2020 HONOURS PROJECTS - School of Information Technology Updated: September 2019 - Deakin University
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
School of Information Technology 2020 HONOURS PROJECTS Updated: September 2019 Deakin University CRICOS Provider Code: 00113B
School of Information Technology
2020 HONOURS PROJECTS
Table of Contents
Code2Vec: learning representations of code using machine learning ............................................................................................. 1
IoT device hacking using SDR ............................................................................................................................................................ 2
Model Driven Engineering of Blockchain systems ............................................................................................................................ 3
Software engineering of AI systems ................................................................................................................................................. 4
Software engineering of AI systems ................................................................................................................................................. 5
Developing engaging digital health platform for elderly .................................................................................................................. 6
Crowdsourcing of traffic incident reporting ..................................................................................................................................... 7
Health apps on Temi ......................................................................................................................................................................... 8
Spatial‐Temporal Data visualisation ................................................................................................................................................. 9
EEG Controlled Robots .................................................................................................................................................................... 10
Self‐learning Swarm intelligence based cyber defense for IoT....................................................................................................... 11
Mesh network for Automotive Industry ......................................................................................................................................... 12
Automated Planning for Smart Transport ...................................................................................................................................... 13
Cyber threat analysis for critical infrastructures through Bayesian classification.......................................................................... 14
Adaptive Game Design for Individualised Experiences................................................................................................................... 15
Using Anonymised Avatars in Video Games for Improved Cooperative Behaviour ....................................................................... 16
Innovative Interaction Methods in Virtual Reality.......................................................................................................................... 17
The Effect of Mixed‐Input Interaction on Collaborative Behaviour in Virtual Reality .................................................................... 18
Deakin University CRICOS Provider Code: 00113B TABLE OF CONTENTS
School of Information Technology
2020 HONOURS PROJECTS
Augmented reality intent capture .................................................................................................................................................. 19
Speech to content ........................................................................................................................................................................... 20
Temporally agnostic virtual learning environment ......................................................................................................................... 21
Augmented Reality OpenXR support .............................................................................................................................................. 22
Network Topology Inference in Swarm Robotics ........................................................................................................................... 23
Goal Reasoning for Robotic Swarms ............................................................................................................................................... 24
Evolving Tailor Made Swarm Controllers ........................................................................................................................................ 25
Digital signatures based on the Rabin cryptosystem ...................................................................................................................... 26
DWL : Identity preservation using blockchain for real world uses (Blockchain and security)........................................................ 27
DWL : Applied costs and benefits study of crypto currencies in everyday life (Blockchain) .......................................................... 28
DWL : Use of AI and chatbots for mental disorders ....................................................................................................................... 29
DWL : IOT ‐ Beacons tracking system using beacons technology ................................................................................................... 30
Robot control using deep interactive reinforcement learning ....................................................................................................... 31
The kingdom of the 7‐segments: Drones as Moving Sculpture ...................................................................................................... 32
Fog Computing Clusters .................................................................................................................................................................. 33
Flying Fog: Drones as Fog Computing Resource Nodes .................................................................................................................. 34
Using state charts to develop maintainable IoT apps ..................................................................................................................... 35
Designing a Trust Management Framework for the Internet of Things ......................................................................................... 36
Developing Anomaly Detection Models for Detecting Sensor Faults in Internet of Things ........................................................... 37
Design of a Multi‐sensor IoT‐enabled Device for Discrete and Outdoor Deployable Gait Monitoring ......................................... 38
Deakin University CRICOS Provider Code: 00113B TABLE OF CONTENTS
School of Information Technology
2020 HONOURS PROJECTS
Developing a Privacy‐Preserving Consensus Algorithm in Blockchain ........................................................................................... 39
Generative Adversarial Nets for Intrusion Detection in Internet‐of‐Things ................................................................................... 40
Cognitive Embodiment for Tactical Training in Team Sports.......................................................................................................... 41
Remote Collaboration Techniques using Extended Reality ............................................................................................................ 42
Emotional Embodiment for Behavioural Change ........................................................................................................................... 43
Augmented Reality Multipresence ................................................................................................................................................. 44
Automated threat modelling for enterprises ................................................................................................................................. 45
Supporting Internet of Things (IoT) Swarms ................................................................................................................................... 46
Conflict Resolution for Automated Smart‐Cars .............................................................................................................................. 47
Collaboration for drones and robots .............................................................................................................................................. 48
Learning camera fingerprints from images ..................................................................................................................................... 49
Investigation into adversarial attacks on deep neural networks and the counter measures ........................................................ 50
Automatic Detection of Deep Fakes ............................................................................................................................................... 51
Anomaly Detection in Multilayer Network Data via Knowledge Graph ......................................................................................... 52
Fog Computing for Smart e‐Health Systems ................................................................................................................................... 53
Fog Computing for VR/AR Systems ................................................................................................................................................. 54
Software Development and Testing for Microservices based Fog Computing Systems ................................................................ 55
Discovering Catalytic Events in Soccer Games ................................................................................................................................ 56
Predicting Sensitivity of Insurance Products................................................................................................................................... 57
Data Fusion and External (Environmental) Situation Awareness for Autonomous Underwater Vehicles..................................... 58
Deakin University CRICOS Provider Code: 00113B TABLE OF CONTENTSSchool of Information Technology
2020 HONOURS PROJECTS
Knowledge Graph for Adaptive Education...................................................................................................................................... 59
Internal Situation Awareness (SA) and Fault Tolerance for Autonomous Underwater Vehicles ................................................... 60
Developing a Motion Planning System for Underwater Vehicles ................................................................................................... 61
Application of Image Processing and Deep Learning on the Marine Environment Monitoring and knowledge Extraction ......... 62
Sequential anomaly detection in cyber security using deep reinforcement learning.................................................................... 63
Mobile edge caching using deep reinforcement learning .............................................................................................................. 64
Cloud‐based malware detection using deep reinforcement learning‐based offloading process .................................................. 65
Security and Privacy Issues of Blockchain‐based Applications ....................................................................................................... 66
A Novel Wireless Edge Analytics Scheme for Virtual and Augmented Reality ............................................................................... 67
Develop a novel multipath data scheduling for future manufacturing .......................................................................................... 68
Can Machine Learning be used to detect Traffic Anomaly at the SDN Edge? ................................................................................ 69
Analysis of Eye Fixation Data .......................................................................................................................................................... 70
Deep Learning for Insider Threat Detection ................................................................................................................................... 71
Software Engineering for Artificial Intelligence (“SE for AI”) .......................................................................................................... 72
Construct, Solve, Merge and Adapt for Car Sequencing................................................................................................................. 73
Optimisation of attack trees ........................................................................................................................................................... 74
Fairness Protocols for Trust Establishment in the IoT .................................................................................................................... 75
Facilitating Collaboration Using VR/AR ........................................................................................................................................... 76
Context Provider/Service Annotation for Context‐as‐a‐Service IoT Platform ................................................................................ 77
Context Provider/Service Selection for Context‐as‐a‐Service IoT Platform ................................................................................... 78
Deakin University CRICOS Provider Code: 00113B TABLE OF CONTENTSSchool of Information Technology
2020 HONOURS PROJECTS
Context Service discovery, composition, and chaining for IoT platform CoaaS ............................................................................. 79
User‐centric interface for smart context provider registration in Context‐as‐a‐Service IoT platform .......................................... 80
CASE tool for developing SLAs between context providers and Context‐as‐a‐Service IoT platform ............................................. 81
Context‐as‐a‐Service IoT platform performance dashboard .......................................................................................................... 82
Smart Context Definition Query Language Editor V2 ..................................................................................................................... 83
Turning Evil to Good: Protect IP of Deep Neural Networks with Backdooring .............................................................................. 84
The School of Information Technology also offers students the opportunity to propose their own research projects, or to suggest modifications to
the suggested projects above, so as to better align to individual learning objectives.
Deakin University CRICOS Provider Code: 00113B TABLE OF CONTENTSSchool of Information Technology 2020 HONOURS PROJECTS Code2Vec: learning representations of code using machine learning Supervisor: A/Prof Mohamed Abdelrazek Email: mohamed.abdelrazek@deakin.edu.au Campus: Burwood Start Date: January or July In order to adopt machine learning in software engineering, programming, vulnerability analysis, etc. we need to be able to convert code into numerical representation – e.g. vector that we can then use as input to many of the existing machine learning models – e.g. clustering, SVM, CNN, LSTM, etc. There are many word embedding techniques available, however source code introduce unique attributes that do not exist in natural language (text) – e.g. code has hierarchies – nested code, etc. which make these techniques less effective. Hence, there is now research in how to create code2vec techniques that best capture code representation and thus be able to apply these representations to develop many relevant software engineering use cases. This project aims at understanding the current state‐of‐the‐art in code2vec, benchmark existing techniques and possibly optimise some of these, and develop ML models for one or two applications – e.g. vulnerability analysis, code recommender systems, intent understanding (e.g. auto generate documentation from code). It is a very exciting project that has many potential applications, and can potentially lead to a PhD project. Necessary Skills: Strong programming skills, good understanding of machine learning (or at least interest to learn). Deakin University CRICOS Provider Code: 00113B Page 1
School of Information Technology 2020 HONOURS PROJECTS IoT device hacking using SDR Supervisor: A/Prof Mohamed Abdelrazek Associate Supervisors: Dr Amani Ibrahim, Dr Alessio Bonti Email: mohamed.abdelrazek@deakin.edu.au Campus: Burwood Start Date: January or July According to recent articles, we expect billions of devices to be in use. However, it is very challenging to assess the security of these sensors/devices. In this project, we want to conduct a detailed threat modelling of IoT systems, identify the set of tools necessary to conduct security assessment of IoT devices, and then use software defined radio to assess the security of existing IoT devices against different variations of replay attacks. The ultimate goal is to create an IoT hacking lab/infrastructure that can then be used for consulting purposes for companies that want to use IoT. The team has SDR kit that will be used during the project. Necessary Skills: Strong programming skills. Good understanding of cyber security and IoT protocols. Deakin University CRICOS Provider Code: 00113B Page 2
School of Information Technology 2020 HONOURS PROJECTS Model Driven Engineering of Blockchain systems Supervisor: A/Prof Mohamed Abdelrazek Email: mohamed.abdelrazek@deakin.edu.au Campus: Burwood Start Date: January or July This project aims at using model‐driven engineering to (semi)auto generate blockchain systems. The idea is how to capture blockchain system details in a text/model/diagram‐like specification that can then be used to automatically generate the necessary system configuration and blockchain setup, and all supporting APIs that can then be used to develop business functionalities without having to understand or learn how to use blockchain technology. Necessary Skills: Strong programming skills. Good understanding of blockchain (or at least interest to learn). Deakin University CRICOS Provider Code: 00113B Page 3
School of Information Technology 2020 HONOURS PROJECTS Software engineering of AI systems Supervisor: A/Prof Mohamed Abdelrazek Email: mohamed.abdelrazek@deakin.edu.au Campus: Burwood Start Date: January or July This project aims at understanding how to capture and model software requirements in AI/ML based systems. Most of the existing requirements specification techniques are mainly designed for deterministic and well‐defined systems and do not address challenges in AI systems. In this project, we will investigate how to specify AI requirements and how to develop UX/UI of AI systems. Necessary Skills: Strong programming skills. Good understanding of Machine Learning. Deakin University CRICOS Provider Code: 00113B Page 4
School of Information Technology 2020 HONOURS PROJECTS Software engineering of AI systems Supervisor: A/Prof Mohamed Abdelrazek Associate Supervisors: Dr Alessio Bonti, Prof Phil Rays Email: mohamed.abdelrazek@deakin.edu.au Campus: Burwood Start Date: January or July Mental health is becoming a major issue in work environments. According to recent studies heart attack was 11 percent higher on Mondays than other days in the week. In this project, we aim to develop a mental health sensing app that learns from user behaviour potential mental health behaviour biomarkers issue that might be developing and potential introduce necessary interventions. Necessary Skills: Strong programming skills. Good understanding of Machine Learning. Deakin University CRICOS Provider Code: 00113B Page 5
School of Information Technology 2020 HONOURS PROJECTS Developing engaging digital health platform for elderly Supervisor: A/Prof Mohamed Abdelrazek Associate Supervisor: Dr Alessio Bonti Email: mohamed.abdelrazek@deakin.edu.au Campus: Burwood Start Date: January or July Hundreds of health apps and wearables are currently in the market. However, the adoption of the technology by elder people is very limited. In this project, we want to develop engaging digital health platform that takes into consideration user emotions as first class citizen when engineering the platform. Necessary Skills: Strong programming skills. Deakin University CRICOS Provider Code: 00113B Page 6
School of Information Technology 2020 HONOURS PROJECTS Crowdsourcing of traffic incident reporting Supervisor: A/Prof Mohamed Abdelrazek Associate Supervisor: Dr Alessio Bonti Email: mohamed.abdelrazek@deakin.edu.au Campus: Burwood Start Date: January or July In this project, we aim to develop a new crowd sourcing platform for traffic incident reporting. The idea is to develop a mobile app that can record traffic incident in real time and report this to authority for investigations – imagine being able to report a car travelling without number plate, misbehaviour on the road, accidents, or in case of terrorism attacks being able to share videos around the place with the police. This needs to be easy to use – e.g. using voice commands to activate the recording. Necessary Skills: Strong programming skills. Deakin University CRICOS Provider Code: 00113B Page 7
School of Information Technology 2020 HONOURS PROJECTS Health apps on Temi Supervisor: A/Prof Mohamed Abdelrazek Associate Supervisor: Dr Alessio Bonti Email: mohamed.abdelrazek@deakin.edu.au Campus: Burwood Start Date: January or July The team now has access to Temi robot. The idea is to develop digital health platform that assist elder people living alone to keep up and connect with care givers and GPs via Temi. Temi supports video calls, follow‐me, and can run apps. Necessary Skills: Strong programming skills. Deakin University CRICOS Provider Code: 00113B Page 8
School of Information Technology 2020 HONOURS PROJECTS Spatial‐Temporal Data visualisation Supervisor: A/Prof Mohamed Abdelrazek Associate Supervisor: Dr Alessio Bonti Email: mohamed.abdelrazek@deakin.edu.au Campus: Burwood Start Date: January or July This project aims to develop spatial‐temporal data visualisation tool that uses maps (indoor & Google maps) to apply interactive visualisation layers – e.g. vehicle tracking, area congestion modelling, etc. The user should be able to interact with the visualisations – e.g. drill down, roll over, zoom in and out, double click, etc. Necessary Skills: Strong programming skills. Deakin University CRICOS Provider Code: 00113B Page 9
School of Information Technology 2020 HONOURS PROJECTS EEG Controlled Robots Supervisor: A/Prof Mohamed Abdelrazek Email: mohamed.abdelrazek@deakin.edu.au Campus: Burwood Start Date: January or July The idea of this project is to develop an understanding of how EEG signal works, how to measure and interpret using ML models, and how to use the signal to take actions. There are many applications of using EEG in entertainment and games industry as it helps to free user hands and make games/VR/etc more intuitive and engaging. It also can be used in hospitals to help doctors to grab information they need in real time. In this project we want to investigate if it would be possible to use EEG signals to control a group of robots and get them to achieve a given task. Necessary Skills: Strong programming skills, experience in interfacing with hardware and robots. Deakin University CRICOS Provider Code: 00113B Page 10
School of Information Technology
2020 HONOURS PROJECTS
Self‐learning Swarm intelligence based cyber defense for IoT
Supervisor: Dr Adnan Anwar Associate Supervisor: Dr Zubair Baig
Email: adnan.anwar@deakin.edu.au
Campus: Waurn Ponds
Start Date: January or July
Internet of Things (IoT) is gaining its popularity because of its enormous potential across a wide range of applications. However, one major challenge of IoT
adoption is to ensure its security that includes the processes, technologies, and measures necessary to ensure the protection of IoT devices as well as the
communication network. Swarm intelligence which has been applied successfully in different application area, can be used for designing an intrusion detection
system. The autonomous and self‐learning property will enhance its capability towards designing a sophisticated cyber defense mechanism for IoT. Key objective of
this project is as follow:
1. Model Swarm intelligence
2. Enhance the model with autonomous and self‐learning capability
3. Develop IDS using intelligent swarm technique
4. Gather IoT attack dataset
5. Apply the developed IDS on IoT dataset
6. Testing and validation
7. Reporting based on the obtained results.
Necessary Skills:
Software skills in Python or C/C++, Matlab.
Deakin University CRICOS Provider Code: 00113B Page 11School of Information Technology 2020 HONOURS PROJECTS Mesh network for Automotive Industry Supervisors: Dr Iman Avazpour, Dr Niroshinie Fernando Email: iman.avazpour@deakin.edu.au, niroshinie.fernando@deakin.edu.au Campus: Burwood Start Date: January The software systems in cars and transport vehicles are becoming more complex. Manufacturers have to employ complex array of sensors and actuators spread across a very small realestate inside vehicle. Multiplex systems have helped to reduce the complexity by providing a common bus for transfer of information between sensors, actuators and vehicle’s computers. However, there is still requirement for all the parts to be physically and electronically connected in most vehicles. This project seeks to investigate and develop small mesh networks purposefully build for vehicle sensor/actuator communication. The network is required to not only provide facilities for communication between sensors, actuators, and a vehicles ECU, but also provide means for vehicle to vehicle, and vehicle to base communication. Necessary Skills: Understanding of networking protocols, basic programming knowledge, experience with IoT small devices like particle/raspberry pi. Deakin University CRICOS Provider Code: 00113B Page 12
School of Information Technology 2020 HONOURS PROJECTS Automated Planning for Smart Transport Supervisors: Dr Iman Avazpour, Dr Kevin Lee Email: iman.avazpour@deakin.edu.au, kevin.lee@deakin.edu.au Campus: Burwood, Cloud Start Date: January or July In the next few years, smart vehicles are going to appear on the public transport system. These, vehicles include freight trucks, smart cars and even automated delivery drones will be capable of navigating from one location to the destination without human intervention. They will have to navigate traffic problems, hazards and uncertainties. This project will investigate and develop algorithms to support this automated planning process. It is expected that it will involve the integration of hardware, software, simulation and the use of planning and optimisation algorithms. Depending on progress in the project, it is anticipated that the work will be published in a distributed systems or software engineering conference. Necessary Skills: Arduino/RPI/sensors hardware. Event‐based programming such as MQTT or ROS. Any BSE student should have the necessary skills. Deakin University CRICOS Provider Code: 00113B Page 13
School of Information Technology
2020 HONOURS PROJECTS
Cyber threat analysis for critical infrastructures through Bayesian classification
Supervisor: Dr Zubair Baig Associate Supervisor: Dr Adnan Anwar
Email: zubair.baig@deakin.edu.au
Campus: Waurn Ponds
Start Date: January or July
Critical infrastructures play a critical role in our daily lives. Energy, water and the Internet infrastructure are all such infrastructures whose role in our lives cannot be
understated. The protection of valuable critical infrastructure assets (both tangible and non‐tangible) from cyber threats, is essential for its smooth functioning.
Bayesian networks are probabilistic graph models that help establish conditional model dependences in data. Through this project, the task of applying Bayesian
networks and its variants to cyber security data sets of the critical infrastructure, will be accomplished.
In particular, the following phases will comprise this project:
1. Risk assessment of a critical infrastructure as decided between the student and the supervisors. E.g., Water, Electricity, Border Security
2. Analysis of 2‐3 variants of Bayesian networks and their application to cyber security datasets, for anomaly detection
3. Implementation of the Bayesian networks for designing the anomaly detection system
4. Testing the Bayesian network classifier on the cyber security datasets of a critical infrastructure of choice
5. Running tools including GNUPlot, to prepare visualisations of the results obtained
6. Reporting on the findings obtained
Necessary Skills:
Software skills in Python or C/C++.
Deakin University CRICOS Provider Code: 00113B Page 14School of Information Technology
2020 HONOURS PROJECTS
Adaptive Game Design for Individualised Experiences
Supervisor: Dr Alexander Baldwin
Email: alexander.baldwin@deakin.edu.au
Campus: Burwood
Start Date: January or July
Video games are used for a variety of purposes including entertainment, education, health, immersion and affect. However, a barrier in achieving optimal player
experience is the individualised nature of how players will choose to interact with the game and their subjective experiences and preferences. This is a significant
challenge as designers and developers are not able to fully tailor an experience to one demographic without potential negative effects on another.
Adaptive game design can hold the key to improving player experience through allowing games to:
1. measure player experience or interactions;
2. adjust elements of the design in response to provide a better match for that unique player.
This project would seek to investigate and measure new methods of improving player experience through adaptive game design. This would build off the
foundations of existing research in areas such as Dynamic Difficulty Adjustment and MDDA (Multiplayer Dynamic Difficulty Adjustment). During this project you may
engage in:
• qualitative and quantitative data collection and analysis;
• prototyping;
• A/B testing;
• formal research methodologies;
• psychological player experience measurement.
Necessary Skills:
Prior study or experience with game design or development. Experience with Unreal Engine 4 or Unity for potential prototyping.
Deakin University CRICOS Provider Code: 00113B Page 15School of Information Technology
2020 HONOURS PROJECTS
Using Anonymised Avatars in Video Games for Improved Cooperative Behaviour
Supervisor: Dr Alexander Baldwin
Email: alexander.baldwin@deakin.edu.au
Campus: Burwood
Start Date: January or July
Human‐to‐human interactions are highly influenced by biases and assumptions derived from identification of traits such as gender, ethnicity, nationality, age and
more. However, virtual environments can allow the use of player avatars that differ in appearance from the user in control and expose or hide these traits.
It is suspected that anonymising these traits may have the effect of improving cooperative behaviour in problem‐solving tasks, with the potential to be transferrable
to subsequent real‐world interactions, such as improving pro‐social behaviours.
This project will involve the use of anonymised avatars in a multiplayer video game to determine the effect of anonymity on cooperative behaviour in school‐age
children.
During this project you may engage in:
• cross‐disciplinary research;
• qualitative and quantitative data collection and analysis;
• formalised research methodologies.
Necessary Skills:
Prior study or experience with game or interaction design. Experience with game development software would be beneficial.
Deakin University CRICOS Provider Code: 00113B Page 16School of Information Technology
2020 HONOURS PROJECTS
Innovative Interaction Methods in Virtual Reality
Supervisor: Dr Alexander Baldwin
Email: alexander.baldwin@deakin.edu.au
Campus: Burwood
Start Date: January or July
Virtual Reality (VR) holds significant potential for highly immersive experiences. However, input devices and methods of interaction for virtual reality are still not
standardised as new discoveries are made and new use cases are discovered or proposed.
In this project, new methods of interaction for virtual reality will be investigated and prototyped for the purpose of enabling new experiences for particular
purposes such as games and entertainment, training, users with limited physical mobility, etc.
Potential types of interactions may include:
• locomotion (navigation);
• object interaction;
• social interaction in virtual reality.
During this project you may engage in:
• prototype development;
• qualitative and quantitative data collection and analysis;
• user experience (UX);
• formalised research methodologies.
This project will involve an element of prototyping using game engine technologies such as Unity or Unreal Engine 4.
Necessary Skills:
Prior study or experience with game or VR development. Experience with Unreal Engine 4 or Unity.
Deakin University CRICOS Provider Code: 00113B Page 17School of Information Technology
2020 HONOURS PROJECTS
The Effect of Mixed‐Input Interaction on Collaborative Behaviour in Virtual Reality
Supervisor: Dr Alexander Baldwin
Email: alexander.baldwin@deakin.edu.au
Campus: Burwood
Start Date: January or July
Multi‐user virtual reality is highly sought after for entertainment, business and education purposes. However, the range of different input devices and their
limitations as well as access to high‐fidelity VR equipment often means that different users in a virtual environment will not all be able to interact with the same
methods. This has the potential to influence collaborative behaviour as the range of communication methods such as voice, gesture, body language, facial
expression, etc will vary from user to user.
This research will investigate the influence of mixed input between users in a multi‐user environment on collaborative behaviours, presence and user experience.
During this project you may engage in:
• qualitative and quantitative data collection and analysis;
• user experience (UX);
• formalised research methodologies.
Necessary Skills:
Prior study or experience with game or VR design.
Deakin University CRICOS Provider Code: 00113B Page 18School of Information Technology 2020 HONOURS PROJECTS Augmented reality intent capture Supervisor: Dr Shaun Bangay Email: shaun.bangay@deakin.edu.au Campus: Burwood Start Date: January or July Existing motion capture technologies capture pose, motion, and externally visible features (such as facial expressions). Motion capture while encumbered with a virtual reality headset is also an area of recent active research. Augmented reality experiences typically avoid the need to modify the participants, but still incorporate tracking facilities that could readily be adapted for motion capture purposes. However, all of these processes capture the particular actions undertaken, rather than the intent behind them. Agents reconstructed from these recordings are unable to adapt, and mindlessly repeat the same actions regardless of changes in their environment and circumstances. The next phase of motion capture is to record the intent behind the actions. This research project aims to record participants for the purpose of constructing adaptive (intelligent) agents that would provide meaningful proxies for those participants in their absence. One potential strategy would be to combine recorded actions with a “deepfakes” process to adapt to new situations. Another strategy could involve inferring a model of the hidden mechanisms that would drive the behaviour of the agent. These processes would be evaluated with respect to their ability to accommodate changes in their environment, and to respond in a way consistent with the individual they are recorded from. Necessary Skills: Experience with the environmental and participant capture processes utilized in SIT383 would be an advantage, as would familiarity with topics in artificial intelligence. Deakin University CRICOS Provider Code: 00113B Page 19
School of Information Technology 2020 HONOURS PROJECTS Speech to content Supervisor: Dr Shaun Bangay Email: shaun.bangay@deakin.edu.au Campus: Burwood Start Date: January or July Our recent research has identified that a significant requirement for widespread use of virtual reality is the ability for individual specialists to be able to author their own experiences. Existing technologies address this in a mundane fashion, converting spoken descriptions into feature descriptors that can be used to retrieve standard media types from online databases. The strategy proposed with this project is to make use of procedural content generation and modification techniques to support advanced virtual reality authoring facilities. These extend beyond just populating a scene by allowing control over the behaviour, interactions, scenarios and user experience. The research process required for this project involves identifying one specific authoring strategy, adapting and applying a procedural generation strategy, and evaluating the performance and constraints under challenging test conditions. Necessary Skills: Experience with virtual reality or game application development is a requirement. Familiarity with virtual reality experience design, or a particular experience authoring context would be an advantage. Deakin University CRICOS Provider Code: 00113B Page 20
School of Information Technology
2020 HONOURS PROJECTS
Temporally agnostic virtual learning environment
Supervisor: Dr Shaun Bangay
Email: shaun.bangay@deakin.edu.au
Campus: Burwood, Cloud
Start Date: January or July
The three factors impeding effective learning:
‐ Having to go to classes. Long travel times or overlapping commitments. You’ve missed the class and now have to put up with a low quality non‐
interactive recording.
‐ The lecturer is never available when you need them. For some reason these people also have lives.
‐ Group work is dreaded. Apart from having to coordinate meetings, some team members don’t contribute or even pitch up leaving you to do the work
of an entire team.
Fortunately, computer science has the answer.
Temporally (time) agnostic virtual worlds provide opportunities to participate in a learning environment at your own schedule. Rather than having to coordinate
directly with others you get to interact with their imposters; agents re‐enacting recorded events. These worlds grow and develop every time someone participates.
Key challenges include devising and evaluating the algorithms and data structures for recording and replaying events in a virtual environment. Criteria include
measuring presence‐fails; the point at which no suitable content is available. This measure should improve the longer the environment is used and provide
significant insight into the costs of authoring lessons in this form of virtual reality experience.
Necessary Skills:
A computer science or information technology background is essential. There are opportunities with this project to focus on the algorithmic principles, and/or on the
components involved in capturing and replaying learning experiences.
Deakin University CRICOS Provider Code: 00113B Page 21School of Information Technology 2020 HONOURS PROJECTS Augmented Reality OpenXR support Supervisor: Dr Shaun Bangay Email: shaun.bangay@deakin.edu.au Campus: Burwood Start Date: January or July The OpenXR 1.0 standard has recently been released by the Khronos Group (https://www.khronos.org/openxr). This provides a hardware agnostic software layer for supporting development of extended reality applications that need to work across multiple devices. This project is intended to access the suitability of this standard for use in augmented reality applications; in particular handheld mobile devices using the facilities provided by ARCore, and pass‐through mixed reality headsets such as the Oculus Quest. This investigation requires proof‐of‐concept implementations of relevant portions of the OpenXR standard for these platforms. Key outcomes include a critical assessment of how well this standard would support current research directions related to custom augmented reality devices and interaction strategies. Necessary Skills: Candidates should ideally have prior experience with developing augmented reality applications (SIT383). Application development experience, such as prior involvement with an open source codebase, is an advantage. Deakin University CRICOS Provider Code: 00113B Page 22
School of Information Technology 2020 HONOURS PROJECTS Network Topology Inference in Swarm Robotics Supervisor: Dr Jan Carlo Barca Email: jan.barca@deakin.edu.au Campus: Burwood Start Date: January or July Swarm robotics refer to the implementation of swarm intelligence features like autonomy and self‐organization to a collective of robots. This work builds on state‐of‐the‐art automated surveillance methods to visually infer topological graphs that reveal communication structures and leadership in robotic swarms. The resulting technology can be used to ensure the safety of airports and other high value sites under attack by robotic swarms. Preliminary work has already been published at a world‐renowned robotics conference and the selected student will be given an opportunity to build on this work. It is therefore expected that output from this research project also will be accepted for publication at a high‐quality venue. Necessary Skills: Programming and mathematical skills are required. Deakin University CRICOS Provider Code: 00113B Page 23
School of Information Technology 2020 HONOURS PROJECTS Goal Reasoning for Robotic Swarms Supervisor: Dr Jan Carlo Barca Associate Supervisor: Dr Kevin lee Email: jan.barca@deakin.edu.au Campus: Burwood Start Date: January or July At present robotic swarms must be told what goals to achieve and how goals can be decomposed into sub‐goals. This constraint is limiting for swarms that perform missions in complex environments when it is not feasible to manually engineer/encode complete knowledge of what goal(s) should be pursued for every conceivable state. This is also a major drawback in situations where actions fail, new opportunities arise or events take place that strongly motivate changing the goals. This exciting project will address the above issue by investigating how intelligent robotic swarms can reason about, formulate, select and manage their goals/objectives autonomously. Algorithms will be developed using Gazebo, the preferred simulator in DARPAs virtual robotics challenge. Depending on time and progress, the work can be evaluated using the Crazyflie drone platform. The selected student will be given an opportunity to publish his/her work internationally if the output is of high quality.. Necessary Skills: C++ programming skills are required. Deakin University CRICOS Provider Code: 00113B Page 24
School of Information Technology 2020 HONOURS PROJECTS Evolving Tailor Made Swarm Controllers Supervisor: Dr Jan Carlo Barca Associate Supervisor: Prof Maia Angelova Turkedjieva Email: jan.barca@deakin.edu.au Campus: Burwood Start Date: January or July At present swarm controllers must be manually designed. This process is time consuming and requires a deep understanding of how local interactions lead to global behaviours. This project addresses this issue by building on state‐of‐the‐art machine learning and evolutionary processes, which makes it possible to automatically generate tailor made swarm controllers from small building blocks extracted from cutting edge algorithms used to operate swarms. This project will focus on generating customized particle swarm optimizers, but the work can be generalized to generate cutting edge swarm controllers for drones, unmanned ground vehicles or unmanned underwater vehicles. The selected student will work with one of the PhD students within the school and will be given an opportunity to publish his/her work internationally, if the output is of high quality. It is highly desirable that the student has a desire to progress into a PhD program. Necessary Skills: Prior experience with MatLab, or a willingness to learn how to use Matlab is required. Deakin University CRICOS Provider Code: 00113B Page 25
School of Information Technology 2020 HONOURS PROJECTS Digital signatures based on the Rabin cryptosystem Supervisor: Prof Lynn Batten Associate Supervisor: Dr Leo Zhang Email: lmbatten@deakin.edu.au Campus: Burwood Start Date: January The Rabin cryptosystem is based on RSA with encryption exponent 2, resulting in a decryption with four possible outcomes. A number of research papers since 1980 have developed ways of determining the original message from these four. The first part of your research would include finding, understanding and explaining how these papers work. The second part of the research involves applications of Rabin signatures in authentication of communications by means of digital signatures. A major topic of interest in digital security today for example, is authenticating communications between Internet of Things devices. Necessary Skills: Familiarity with RSA schemes and the underlying number theoretic results. Deakin University CRICOS Provider Code: 00113B Page 26
School of Information Technology
2020 HONOURS PROJECTS
DWL : Identity preservation using blockchain for real world uses (Blockchain and security)
Supervisor: Dr Alessio Bonti Associate Supervisor: A/Prof Mohamed Abdelrazek
Email: a.bonti@deakin.edu.au
Campus: Burwood
Start Date: January or July
DWL : Deakin Wide Lab is an initiative that aims at creating novel and innovative solutions that will use the University as a test bed for experiments in order to
solve real life problems that can scale up to the world.
Blockchain technologies have allowed for creation of new services, but also for changes in our current ones. The promise of traceable immutable system of records
based on decentralized trust have changed the way we design and envision our future social interactions and business relationships.
This project aims at identifying the benefits and costs of implementing a system that will use blockchain to allow users, groups and companies to safely share
information either without revealing their identity, or by allowing easy to track audited documents.
This will greatly increase the value of data and enhance its protection, allowing the user to fully assess the current state of his privacy and own his data, may he
want to freely disclose it, or sell it
The student will make use of the existing resources, including:
Use and write Smart contracts
Leverage existing infrastructure and relationships
Necessary Skills:
Nodejs or alternative programming language (will require to learn nodejs), learn basic Hyperledger composer.
Deakin University CRICOS Provider Code: 00113B Page 27School of Information Technology
2020 HONOURS PROJECTS
DWL : Applied costs and benefits study of crypto currencies in everyday life (Blockchain)
Supervisor: Dr Alessio Bonti Associate Supervisor: A/Prof Mohamed Abdelrazek
Email: a.bonti@deakin.edu.au
Campus: Burwood
Start Date: January
DWL : Deakin Wide Lab is an initiative that aims at creating novel and innovative solutions that will use the University as a test bed for experiments in order to
solve real life problems that can scale up to the world.
Blockchain technologies have allowed for creation of new services, but also for changes in our current ones. The promise of traceable immutable system of records
based on decentralized trust have changed the way we design and envision our future social interactions and business relationships.
This project aims at identifying the benefits and costs of implementing a parallel economy within the university, understanding the pros and cons of moving
towards a model where the currency is replaced by a localised token (the Deakin Dollar)
The student will make use of the existing resources, including:
A prototype economy
Use and write Smart contracts
Leverage existing infrastructure and relationships
Necessary Skills:
Nodejs or alternative programming language (will require to learn nodejs), learn basic Hyperledger composer.
Deakin University CRICOS Provider Code: 00113B Page 28School of Information Technology
2020 HONOURS PROJECTS
DWL : Use of AI and chatbots for mental disorders
Supervisor: Dr Alessio Bonti Associate Supervisor: A/Prof Mohamed Abdelrazek
Email: a.bonti@deakin.edu.au
Campus: Burwood
Start Date: January or July
DWL : Deakin Wide Lab is an initiative that aims at creating novel and innovative solutions that will use the University as a test bed for experiments in order to
solve real life problems that can scale up to the world.
Mental disorders, including depression, create difficult environments for people to live in. The stigmata associated with them also pushes people to hide having
such problems.
Our chatbot, based on a past successful engagement with a leading organization, will further explore how AI can help to reduce the burden for people and for
society.
The student will make use of the existing resources, including:
Leverage existing infrastructure and relationships
Explore how to train and enahnce chatbots
Identify key metrics to be used in conversation design
Necessary Skills:
Nodejs or alternative programming language (will require to learn nodejs), learn basic Chatbot and conversation design.
Deakin University CRICOS Provider Code: 00113B Page 29School of Information Technology
2020 HONOURS PROJECTS
DWL : IOT ‐ Beacons tracking system using beacons technology
Supervisor: Dr Alessio Bonti Associate Supervisor: A/Prof Mohamed Abdelrazek
Email: a.bonti@deakin.edu.au
Campus: Burwood
Start Date: January or July
DWL : Deakin Wide Lab is an initiative that aims at creating novel and innovative solutions that will use the University as a test bed for experiments in order to
solve real life problems that can scale up to the world.
Indoor tracking, or high resolution geopositioning can be very useful to create new services based on the tracking data of users.
We have developed a prototype infrastructure to track users inside Greenwood Park, using Estimote beacons. We currently use this data for temperature control
and for other projects which leverage human activity.
The student will make use of the existing resources, including:
Leverage existing infrastructure and relationships
Explore the use of Beacons technologies
Create new ways of using the data
Analyze the data to find patterns
Necessary Skills:
Nodejs or alternative programming language (will require to learn nodejs).
Deakin University CRICOS Provider Code: 00113B Page 30School of Information Technology 2020 HONOURS PROJECTS Robot control using deep interactive reinforcement learning Supervisor: Dr Francisco Cruz Associate Supervisor: A/Prof Richard Dazeley Email: francisco.cruz@deakin.edu.au Campus: Waurn Ponds Start Date: January or July Intelligent robots have recently taken their first steps into domestic scenarios. It is thus expected that robots learn to perform tasks, which are often considered rather simple for humans. However, for a robot to reach human‐like performance diverse subtasks need to be accomplished in order to satisfactorily complete a given task. These subtasks include perception, understanding of the environment, learning strategies, knowledge representation, awareness of its own state, and manipulation of the environment. Reinforcement Learning (RL) [1] is a learning approach supported by behavioural psychology where an agent, e.g., a person or a robot, interacts with its environment trying to find an optimal policy to perform a particular task. In every time step, the agent performs an action reaching a new state and, sometimes, may obtain either a reward or a punishment. The agent tries to maximize the obtained reward by choosing the best action in a given state [2]. On the other hand, deep learning [3] is composed of many processing layers and has been successfully tested, among others, in image classification by representing different levels of abstraction [4]. Moreover, deep reinforcement learning [5] has combined the two aforementioned approaches to learning a motor policy mapping from a set of states to a set of actions. Deep reinforcement learning uses a neural network to learn the sum of direct rewards and expected future rewards for each action‐state either in discrete or continuous domains [6]. In this project, the student will work with the deep reinforcement learning approach with interactive feedback applied to a domestic robot scenario. In this context, it is expected to develop a simulated human‐robot scenario where the robot observes the environment states by using deep learning approaches and decide actions to perform by means of interactive reinforcement learning. 1. R. S. Sutton and A. G. Barto. Reinforcement Learning: An Introduction. Cambridge, MA, USA: Bradford Book, 1998. 2. Francisco Cruz, Sven Magg, Yukie Nagai, and Stefan Wermter. Improving interactive reinforcement learning: What makes a good teacher? Connection Science, In Press, 2018. 3. I. Goodfellow, Y. Bengio, and A. Courville. Deep learning. Cambridge: MIT press, 2015. 4. Y. LeCun, Y. Bengio, and G. Hinton. Deep learning. Nature, Vol. 521, Nr. 7553, pp. 436‐444, 2015. 5. V. Mnih, K. Kavukcuoglu, D. Silver, A. Graves, I. Antonoglou, D. Wierstra, and M. Riedmiller. Playing atari with deep reinforcement learning. arXiv preprint arXiv:1312.5602, 2013. 6. M. Kerzel, H. Beik‐Mohammadi, M. A. Zamani, S. Wermter. Accelerating Deep Continuous Reinforcement Learning through Task Simplification. In Proceedings of the International Joint Conference on Neural Networks (IJCNN), pp. 139‐144, 2018. Necessary Skills: Good programming skills. Deakin University CRICOS Provider Code: 00113B Page 31
School of Information Technology
2020 HONOURS PROJECTS
The kingdom of the 7‐segments: Drones as Moving Sculpture
Supervisors: Prof Peter Eklund
Email: peter.eklund@deakin.edu.au
Campus: Burwood
Start Date: January or July
“Franchise Freedom, a Flying Sculpture” (2018) at the Burning Man Festival ‐ an event held annually in the western United States at Black Rock City ‐ is a good
example of the use of 300 drones and lighting to synthesise a murmuration of starlings see https://youtu.be/ChnKSt4SbOI
This project proposes an a less ambitious number of drones but with a similar purpose, namely to use drones to entertain with movement and lighting.
Programable LED lights will be attached to quadcopter drones that will ray trace specific shapes in flight in an attempt to create a 3D lighting sculpture effect.
Success of the project will be in determining whether observers can identify and name the object being rendered by the drones. For instance, imagine 11 drones
tracing the letter ‘E’ below in the style of a 7‐seqgment display, where each ‘X’ identifies a drones position in the 2D positioning in the aspect of the reader, namely
one for each apex and 2 to fill the vertical line segments on the left, and horizontal line segment for each of the top, middle and bottom segments.
In 3D space, recognition of the letter ‘E’ might only result from a
murmuration behaviour in which the entire letter ‘E’ rotates slowly over
180 degrees w.r.t. the position of the observer so that the observer at any
position in the viewable space will see the emergence of the letter ‘E’ at
some point in its rotation.
Necessary Skills:
Excellent programming skills (Arduino/Python, Java/Android would be highly beneficial), software validation/test, experiment design, academic writing.
Deakin University CRICOS Provider Code: 00113B Page 32You can also read
