation of the task scheduling problem with a Column Generation algorithm.Students will benefit fromfamiliarity with Linear Programming and solvers, e.g. CPLEX or Gruobi

Location: Control Systems (405.722, Lab)

Globally, metropolitan cities are rushing towards Net Zero Emissions by 2030. Use of active mode is becoming easier with new micro-mobility modes such as scooters and e-bikes. Despite these changes, a large proportion of the population in car-centric cities do not cycle. Changes in infrastructure are proving to be not enough to change safety perception and the ease of cycling.

In this research we will develop a VR environment to investigate cyclist’s safety perception of road and cycleway infrastructure and how it can be improved. In order to ensure a sufficient scale, we will use procedural generation methods and ideally create “infinite cities” where roads and cycleways contain pre-defined properties, but do not have a natural boundary and properties can be modified by the study supervisor.

Location: MDLS (405-559)

RFID tags have been used as ID tokens in many applications. The project aims to explore the potential of using RFID technology as a basis for battery-less transient power sensor nodes for various sensing applications, where retrieving sensor nodes or replacing batteries are impossible. The end goal is to make these devices as low cost as possible to allow dense deployments with a high level of redundancy. A customized drone will need to be equipped with high power RFID transceivers to provide enough power to these battery-less nodes to allow them to perform one or multiple sensing tasks and report the data back to the drone.

@font-face {font-family:"Cambria Math"; panose-1:2 4 5 3 5 4 6 3 2 4; mso-font-charset:0; mso-generic-font-family:roman; mso-font-pitch:variable; mso-font-signature:-536870145 1107305727 0 0 415 0;}@font-face {font-family:Calibri; panose-1:2 15 5 2 2 2 4 3 2 4; mso-font-charset:0; mso-generic-font-family:swiss; mso-font-pitch:variable; mso-font-signature:-469750017 -1073732485 9 0 511 0;}p.MsoNormal, li.MsoNormal, div.MsoNormal {mso-style-unhide:no; mso-style-qformat:yes; mso-style-parent:""; margin:0cm; mso-pagination:widow-orphan; font-size:12.0pt; font-family:"Calibri",sans-serif; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-fareast-font-family:Calibri; mso-fareast-theme-font:minor-latin; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin; mso-bidi-font-family:Arial; mso-bidi-theme-font:minor-bidi; mso-ansi-language:EN-GB; mso-fareast-language:EN-US;}.MsoChpDefault {mso-style-type:export-only; mso-default-props:yes; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-fareast-font-family:Calibri; mso-fareast-theme-font:minor-latin; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin; mso-bidi-font-family:Arial; mso-bidi-theme-font:minor-bidi; mso-font-kerning:0pt; mso-ligatures:none; mso-fareast-language:EN-US;}div.WordSection1 {page:WordSection1;}

Location: Embedded Systems (405.760, Lab)

Lake Waikare (Hamilton, Wakaito) is a heavily polluted lake in the North Island. Intensive farming has caused excessive level of nitrogen and phosphorus to build up which caused negative impact on the aquatic life. The project aims to use sensors deployed on an autonomous watercraft to collect spatial and temporal data of the lake on a regular basis to establish a digital twin model of the lake, which has the potential to offer a better understanding on the flow of water and pollutants in the lake. The targeted sensors include, but not limited to: temperature, flow, turbidity, dissolved oxygen, pH, conductivity.

Location: Embedded Systems (405.760, Lab)

mmWave is an emerging technology and has several benefits in comparison to existing sensors. However, it is currently rarely used for smart environment applications. This project aims to explore the capability of mmWave radar for the purpose of performing various types of indoor human activity recognition. The team will need to investigate and design a mmWave radar based system to recognise common human activities.

Location: Embedded Systems (405.760, Lab)

Music is an important way for humans to express themselves, unite, relate, and gain enjoyment. While many people desire for themselves, or their loved ones, to achieve some form of musical prowess, the discipline remains encumbered by low motivation and high dropout rates. This is despite long established trends between exposure to learning an instrument and enhanced cognitive development. Existing research suggests that such issues may stem from the lack of perceived progress during the early stages of music instruction.

In this project we will use Augmented Reality to make the piano learning process easier and more enjoyable. We will evaluate the current state of the art of AR piano tutoring and based on this propose novel concepts to improve learning outcomes. The research will build on previous projects conducted by the supervisor’s research group. Possible research goals are increasing motivation using gamification, teaching sight reading, the use of multi-modal feedback and sensor fusion to improve learners’ skills, or a combination of multiple of these topics. The exact research question will depend on the results of the literature review and the student’s skills, preferences, and opportunities for creating publishable research.

Location: MDLS (405-559)

This project focuses on the development of user interaction in Augmented Reality (AR) by providing a highly customizable and user-friendly application that leverages the power of AR to enhance users' interaction with smart home devices in the field of Computer Human Interaction.

Users will initiate the experience by registering all target devices, such as Wi-Fi lights and smart air conditioners, subsequently inputting static custom gestures for each device. Once configured, performing the corresponding gesture will enable seamless interaction with the selected smart device. Utilizing state-of-the-art gesture recognition and model training technologies, the project focus is on developing a comprehensive top-level interface that is capable of integrating a wide range of smart home devices. This interface features intuitive parameter ports, allowing lower-level devices to be connected effortlessly.

By prioritizing ease of use and broad compatibility, our project aims to deliver a universal solution that enhances daily life through personalized, gesture-based interactions with smart home technology, making advanced home automation accessible and convenient for all users.

Location: Embedded Systems (405.760, Lab)

Digital substations represent a significant advancement in the modernization of power systems infrastructure, leveraging technologies to enhance efficiency, reliability, and monitoring capabilities. Unlike conventional substations architectures that rely on point-to-point wiring, digital substations employ standardized communication protocols to facilitate seamless integration of intelligent electronic devices (IEDs) and substation automation systems. At the forefront of this transition is the International Electrotechnical Commission (IEC) 61850 standard, which provides a comprehensive framework for the design, engineering, and operation of digital substations. IEC 61850 defines a common language and data model for communication between IEDs, control systems, and utility SCADA (Supervisory Control and Data Acquisition) networks, promoting interoperability, flexibility, and scalability across diverse substation architectures. The adoption of IEC 61850-compliant solutions, is a key strategy for utilities to unlock the full potential of smart grid technologies.

However, as the threat landscape in the digital age evolves, safeguarding critical infrastructure such as electric utilities become paramount. To defend against growing cybersecurity threats, many electric utilities are beginning to integrate intrusion detection systems (IDS) to monitor electrical substation automation communications networks. These systems examine Ethernet network traffic to detect unusual or malicious behaviour that could indicate a cyber-attack and thus, must integrate with systems that employ IEC 61850.

IDS solutions employ two primary techniques for cyber-attack detection: signature-based and anomaly-based detection. Signature-based detection compares network traffic against predefined pattern rules indicative of malicious activity. In contrast, anomaly detection learns normal network behaviour over time and triggers alerts when deviations from this baseline occur.

While anomaly detection holds promise due to its adaptability, concerns have been raised regarding its suitability for digital substations (Cybersecurity in Substations – Attack Vectors on Substations and their Countermeasures - OMICRON (omicronenergy.com). Unlike traditional IT environments, digital substations exhibit unique communication patterns, characterized by long periods of stability interspersed with bursts of activity during primary electric events such as switching or power system faults. This dynamic behaviour may inadvertently trigger false alarms in anomaly-based IDS, potentially overwhelming cybersecurity analysts with unnecessary alerts.

References/Background:

• Kang, B., Maynard, P., McLaughlin, K., Sezer, S., Andrén, F., Seitl, C., ... Strasser, T. “Investigating Cyber-Physical Attacks against IEC 61850 Photovoltaic Inverter Installations.” Proceedings of IEEE 20th Conference on Emerging Technologies & Factory Automation (ETFA). (pp. 1-8). 2015, Luxembourg.
• Hoyos, J., Dehus, M., Brown, T.X. - “Exploiting the GOOSE Protocol: A Practical Attack on Cyber[1]infrastructure”. GC'12 Workshop: Smart Grid Communications: Design for Performance 2012
• D.L.P Jenkins "How to hack an IEC 61850 system (or protect one)" Study Committee B5 Colloquium September 11-15, 2017 Auckland, New Zealand

Location: Power Systems (405.628, Lab)

There is a large and growing number of software vulnerabilities being discovered and reported, resulting in CVE (common vulnerability exposure) entries in the National Vulnerability Database (NVD). In open source projects, CVEs are usually quickly addressed with patch commits, and subsequent releases of new versions. Often those commits include regression tests to give some evidence that the vulnerability is fixed.

The objective of this research project is to investigate whether novel techniques can be developed to automatically assess those commits. In particular, comparing the source code and test executions before and after the patch could increase or decrease our confidence that the patch fixed the vulnerabilities. Analysing how developers of open source projects patch and test vulnerabilities might help us to achieve this goal. The project can use public vulnerability datasets such as  https://github.com/tuhh-softsec/vul4j to conduct experiments, and evaluate an automated tool to be developed.

Location: HASEL (405.662, Lab)

In recent years Large Language Models (e.g., ChatGPT) are showing extraordinary performance. Many Software Engineering researchers around the world are now investigating how to exploit LLMs to automating software engineering tasks like source code or test case generation, showing promising results. When doing so, researchers are facing the challenge of automatically executing and testing the LLM-generated code, which often does not even compile. Indeed, it is an open research problem on how to automatically fix, validate, and improve LLM-generated code.

This research project aims to design and develop a novel framework for code and test generation with LLMs. The framework will intercept the interactions of developers and LLMs, to automatically fix, validate, or improve the code or test cases generated by LLMs. The framework is envisioned to be used by developers and researchers who want to obtain code from LLMs that is readily compilable and runnable with also some guarantees of the correctness of the code.

Location: HASEL (405.662, Lab)

Behaviour-Driven Development (BDD) is a collaborative software development methodology involving developers, QA professionals, and non-technical stakeholders. It centres on constructing human-readable scenarios in Gherkin, detailing the desired system behaviour. Selenium test cases, specifically designed for the user interface (UI), play a pivotal role in BDD by directly implementing Gherkin scenarios—establishing a clear and direct one-to-one mapping between the articulated behaviour and its corresponding test implementation. This approach ensures a seamless alignment of software development with specified business objectives.

This research project aims at designing and implementing a novel technique to automating the creation of Gherkin scenarios and their corresponding Selenium test cases within the context of BDD. By mining Gherkin scenarios from GitHub repositories, we aim to analyse existing patterns and structures, subsequently developing techniques for the automatic generation of Gherkin scenarios. Additionally, the project aims to propose methods to seamlessly derive Selenium test cases from the generated scenarios. The methodology involves data collection, natural language processing, machine learning, and integration with Selenium. We plan to investigate the use of LLMs (e.g., ChatGPT) to automate some of the scenarios and test cases generation. You are recommended to choose this project if you are also enrolled in SOFTENG754, as BDD will be covered in the course.

Location: HASEL (405.662, Lab)

A research project the a the UoA is developing a novel behavioural monitoring system for beehives. Part of this system will require CO2 sensing and we need to select the most appropriate sensor for use in the hive. This project will involve testing a range of available CO2 sensors in a range of environments to enable an informed decision about which is most appropriate for use in monitoring bee behaviour. This may or may not be extended to include pheromone sensing. It will require Raspberry Pi / Arduino programming, sensor testing, and data analysis.

All applicants are required to meet the supervisor before submission of the project preferences.

Location: Control Systems (405.722, Lab)

A research project the a the UoA is developing a novel behavioural monitoring system for beehives. A part of this system will require the analysis of large sets of multivariate time-series data. This project will utilise existing datasets to develop ML methods that might be best suited to identifying changes in circadian rhythms that are predictive of certain events. It will require understanding some biological principles, data cleaning, and advanced analysis.

All applicants are required to meet the supervisor before submission of the project preferences.

Location: Control Systems (405.722, Lab)

There has been a lot of research in the past about detecting bias in Large Language Models (LLMs) but it is very US centric. The goal of this project is to develop software that also detects bias in the NZ context. The project will be part of an ongoing research project, which identifies and tackles bias in AI. The output of the project will form the basis for an AI bias detetction tool that is specific to NZ industry.

Location: MDLS (405-559)

We are working with Parininihi ki Waitotara (PKW), a Māori organisation based in Taranaki. Like many other organisations, PKW have shareholders who transfer shares over time, typically between family members. They wish to better understand these transfers by creating visualisations of the 'flow' of shares between people over time. Creating an easy-to-follow visualisation requires careful layout of the people (nodes) to avoid a spaghetti-style mishmash of flows. We wish to develop optimisation models to create a visually-optimal network representation. This is an example of a graph-layout problem. These flow graphs are known as Sankey diagrams.

We have been working on visualisations that use the Unity games engine and C#. These have focussed on showing early PkW shareholders. We wish to extend and enhance these visualisations by creating appealing interactive visualisations that show a range of data sets, including the share transfers over times. We have not previously generated Sankey diagrams using these 3D visualisation tools.

A second problem faced by PKW and other organisations is the handling of missing data. For example, a company typically reports its shareholding annually, but does not report the transfers that occurred during the year. Groupings of shares are typically transferred in predictable ways, such as being transferred to a single individual, or being split evenly between a small number of individuals. These patterns make it possible to deduce likely share transfers, and thus infer the missing share transfer data.

If time allows, we also wish to experiment with modern Web-based visualisations including WebGL and WebAssembly.

Location: Control Systems (405.722, Lab)

There are many sources of information that you might want to access when looking at a piece of land you could build on. The council holds information on the land and its allowable use, they also can provide information on the neighbouring properties and their owners, LINZ hold information on the actual boundaries, Google and others have photographic views of the land, MBIE has the codes that pertain to that type of land, Waka Kotahi on the roading network and connections to the area, etc, etc. Being able to ask questions about a piece of land and get accurate answers and visuals of the land and its surrounds would be useful for anyone wishing to purchase some land. This project will investigate how to provision textual and graphical information based on a user’s natural language queries.

Location: MDLS (405-559)

Dedicated mobile devices are widely used for tracking people who need special care or assistance, for example, patients suffering from dementia. Apart from wireless data transfer, it has been a challenging task to maintain the power needed for driving these devices. This project is to develop a wireless/contactless platform for charging the battery or supercapacitors in UWB (Ultra-WhiteBand) devices for tracking dementia patients. The work involves a literature review of existing techniques, the design of a new inductive or capacitive power supply with energy storage, circuit simulation, and practical evaluation.

Location: Power Electronics (405.614, Lab)

Wheel hub motors are used for electric bikes, robots, AGVs (Automatic Guided Vehicles), and other light electric vehicles. This project is to study the feasibility of driving wheel hub motors wirelessly for increased safety and reliability. The work involves a literature review and preliminary study to clarify the power and control requirements of wheel hub motors, the design of a new wireless/contactless power supply with wireless control capability, circuit simulation, and practical testing.

Location: Power Electronics (405.614, Lab)

LLMs (Large Language Models), such as ChatGPT,  are technologies with unprecedented text analysis capabilities. Despite the performance of LLMs is continuously improving, they still suffer from hallucinations and somehow limited capabilities when performing complex reasoning tasks. To explain LLMs results and perform complex reasoning tasks in specific domains, it is possible to adopt an Ontology. In general, Ontologies represent knowledge in a standard form using description logic. An ontology includes the main concepts and the possible relationships among them. Well established reasoning software can exploit this logic to perform complex deductions and check if statements are logically correct according to the ontology. This process can be used to verify the coherence of LLMs responses. Unfortunately building an ontology for a specific domain and populating it with relevant concepts and relationships is a labour-intensive task for humans.

This research project wants to investigate the idea to use LLM technologies (i.e., prompt engineering) to automatically build and populate an ontology for a specific domain (e.g., financial news, sport news etc.). Starting with a minimal set of concepts the approach will perform an Ontology Learning task looking for a cost-effective and scalable solution for knowledge acquisition. This will enable trusted and verified decision-making support in different domains exploiting LLM. The objective of the project is to design and implement a novel technique for Ontology Learning with LLMs to improve their effectiveness. Rigorous experiments will need to be designed and conducted to evaluate the effectiveness of the proposed approach.

Location: HASEL (405.662, Lab)

Software testing is a costly yet crucial task in software engineering, ensuring confidence in the expected behaviour of implemented software. The automated generation of test cases is a significant and actively researched aspect of software engineering. Search-based software engineering has demonstrated effectiveness in automatically generating test cases, achieving high code coverage, and uncovering faults. EvoSuite, currently the most effective test case generator for Java (available at https://github.com/EvoSuite/evosuite), employs an evolutionary algorithm inspired by Charles Darwin's theory of evolution. This algorithm evolves a population of test cases to enhance coverage and fault detection.

Despite EvoSuite's effectiveness, a substantial gap persists between automatically generated tests and those written by humans, particularly in terms of readability and fault revelation.

The emergence of Large Language Models (LLMs) like ChatGPT has prompted researchers to explore leveraging these models to enhance the effectiveness of search-based test generators such as EvoSuite. However, this is a relatively new research topic, and researchers have only scratched the surface of what LLMs can contribute to improving test case generation. The goal of this research project is to design and explore novel methods for integrating LLMs with EvoSuite. The devised approach will be implemented on top of EvoSuite, which implementation is open-source and publicly available. The effectiveness of test case generation using this combined approach will be compared with previous attempts that integrated search-based test generation and LLMs. Proficiency in Java is essential, as EvoSuite is implemented in Java.

Location: HASEL (405.662, Lab)

Smart Meter penetration across the globe is ever increasing, for example in New Zealand we have close to 72% already. There are several implementations and publications around utilizing Smart meter data for various applications. Alongside, there is also a pick-up of penetration of distributed energy technologies like PV and Storage batteries, which makes the distribution networks bidirectional, influencing changing load profiles named as "California duck curves" and "Indian Giraffe curves"  etc.This project will comprehensively assess and classify all possible applications/implementation potential for Smart Meter Data for Electricity Distribution Network utilities and/or retailers.  These will include both real-time, online and time-insensitive applications spanning monitoring, operation, maintenance and trading applications.For handling the large scale of data, big-data analytics will need to be assessed and for smart grid and self-learning algorithm development all possible Artificial Neural Network (ANN) techniques will need to be investigated.The Power systems Group (PSG) have close working relationship with the various retailers and distribution utilities of New Zealand. Thus, once the project group identifies a few use-cases, if needed getting data and industry engagement will not be an issue.  The PSG PhD researchers working in Smart Grid analytics will be closely working with this project group.

Location: Power Systems (405.628, Lab)

It is common for many of us to change the way we speak so that voice-activated systems (such as Siri, Alexa) understand us! This is called bias in speech technology, where the technology works best only for speakers of well-resources languages such as American English. For speakers of other languages/accents such as Ne Zealand English, the technology does not work as well.

As voice-activated systems become increasingly integral to our daily lives, understanding and mitigating bias in these technologies is paramount. This project aims to uncover the hidden biases present in speech technology, utilising a multidimensional approach that encompasses linguistic, cultural, and demographic factors.

Firstly, we will try to define what bias in speech technology is (we all know that the technology is biased, but can we define what bias is!). Secindly, we will propose aproaches to reduce bias in speech technology.

The skills needed for the project are a vision contribute to fair technology development using your engineering education, coding skills (R, Python), basics of machine learning and the willingness to learn. Intrest in linguistics and signals (such as music) is a bonus!

The project will contribute valuable insights to the ongoing discourse on ethical AI and inclusive technology development.

Location: Control Systems (405.722, Lab)

IoT sensors are widely spread and collecting enormous amount of data. However, sensor data may contact noises caused by different factors, which will impact the data processing steps later on for providing intelligent services. This project aims to explore the existence and the type of anomalies (i.e. abnormal data) in IoT sensor time series, and investigate the potential to identify and classify them.

Location: Embedded Systems (405.760, Lab)

Recent developments in transportation wireless power systems are focused on high power stationary systems suitable for fast charging stations for light, medium and heavy duty vehicles. This is placing considerable strain on the primary converter that typically uses a number of switches (within a full bridge inverter) to convert high voltage DC and current to AC through a tuned resonant network made up of inductors and capacitors operating at 85 kHz.

In particular the output bridge inductor is placed under considerable stress as it is usually placed within the inverter cabinet – needs to be as small and light weight as possible but must be designed to handle high volt-amps. For the system designs typically used, a tuned L-C-L network is used at the output of the inverter. Under perfect tuning the bridge inductor (the first L),  ideally has a volt-amp rating that is the same as the kW rating of the supply. However in wireless charging applications, misalignments between the ground side magnetics and the vehicle magnetics along with battery voltage variations, mean that up to 50% increase in volt-amps is needed to handle all potential operating conditions.

At present 50-70kW systems are being used and the inductor is rated to 100kVA in these systems, but we are needing to scale up to 100kW, meaning a target of 150kVA is necessary.

This project requires  evaluation of the present design, and re-evaluation of potential new designs for the bridge inductor. PLECs and ANSYS are needed to establish the current and flux ratings of the inductor core to consider wiring and gapping arrangements in the core and how this impacts the internal and surrounding flux. Either ferrite or nanocrystalline cores with the best Litz wire available should be considered.

The project will establish a new high power inductor design that will be built and tested at 100kW in the high power labs under controlled guidance.

Location: Power Electronics (405.614, Lab)

In this project you will work to develop and optimise a supercapacitor energy management system for energy harvesters. Multiple energy harvesting sources feed into this system to extract ambient energy in whichever possible form to then store and distribute energy purely using supercapacitors. The target is low power IoT applications that require power in the mili to microwatt range.

Supercapacitors, renowned for their high power density and rapid charge-discharge cycles, form the cornerstone of our endeavor. They offer a promising avenue for efficient energy storage in systems reliant on energy harvesting.

The energy stored in supercapacitors should be effectively managed to maximize the performance, reliability, and longevity of energy harvesting systems. The ultimate goal is to pave the way towards sustainable and autonomous energy solutions capable of powering the technologies of tomorrow.

Advanced algorithms and control strategies tailored to the unique characteristics of supercapacitors will be developed as part of this project. Through meticulous research and experimentation, the full potential of these energy storage devices are expected to be unlocked.

Location: Green Electronics (405.712, Lab)

STEM is a hot topic among parents and communities where there is a lot of focus on promoting Science, Technology, Engineering, and Mathematics (STEAM) among the future generation. The predecessor of this project has developed a series of activities, which target high school students. These activities are based around computer hardware, allowing students to investigate the backend of computing devices and hardware programming. In this year's project we aim to investigate the effectiveness of these activities by taking them to high schools and collect data. Ethics approval has been granted for this project already.

Location: Green Electronics (405.712, Lab)

Colours, as music, affect our mood and are associated with emotions. For example, we say: “I feel blue” when we are sad. Certain aspects of colour emotions association are universal and probably have biological origins. However, other aspects of colour emotion association depend on culture. For example, while green symbolizes happiness for Muslims, this colour corresponds to envy in English culture and can have negative connotations. Moreover, people of the same culture relate colours to mood and emotions differently. These differences are due to specific experience of individuals with colours and can arise from traumatic experience with a certain colour. Therefore, the colours that may help to calm down one person might yield strong negative emotions in another individual.

The aim of this project, in collaboration with researchers from Optometry & Vision Science, is to introduce colours into digital apps for treating anxiety and depression. This involves developing different tools for colour selection (colour pickers), developing suitable platform independent interfaces, and investigating how colour display and perception is affected by device parameters and environmental conditions.

Location: Green Electronics (405.712, Lab)

Compressive Sensing (CS) is a relatively new signal processing technique that challenges the conventional Shannon-Nyquist theorem, asserting that signals can be accurately reconstructed from significantly fewer samples than dictated by the traditional sampling rate. Compressive sensing exploits the inherent sparsity of many real-world signals, where only a small fraction of the data carries significant information. This is achieved through a process of random sampling and reconstruction through optimization algorithms.

Compressive sensing involves concepts from several subjects and is at the intersection of various fields, including mathematics, signal processing, and computer science. Mathematically, it involves linear algebra and domain transformations, while in computer science, the use of efficient algorithms is required for signal reconstruction and computation. This multidisciplinary nature requires a nuanced understanding of different concepts. However, the complexity of CS concepts, coupled with its mathematical foundation, often renders it less accessible to undergraduate students. The mathematical intricacies involved in optimization problems, theorems related to sparse signal representations, and the novel approach to sampling contribute to a steep learning curve. The lack of emphasis on the visual and graphical interpretation of CS in undergraduate education further compounds the challenge, making it difficult for students to develop an intuitive understanding of the concepts.

Recognising these challenges, there is a compelling need to create an educational aid that leverages the graphical and visual nature of Compressive Sensing. By developing visual and graphical representations, students can more intuitively grasp the core concepts of CS, bridging the gap between theoretical understanding and practical application. The educational aid aims to provide a tangible tool for educators to enhance the accessibility of CS, making it a more engaging and comprehensible topic for undergraduate students and postgraduate students.

Location: Control Systems (405.722, Lab)

The project advocates for a transformative method in livestock monitoring by marrying computer vision with intelligent wearable devices. It aims to circumvent the constraints posed by traditional sensor-reliant systems and the prohibitive cost of hardware, suggesting a method for precise behavior tracking and categorization in cows, such as grazing and ruminating, via a computer vision model. This methodology is designed not just to advance behavior monitoring technologies but to also refine predictions of feed intake, conversion efficiency, and methane emissions, thereby offering an economical and scalable alternative to enhance livestock management and promote ecological sustainability.

Location: Robotics (405.652, Lab)

Online developer forums such as StackOverflow.com (SO) have transformed how code is written by fostering a culture of searching and reusing code snippets. While various techniques have emerged to facilitate the reuse and analysis of SO code snippets, the automatic testing of these snippets remains an unexplored challenge. This research project seeks to address this gap by designing, implementing, and evaluating a novel automated technique specifically tailored for testing SO code snippets. The objective is to understand the reliability and quality of code snippets sourced from SO, contributing to the overall improvement of software development practices.

Location: HASEL (405.662, Lab)

This project explores cost effective and user friendly technologies to identify and track the location of assets, for example packages in a warehouse, equipment used in industry/universities, etc. The key focus will be to minimize the need for costly infrastructure needed for localization, while maximizing usability. The research will compare the use of Bluetooth beacons against using indoor localization methods with barcode scanning as well as other feasible methods to track assets in an indoor setting. The project seeks to identify the most effective and practical solutions for accurate tracking and management of assets. The research will also involve conducting user surveys to understand current challenges, practices for asset tracking, and user preferences.To demonstrate the viability of the proposed solution, a laboratory asset management system will be implemented. This system will be designed to track equipment used in the research laboratories at the ECSE department.

Location: Green Electronics (405.712, Lab)

There is a widening gap between available health funding and mental health needs both globally and also in New Zealand. Digital health is an opportunity for addressing this widening gap. Phones, wearables, and deep learning, can effectively detect, prevent, and manage health issues. This is also consistent with UN’s “Be He@lthy and Be Mobile”. Current digital methods of monitoring mood are often subjective.

AI-based modelling for the classification of wellness/mood states is a hot-topic in mental health. Funded by the New Zealand Health Research Council, we have a developed a platform called MoodAI jointly with Dr. Fred Sundram (FMHS) and Dr. Amy Chan (FMHS). MoodAI is designed to collect both active and passive data from individuals using Fitbits (heart rate, movement, sleep data) and mobile devices (speech data). This data is then used for mood state classification.

An initial dataset has been collected for healthy participants (n=19) and a depressed sample (n=2) with more anticipated to come onboard this year.

The objectives of this project will be to:

Please add

• To work alongside primarily our post-doc (Luman) and other members of the team on several stages of data curation and processing
• To pre-process acquired data – analyse quality of dataset, identify gaps/missing data/completeness of data, consider methods to either exclude missing data or include the data with certain provisions
• Once the data quality considerations have been explored, to consider methods to meaningfully combine the actively and passively acquired data which are of different modalities (physiological, physical activity, sleep and speech) for further analysis
• To develop initial explainable AI models with our MoodAI dataset and related datasets (MODMA)
• To produce relevant reports/summaries of the various data curation stages and work completed
• To evaluate already developed models and propose new models

Location: Radio Systems (405.736, Lab)