){kind=small}
Green Consensus Mechanisms for Energy-Efficient Decentralized Systems
Participants
Summary
Blockchain technology has revolutionized digital transactions and data security, but its high energy consumption remains a major concern. This research focuses on developing and analyzing green blockchain solutions, emphasizing energy-efficient consensus mechanisms to reduce the environmental impact of decentralized systems. Traditional methods, such as Proof of Work (PoW), require vast computational power, leading to excessive carbon footprints. This study explores alternative consensus protocols, including Proof of Stake (PoS), Delegated Proof of Stake (DPoS), Proof of Authority (PoA), and hybrid models, which significantly lower energy consumption while maintaining security and decentralization. Additionally, the project investigates the integration of renewable energy sources, carbon offset strategies, and hardware optimization to enhance blockchain sustainability. By addressing the balance between efficiency, security, and environmental responsibility, this research aims to pave the way for eco-friendly blockchain applications, benefiting industries such as finance, supply chain management, and smart cities.
Funding
){kind=logo}
Partners
){kind=logo}
){kind=icon}
)
)
){kind=logo}
)
Publications
[1] Alharbi, O., Shaikh, R. A., & Asif, R. (2025). Data-Aided Intrusion Detection Systems: Leveraging AI, Blockchain and Digital Twin Technology. In 2024 IEEE International Conference on Big Data (BigData) (pp. 8214-8215). (Proceedings - 2024 IEEE International Conference on Big Data, BigData 2024). The Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/BigData62323.2024.10825899
[2] Asif, R., Hassan, S. R., & Parr, G. (2023). Integrating a blockchain-based governance framework for responsible AI. Future Internet, 15(3), Article 97. https://doi.org/10.3390/fi15030097
Private 5G mobile network setup and encrypted data analysis
Participants
Summary
We build a 4G private network testbed, in which free5GC (developed by NCTU) acts as a core network, and the 4G cell station is provided by Gemtek Technology. Free5GC enables rapid and flexible employment of 5G private networks, supporting a variety of applications. Key Capabilities:
Support free 4G mobile network service. Every user in UEA registers a SIM card and then can get free network service within the range of core network, which means users do not need to pay text and call fees to operators.
Real-world Network traffic data gathering. With the test bed, we can get a volume of users’ traffic data from the real world, which can be fed into an AI/ML model for data analysis. For instance, the network operators can gather encrypted network traffic data of different applications on real testbed and train the model for application classification by fingerprint technology.
For security system. Current research already analyzes the DDoS attack on authentication protocol based on the free5GC testbed. Besides, other attacks like modifying and replying to core network traffic are also analyzed in free5GC.
Other experiments like network traffic offloading are supported on the testbed.
)
Funding
)
Publications
[1] Xiong, R., Tong, K.-L., Ren, Y., Ren, W., & Parr, G. (2023). Demo: From 5G to 6G: It is time to sniff the communications between a base station and core networks. In ACM MobiCom '23.
Optimised UAVs for Comprehensive Environmental Mapping
Participants
Summary
The rise of 6G and trends in Robotics for industry and agriculture creates demand for better understanding of the environments they inhabit, examples include drone delivery, crop growth mapping and disaster response. In order to better manage robotics systems, a more complete view of the environment is needed, including spatial, comma signal, atmosphere conditions and device localisation. This work seeks to optimise mapping of these measures with drone systems and deep learning, generating optimised paths and behaviours to make better use of limited resources while still building consistent representations of real world conditions.
PGR Poster (PDF, 1.13MB)
)
Funding
){kind=logo}
Measuring Employee Wellbeing Through Ubiquitous Devices
Participants
Summary
Employee wellbeing is an important issue for both employers and employees to consider, with the HSE estimating that 0.64 days per worker were lost to stress, anxiety, and depression in 2022/23. One particular issue in the workplace is burnout, which is characterised by long-term exhaustion and negative feelings towards work.
This project aims to create an innovative phone app to alert an employee that they may be approaching burnout. Users will be asked to type their feelings towards work, and the application will use the phone’s sensors, along with natural language processing (NLP) analysis on the text typed, to determine the employee’s risk of burnout. This will be a privacy-focused approach, empowering employees to make decisions to benefit their own wellbeing.
A Novel AI Temporal-Spatial Analysis Approach for GNSS Error Source Recognition
Participants
Summary
Global navigation satellite systems (GNSS) error source analysis is crucial for identifying factors that affect the accuracy of positioning, navigation, and timing services (PNT). Detecting and correcting these factors is essential for enhancing overall service accuracy. Traditional methods primarily focus on surface-level receiver output data, which may overlook underlying factors. Additionally, analyzing daily generated data is expensive and requires advanced proficiency. This research uses a novel temporal-spatial analysis approach to analyze GNSS error sources with artificial intelligence (AI) model support. We develop a noise segments dataset categorized into six types, with a particular focus on ionospheric disclosure, a deeper-level receiver data calculating PNT result. By applying clustering combined with a z-score normalization filter (ZFilter), we identify highly consistent noise segments in daily data, which aids in understanding potential causes. We then employ a multi-model deep learning approach to classify the noise segments, as opposed to relying on a single baseline model. Additionally, we experiment with semi-supervised learning through pseudo-labeling to improve classification performance. Our experiments show that our classifier achieves approximately 84% accuracy in identifying the noise segments.
)
Partners
)
Blockchain based Secure AI-Aided Intrusion Detection System
Participants
Summary
Cyber threats are becoming more advanced, making it difficult for traditional security systems to detect and prevent attacks effectively. This research focuses on developing a blockchain-based, AI-powered intrusion detection system to improve cybersecurity. Blockchain technology ensures that security data, such as attack logs, cannot be altered or tampered with, providing a transparent and trustworthy record of cyber threats. At the same time, artificial intelligence is used to detect unusual patterns in network traffic, helping to identify potential attacks in real-time. Smart contracts, which are self-executing programs on the blockchain, will automate security responses, reducing delays in addressing threats. The project will explore ways to make this system both secure and scalable while improving accuracy in detecting cyberattacks. By combining blockchain and AI, this research aims to create a more intelligent, decentralized, and resilient cybersecurity solution for protecting sensitive systems and critical infrastructure from evolving cyber threats.
)
Funding
){kind=logo}
Publications
[1] O. Alharbi, R. A. Shaikh and R. Asif, "Data-Aided Intrusion Detection Systems: Leveraging AI, Blockchain and Digital Twin Technology," 2024 IEEE International Conference on Big Data (BigData), Washington, DC, USA, 2024, pp. 8214-8215, doi: 10.1109/BigData62323.2024.1082589
To Bot or Not to Bot? Understanding and Analysing User Perceptions and Preferences for Gaining Trust in Mental Health Data Disclosures
Participants
Summary
This PhD research explores the potential of a bespoke chatbot, designed with human-like features and Human-Computer Interaction (HCI) heuristics, to facilitate personal data collections for initial mental health assessments. The findings from a dual study, evidence that a bespoke chatbot outperformed standard online forms, when collecting sensitive information from 253 participants, recruited via a local mental health charity and via social media.
Trust cultivated, when enhancing the chatbot's human-like features and HCI heuristics, improved the depth of sensitive personal data extractions, particularly for individuals aged 18-49 and non-binary gender identities. If utilised; these insights may facilitate a more efficient identification of appropriate support and treatment plans, such as referrals for individual or group therapy, Cognitive Behavioural Therapy (CBT), or leveraging existing online mental health applications like Wysa. Alternatively, it could enhance the initial assessment by providing a more holistic overview of the patient.
This approach could enable trained professionals to focus their efforts on delivering the necessary therapeutic support, rather than dividing their attention between administrative tasks and clinical responsibilities. The final stage of the research will focus on implementing the bespoke chatbot in a real-world mental health charity environment, to assess its practical applicability and effectiveness.
)
Funding
){kind=logo}
Partners
){kind=logo}
Publications
[1] Understanding User Preferences for Gaining Trust, When Utilising Conversational Agents for Mental Health Data Disclosures (2023) HCI International 2023 Posters, Communications in Computer and Information Science. 1833,167--174
[2] To Bot or Not to Bot?: Analysing Mental Health Data Disclosures (2024) International Conference on Human-Computer Interaction 14728, 97--115
Guess Who: Developing a Behavioural Fingerprint for Insider Threat Detection
Participants
Summary
When systems become compromised, data is leaked or networks are infiltrated our minds often wander to the dramatic; Hackers lurking in a dark basement, hoodies up, typing furiously. In reality, these attacks are completed by Insiders, where the average employee is driven to extremes by stressors in their lives, well-being, income or worse. This Project aims to develop a behavioural model to determine changes in user behaviour and in turn detect Insider Threat attacks at the source - the Insider themselves.
Organisations are no longer simply under threat from external hackers, the threat is now more tangible and real. The threat of an insider is constantly evolving as the way in which organisations are conducting business are too evolving, further contributing to the devastating impact such attacks have on wider organisational cyberspace.
This project looks to develop what would be needed to create a ‘Behavioural Fingerprint’, a method of benchmarking user behaviours, to determine and measure how users interact with devices and any changes that may occur if a user is compromised. Initial areas of interest include Keystroke Analysis, Mouse Dynamics and Linguistic Profiling.
)