Computer Science and Informatics (PhD)

Nowadays, the combination of big data and time series analysis is on the edge of world-leading research directions. The characteristic advantage is the wide and direct transferability of the methods and the algorithms that are developing from one scientific field to the other as well as to the industry. Applications in mechanical automotive, financial, public health, biological systems, among others, continuously improve the quality of our everyday life. The prediction of the behaviour or even of the sudden changes in these systems could introduce standards to risk assessment or to the over-specification. The purpose of this project is to develop generic new methods and tools in the detection and prediction of anomalies in time series based on novel concepts like correlation coefficient, natural time analysis, and machine learning with applications in a range of diverse systems.

The first priority here is the outcome of the applications to constitute independent studies. Finally, since, the impact of this project will be both direct and valuable will improve the engagement between academia and industry.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Stavros Christopoulos
• Dr Georgios Bargiannis

The project aims to leverage Artificial Intelligence (AI), specifically Machine Learning (ML) models, with the latest computer vision advancements for time-critical and responsive visual event management in a Smart House environment. It seeks to explore how ML models can be effectively deployed on edge devices, enhancing their capabilities and enabling real-time data processing for human-centric applications, for example, elderly care. The objectives of the project can be summarised as follows: 1. Develop vision models and co-inferencing mechanisms for real-time visual event interpretation. 2. Investigate computer vision applications in a Smart House, addressing challenges like space monitoring, human activity analysis, and emergency care detection. 3. Implement advanced AI/ML models, such as Transformers, for sequential IoT data processing. 4. Develop an evaluation matrix for real-world application factors. 5. Formalise the devised framework and models for social and commercial exploration. The project will start by creating dynamic visual analytical models and co-inferencing mechanisms. The project investigator will then enable model deployment within a Smart House Visual Event Management framework. Simultaneously, cutting-edge ML models will be explored to process visual and other IoT data in a Smart House. Semantic event information will be extracted and interpreted based on scene understanding. An evaluation matrix will be established for assessing real-world application factors. The framework will then be formalised for broader exploration by stakeholders and the public.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Zhijie Xu
• Professor Rakesh Mishra

The quest for sustainable urban development necessitates innovative approaches to assessing and enhancing energy performance at the neighbourhood level. This research proposes an AI-driven interactive visualization framework utilizing Network Graph Analysis (NGA) to compare energy performance across neighbourhoods. Traditional methods of energy performance analysis often lack the ability to dynamically visualize complex relationships and patterns among multiple entities. By integrating AI with NGA, this research aims to provide a sophisticated tool for visualizing and analyzing the interconnected energy performance metrics of residential buildings, facilitating more informed decision-making for policymakers, urban planners, and residents. This research seeks to bridge the gap between complex data analysis and practical application by providing a robust, AI-enhanced visualization tool that can drive substantial improvements in energy efficiency at the neighbourhood level.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Shamaila Iram
• Dr Minsi Chen

Industry 4.0 (I4.0) is a new paradigm of smart manufacturing that requires the intelligent integration of a wide range of systems and devices, including sensors, actuators, robots, and other manufacturing equipment in order to obtain greater efficiency, quality and productivity. However, designing Industry 4.0 systems that are interoperable, flexible, scalable, and capable of adapting to changing conditions is the key challenge in achieving smart manufacturing. Similarly, while digital technology is the backbone of I4.0 and underpins the U.K Governments Green Industrial Revolution, it is estimated the ICT sector’s carbon footprint accounts for 0.8-2.3 overall global emissions, and that by 2040 it is expected to account for 14% of the world’s carbon footprint at a time when emissions must shrink. Industry 4.0 refers to the current trend of automation and data exchange in manufacturing technologies, including developments in artificial intelligence, the Internet of Things (IoT), and cyber-physical systems. The goal of Industry 4.0 is to create "smart factories" that are more connected, efficient, and responsive to changing customer demands. Industry 4.0 technologies are expected to lead to significant improvements in productivity and competitiveness, as well as to the creation of new business models and industries. Robotics is an integral element of Industry 4.0, as they are used to automate a wide range of tasks including assembly, inspection, transportation, and packaging. In addition, they can be integrated into the production process through the use of sensors, artificial intelligence, and the Internet of Things (IoT), allowing them to communicate with other devices and systems and to make decisions based on real-time data. While robots play a valuable role in Industry 4.0, they are designed to augment and support human workers, which can pose safety risks to human workers if they are not properly designed, programmed, or maintained. Robots pose challenges to software designers who are required to take care of difficult architectural drivers such as acceptability, trust as well as recurrent software design issues such as interoperability, reusability, and customizability. Reference architectures provide a standard framework for designing, building, and maintaining a system as they specify a high-level, abstract view of the system, and define the key components, their relationships, and their interactions. The goal of a reference architecture is to provide a common language and a shared understanding of the system, which can help to reduce complexity, improve interoperability, and facilitate the development of new capabilities. While a limited number I4.0 reference architectures have emerged, i.e., Reference Architectural Model Industrie (RAMI), Robotics Reference Architecture (RRA), Robot Operating System (ROS) Reference Architecture, and the Robot Framework Reference Architecture their suitability in addressing interoperability, scalability, and sustainability remain an open research challenge. The overall aim of this project is to advance software architectural-level reasoning for pre-system understanding and post-system maintenance and evolution through the development of a Human-Robotic Reference Architecture for Sustainable Industry 4.0.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Minsi Chen

Computer Science at the University of Huddersfield are a Technical Associate Institute of the ATLAS project at CERN. ATLAS is the largest experiment at CERN and is most well-known for the role it played in the discovery of the Higgs Boson. As a part of the High Luminosity upgrade of the Large Hadron Collider (LHC) the ATLAS detector is also being upgraded which will require the processing of very high raw data rates (>5 TB/s) in order to identify events that could contain interesting physics that will then be stored for offline detailed analysis. We are offering two funded PhD studentships to work in this area. We envisage two main areas of activity with the detailed workplans being formulated based on the interests of the successful candidates. The two areas of activity are: 1) Development and characterisation of novel data processing pipeline approaches that can filter the very high data rates off the detector and implement algorithms that can detect events of interest. This project would suit individuals with an interest in high-speed communications systems; FPGA, GPU and other advanced processing solutions (e.g. MPSoC); and modelling and simulation of such systems. 2) Development and characterisation of algorithms to detect events of interest in the filtered data. Traditionally, techniques such as the Hough Transform have provided the basis for the detection, but there is significant interest now in the application of machine learning to this problem area. This project would suit individuals with an interest in pattern detection and machine learning. Both of the projects will involve close collaboration with, and visits to, our ATLAS partners located at prestigious, globally-distributed universities and research facilities, and there will be opportunities to spend time at CERN in Geneva.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Phil Lane
• Professor Richard Hill

The development of co-simulation procedures has led to the development of sophisticated numerical dynamic analysis tools. These are able to couple two different simulations or more, running alongside each other. Such methods allow for the study of more complex systems by coupling different sub-systems or coupling different phenomena in the same system. The aim of this work involves the study and investigation of co-simulation methodologies and its application in numerical dynamic analysis tools. Different approaches are to be implemented and tested under a series or different case scenarios and benchmarks. The final objective of this work includes the development and implementation of a new co-simulation framework on a state-of-the-art Pantograph-catenary dynamic analysis tool. This is able to handle the numerical analyses of pantograph-catenary interaction, where the pantograph is modelled as a multibody system in interaction with finite element OLE model.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Pedro Cabaco antunes
• Professor Joao Pombo

The aim of the project is to invetigate innovative edge software architectures that enhance the integration of AIoT applications in smart environments, for example, smart buildings and smart houses. The objectives of the project can be summarised as follows: 1. Develop an integrative edge-sensor architecture that is scalable, energy-efficient, secure, and adaptable to diverse environments. 2. Create an interpretative meta schema and edge operating system to facilitate efficient resource management in the devised architecture. 3. Implement dynamic model partitioning and co-inferencing mechanisms for executing machine learning models on heterogeneous processors. 4. Test and validate the developed software architecture and process mechanisms through real-world case studies in a smart house research facility. The project will begin by exploring existing challenges in integrating smart sensors and edge servers, leading to the development of the proposed software architecture. The research will involve designing the meta schema for resource management and developing the edge operating system for efficient control and monitoring. The project will also focus on partitioning machine learning models for co-inference on heterogeneous processors, utilising dynamic graph partitioning techniques. The developed framework and mechanisms will be rigorously tested and validated through real-world case studies conducted in a smart house research facility, ensuring their practicality and effectiveness.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Zhijie Xu
• Professor Rakesh Mishra

Pantograph-OLE Interaction plays a fundamental role in the traction of electric railway vehicles. The sliding contact between the overhead line and the pantograph contact strips must be as smooth as possible and uninterrupted. The study of this interaction under aerodynamic loads has become a key factor on the development of new overhead lines and current collection systems. The employment of numerical dynamic analyses tools to study pantograph-OLE interaction is now being accepted by the industry, as these types of software are becoming more reliable and accurate. Though, added model complexity is sought after in order that more complex problems can be analysed. One of these aspects is the inclusion of aerodynamic effects in these types of numerical studies, which today have an impact in the design of new pantographs and overhead line systems. This work aims to study the aerodynamic effects on the pantograph and the overhead line, and the development of a modelling methodology to include them in pantograph-OLE interaction numerical analyses. These methodologies are to be incorporated in a state-of-the-art dynamic analysis tool already developed, so its capabilities are augmented.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Pedro Cabaco antunes
• Professor Joao Pombo

This study will build on existing research mapping twenty-first-century skills to degree provision as the basis of a microcredentialing standard for UK higher education. This work considers a broader application of a microcredentialing standard across formal education.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Rupert Ward
• Professor Richard Hill

Sustainable Software Engineering has been defined as "the ability of software to be developed and used in a way that minimizes its negative impacts on the environment, economy, and society, while ensuring its long-term viability and quality". This project aims to analyse the impact of Aspect Oriented Programming (AOP), specifically the development of reusable aspects designed for efficiency and resilience, on Software Sustainability. This will involve the use of software metrics (including novel metrics developed for this purpose) to measure the impact of AOP on both the software development process and upon the sustainability of the end-product software system itself.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Gary Allen
• Dr Simon Parkinson

Internet of things (IoT) that integrate a variety of devices into networks to provide advanced and intelligent services. By 2020, 50 billions of IoT heterogeneous devices are connected to the internet. These connected IoT devices form an intelligent system of systems that transfer the data without human-to-computer or human-to-human interaction.

While enjoying the convenience and efficiency that IoT brings to us, threats to the IoT devices and applications are on the rise; however patterns within recorded data can be analysed to help predict threats. There are different types of attacks and threats that may diffuse the IoT architecture, such as spoofing, DDoS attacks, unauthorized access, man-in-the-middle attacks and ad hoc networks. One feasible countermeasure against targeted attacks is to apply Intrusion Detection System (IDS) on the IoT network to detect and report intrusion, policy violations, and unauthorized use. Hence a an efficient approach for detecting and predicting IoT attacks is needed.

This PhD project will develop an innovative IDS model for the application and distribution of deep Learning algorithms within IoT networks. There will be a need to develop proof-of-concept simulations for the purposes of benchmarking and evaluation. Applicants should be comfortable with programming in a high level programming or simulation language (C, Python, R, etc.). A background in mathematics/statistics will be useful, as will experience of planning and workflow scheduling systems.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Richard Hill
• Dr Hussain Al-aqrabi
• Dr Muhammad zahid Hussain

Argument mining is a research area within natural language processing. The aim of argument mining is the automatic identification and extraction of argumentative structure from real world textual resources such as legal documents, product reviews, online debates or newspaper articles. This project will involve a detailed research study of the structure of natural language arguments with the aim of devising new and effective computational argument mining techniques. The successful candidate will be expected to focus on an application domain and extend existing natural language or machine learning techniques applied to the argument mining domain.

Essential attributes for the candidate: • Experience of fundamental Natural Language Processing (NLP) or Machine Learning (ML) techniques. • Competent in applying NLP toolkits, such as NLTK or Spacy, or ML toolkits such as Scikit-Learn. • Knowledge of argumentation theory and defeasible reasoning would be advantageous. • Good written and communication skills. • Strong motivation, with evidence of independent research skills relevant to the project.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Quratul-ain Mahesar
• Dr Isa Inuwa-dutse

The realisation of UK’s target of “Net zero energy buildings by 2050” could only be possible by prioritising overall quality of living in the buildings and not just by enhancing energy efficiency. Other factors, such as thermal and visual comfort, fabric efficiency, level of insulation, building orientation, and ventilation are also important. Furthermore, exploring this information together with other attributes such as environmental as well as demographic and behavioural information could play vital role to provide more personalised, adaptive and context aware solutions to enhance overall performance of the buildings. Currently, the focus is greatly shifting towards analysing IoT (Internet-of-Things) sensors data along with the contextual information. The proposed research work aims at analysing IoT data such as temperature, CO2 emission, humidity, electricity and gas usage together with contextual data such as environmental data, building information, and demographic information. The relevant attributes will then be selected to develop the most suitable prediction model to predict the energy performance of the building with respect to contextual information. The outputs of this research work are two fold; 1) to calculate the current energy performance as well as carbon footprints of the building with respect to the contextual information 2) predicting future energy performance and carbon footprints at various level of timescale and with respect to different contextual information. It is expected to develop a platform or toolkit to visualise the outputs of analysis and prediction modelling. This project will provide proof of concept prototype that can be used for benchmarking and evaluating the outputs.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Shamaila Iram
• Professor Richard Hill

The aim of this project is to use natural language processing and information retrieval techniques to extract relevant information from medical notes. This will entail developing a thorough understanding of existing methods and tools, testing them on concrete collections of medical notes (e.g. in mental health) and developing novel methods improving on the state of the art. In some cases, medical texts will be semi-structured, thus making analysis easier, but in other cases text will be free, which poses the biggest challenge. Analysis work may have to be carried out in collaboration with medical experts who will assess the validity and usefulness of the extracted knowledge.

The successful candidate will have thorough computer science education, and will have some specialised knowledge in artificial intelligence, natural language processing or information retrieval.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Tianhua Chen
• Professor Grigoris Antoniou

The Industrial Internet of Things (IIoT) is a challenging technology for businesses. The ability to secure and trust IIoT devices is of paramount importance, and can often be a vulnerability to expose valuable Intellectual Property (IP). There is a need to develop new mechanisms of trust that also provide immutable provenance records. This PhD project will identify security risks in IIoT networks and focus on developing new, smarter, and more agile approaches to ensuing trust in IIoT networks. There will be a need to develop proof-of-concept simulations for the purposes of benchmarking and evaluation, as well as a hardware-in-the-loop demonstrator. Applicants should be comfortable with programming in a high level programming or simulation language (C, Python, R, etc.). A background in mathematics/statistics and network security will be useful, as will experience of interfacing embedded systems and sensors.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Richard Hill
• Dr Hussain Al-aqrabi
• Dr Muhammad zahid Hussain

Requirements engineering (RE) is the process of defining, documenting and maintaining requirements in the design process and is considered the key success factor in software systems projects. While requirements elicitation is a relatively mature area of RE, an inability to frame problems and explicitly relate them to approaches is recognised as one of the most glaring deficiencies of practice and theory (Jackson, 1995). Despite this recognition, only a small number of studies have attempted to understand the factors that influence requirements practitioners in their selection of appropriate approaches(s). Given the plethora of available approaches, understanding the factors that influence the selection of requirements engineering (RE) approaches during a software or systems development project remains an open challenge. We use the term ``approach'' to mean any technique, tool or method that may be used during RE endeavours. The primary aim of this research is to organise the RE approaches into an ontological framework, based on a series of collaborative studies, that gathers the collective experience and knowledge of RE practitioners and researchers in the selection of complementary RE approaches.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Gary Allen

In recent years, the integration of artificial intelligence (AI) into decision-making processes across various domains has raised concerns about fairness and bias. While AI systems offer unprecedented efficiency and accuracy, they can inadvertently perpetuate or amplify biases present in the data used for training. Addressing fairness and bias in AI decision-making is crucial to ensure equitable outcomes for all individuals affected by these systems. One promising approach to enhancing fairness and mitigating bias is using counterfactual explanations. Counterfactual explanations provide insights into how decisions would have been different under alternative scenarios, offering transparency and accountability in AI systems.

The primary objective of this research is to investigate the effectiveness of counterfactual explanations in promoting fairness and mitigating bias in decision-making with AI systems. Specifically, the research aims to:

· Evaluate existing methods of generating counterfactual explanations in the context of fairness and bias.

· Develop novel techniques for generating counterfactual explanations tailored to address fairness concerns in AI decision-making.

· Assess the impact of counterfactual explanations on user understanding, trust, and acceptance of AI-driven decisions.

· Investigate the trade-offs between fairness, accuracy, and interpretability in AI systems equipped with counterfactual explanations.

Entry requirements: A Bachelor’s and/or master's degree (or equivalent) in Computer Science, Data Science/Analytics, Information Technology, or a related field. Proficiency in a programming language commonly used in AI/ML, such as Python, R, or Java is preferable.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Abirami Gunasekaran
• Dr Pritesh Mistry
• Dr Quratul-ain Mahesar

Data analytics and the use of Artificial Intelligence techniques to forecast outcomes is an area of considerable interest to the Business Intelligence community.

As new business models and technologies such as cloud computing mature, the emergent use of connected devices such as smart phones and Internet of Things (IoT) appliances is creating a demand for resilient Business Intelligence systems that can provide robust data analytics services across different security realms, often in multi-cloud settings.

This PhD project will develop a cross-realm BI model for distributed data analytics services. There will be a need to develop proof-of-concept simulations for the purposes of benchmarking and evaluation, as well as a hardware-in-the-loop demonstrator. Applicants should be comfortable with programming in a high level programming or simulation language (C, Python, R, etc.). A background in mathematics/statistics and network security will be useful, as will experience of signal processing and/or time series data analysis.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Muhammad zahid Hussain
• Professor Richard Hill
• Dr Hussain Al-aqrabi

The Internet of Things (IoT) is aimed at connecting billions of things and enable communication and information exchanging among things, where many different types of large directed networks will arise. In large directed networks, there may exist dead-zones, where only incoming edges are available but without outgoing edges from the zone. There may also exist nearly-dead-zones, where the number of outgoing edges is significantly smaller than the number of incoming edges for the whole zone. Finding all such dead-zones and/or nearly-dead-zones in large directed networks is important to analyse and maintain the networks. Large (nearly-)dead-zones may also contain small (nearly-)dead-zones. For example, a given standalone directed network as a whole, is a dead-zone. This means that we can build up a hierarchical structure of (nearly-)dead-zones for any given large directed networks. Your tasks in this project mainly include: (1) Identify research gaps between existing approaches for finding dead-zones and our target - dead-zone finder in large directed networks; (2) Design efficient algorithms to find (nearly-)dead-zones in large directed networks; and (3) Evaluate your proposed algorithms against existing algorithms. The application of the research results from this project will help us to understand the community structures and communication patterns across large directed networks in IoT.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Simon Parkinson

Technologically-induced communities enabled by social network platforms such as Twitter and Facebook support a myriad of diverse users to remain connected, leading to a highly connected and dynamic social media ecosystem. The openness and lack of effective filtering mechanisms result in the proliferation of unwanted and fake information. Since the advent of modern social network which promotes viral propagation of all sort of information. Misleading information leads to catastrophic outcome and hampers the fight about the containment measures. A case in point is the COVID-19, whose devastating impact could be mitigated using social sensors to help in combating the outbreak. As the virus propagate, so is fake news/misinformation about it. Thus, it is equally important to combat it simultaneously as we confront the virus spread. In the context of Twitter, while many detection methods have been proposed in the past, identifying genuine content is still a challenge. The project is aimed at a comprehensive analysis using social media to identify the proliferation of spurious content and how to prevent it. Success of the work will help with real-time monitoring of health-related issues from online news streams.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Sangeet Saha
• Dr Isa Inuwa-dutse

The ways in which we interact with computer systems is changing, with a vast array of alternative input methods now readily available to the 21st Century developer. Alongside these developments there is significant research into hand gesture recognition systems that provide the user with an efficient and effortless method of human-computer interaction. Typically, hand gestures are elicited from target users in Gesture Elicitation Studies (GES). This research project seeks to address two critical issues in traditional GES. Firstly, there is the potential for biases to be introduced by priming the participants before gestures are elicited. Secondly, the analysis of the gestures produced by the participants for higher level interactions is complicated. It is anticipated that this research will lead to a more rigorous framework for gesture analysis.

Key Objectives:

1) Biases Introduced by Priming: Investigate the potential biases introduced by traditional priming methods in GES. Develop strategies to minimise or account for any potential biases, to ensure more accurate and unbiased participant responses.

2) Gesture Analysis Challenges: Scrutinise the intricacies of analysing gestures produced in GES for immersive systems by considering factors such as speed, type, and scale. Develop a framework for the analysis of gestures to support the researcher addressing the challenge of distinguishing between different gestures.

3) Integration and Comparative Analysis: Integrate findings from the examination of biases and gesture analysis to generate a methodological framework suitable for immersive system GES. Conduct a comparative analysis between the traditional and refined methodologies, highlighting the advantages and limitations of each approach.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Mr Christopher Dewey
• Dr Duke Gledhill

This research aims to revolutionize Structural Health Monitoring (SHM) by developing novel attention mechanisms to enhance the accuracy, efficiency, and interpretability of SHM systems. By leveraging advanced machine learning techniques, such as self-attention and multi-head attention, the study seeks to improve the detection and characterization of structural anomalies. The integration of these attention mechanisms with existing SHM sensor networks will facilitate real-time data acquisition and analysis, enabling more reliable and actionable insights for maintaining the safety and integrity of critical infrastructure. Through extensive laboratory experiments, field deployments, and comparative analyses with traditional SHM methods, this research will demonstrate the potential of attention mechanisms to transform SHM practices and contribute significantly to the fields of civil engineering and machine learning.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Muhammad zahid Hussain
• Dr Tariq Alsboui
• Professor Richard Hill

It is increasingly important that the skills gained within learning are clearly mapped to job roles and that they are communicated clearly to employers, governments and educational providers. Using a 21st Century Skills framework this research will explore a range of courses and job roles to categorise learning and to therefore support learning analytics approaches based on digital skills development.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Rupert Ward
• Professor Richard Hill

Autism Spectrum Disorder (ASD) is a complex neurodevelopmental condition with a wide range of symptoms and potential interventions. Information on ASD is scattered across diverse sources, including research literature, clinical guidelines, and therapeutic practices. Clinicians and researchers face the challenge of retrieving ASD-related information by manually analysing and combining this voluminous and heterogenous knowledge. This task can be time-consuming (in some cases even infeasible) and hinders decision making or the development of effective support strategies. This project aims to develop a comprehensive Autism Spectrum Disorder Knowledge Graph (ASD-KG) to streamline knowledge access and integration, empowering ASD research and clinical decision-making.

The development of a knowledge graph about autism spectrum disorder has the potential to significantly advance autism research and clinical practice. The proposed resource will aid in organising and interlinking medical knowledge about ASD, addressing the challenge of linking synonymous medical concepts from different sources. Utilising semantic search over ASD-KG it will promote efficient question answering, drastically reducing the time clinicians and researchers spend on manually analysing medical data sources. Furthermore, the ASD-KG will support complex queries that draw insights from multiple data sources, enhancing the clinical understanding of ASD.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Emmanouil Papadakis
• Dr Ilias Tachmazidis
• Dr Georgios Bargiannis

Industry 4.0 (I4.0) is a new paradigm of smart manufacturing that requires the intelligent integration of a wide range of systems and devices, including sensors, actuators, robots, and other manufacturing equipment in order to obtain greater efficiency, quality and productivity. However, designing Industry 4.0 systems that are interoperable, flexible, scalable, and capable of adapting to changing conditions is the key challenge in achieving smart manufacturing. Similarly, while digital technology is the backbone of I4.0 and underpins the U.K Governments Green Industrial Revolution, it is estimated the ICT sector’s carbon footprint accounts for 0.8-2.3 overall global emissions, and that by 2040 it is expected to account for 14% of the world’s carbon footprint at a time when emissions must shrink. The future of I4.0 is highly dependent on a resilient and sustainable eco-system of software systems. In this new environment, standardization, interoperability, scalability, security, and sustainability are critical factors in achieving the I4.0 vision, which paves the way for leveraging the Industrial Internet of Things, Big Data analysis, simulation, Cloud Computing, and augmented reality. The first step towards achieving system integration is to reason about the elements, their properties and relationships that make up the eco-system. Software system design is a key component, which starts with software architecture as it lays the foundation for the successful implementation, maintenance and evolution in a continually changing execution environment by providing a mechanism for reasoning about core system quality requirements including interoperability scalability, and sustainability. Reference Architectures have been used for the aggregation of knowledge in a range of specific domains, promoting the reuse of design expertise and facilitating the development, standardization, and evolution of software systems. Examples include AUTOSAR for the automotive sector (AUTOSAR, 2021), ARC-IT for transportation systems (U.S. Department of Transportation, 2019), EIRA for interoperable e-Government systems (Joinup, 2021), and SOA RA for service-oriented systems (Open Group, 2021). The main benefits of these architectures include increased interoperability among systems and subsystems, reduction of development costs and time by enabling reuse, reduction of risks in software projects, improvement in communication, and adoption of best practices. While a limited number I4.0 reference architectures have emerged, i.e., Reference Architectural Model Industrie (RAMI) and Industrial Internet Reference Architecture (IIRA), their suitability in addressing interoperability, scalability, and sustainability remain an open research challenge. The overall aim of this project is to advance software architectural-level reasoning for pre-system understanding and post-system maintenance and evolution through the development of a Reference Architecture for Sustainable Industry 4.0

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Gary Allen
• Dr Colin Venters

Fog computing encourages a shift of handling everything in the core of cloud computing to handling at the edges of a network. To be specific, instead of sending all the data collected to the cloud, fog computing suggests to process the data at leaf nodes or at the edges. This idea is also called ‘edge analytics’. In smart healthcare at home, sensing data about inhabitants or patients should be analysed and processed in a near-real-time manner. To enable this, fog computing technologies can be used to distribute the healthcare data analytics onto fog computing nodes and communication with remote cloud servers only takes place when necessary. In such scenarios, it will be critical to process smart healthcare sensing data at home efficiently and effectively on fog nodes. To this end, this project aims to develop novel home healthcare and remote monitoring solutions. Your main tasks in this project include: (1) Study some smart home simulation tools, such as OpenSHS; (2) Incorporate smart healthcare into the simulation tools; (3) Develop novel approaches based on fog computing paradigms to achieve near-real-time processing of healthcare sensing data.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Muhammad zahid Hussain
• Professor Richard Hill

Dementia significantly impairs an individual's ability to perform daily tasks independently, posing challenges for both patients and caregivers. This research aims to develop a computer vision-based assistive technology system designed to support independent living for dementia patients. By leveraging advancements in computer vision and machine learning, the proposed system will monitor and assist patients in real-time, ensuring their safety and improving their quality of life while reducing the burden on caregivers.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Shamaila Iram
• Dr Muhammad zahid Hussain

A major challenge of additive manufacturing (AM) technology is that AM processes are not robust enough and AM production machines lack sufficient process control, which consequently bring various shortcomings that are commonly seen in AM products, such as poor as-built surface finish, porosity, and mechanical properties not equivalent to those of bulk materials. X-ray Computed Tomography (XCT) is often employed to inspect the porosity of AM parts, providing more information (e.g. size, location, morphology) in comparison to the traditional Archimedes method. However, the accuracy and performance of XCT on AM porosity inspection need to be verified. This research project will conduct an experimental study to target the optimum scanning configuration for AM porosity measurement and compare with other porosity inspection methods, e.g. Archimedes, scanning electron microscope and ultrasonic. XCT simulation will complement to experimental work, which allows investigate the impact of major scan parameters on XCT porosity measurement. Physical and virtual AM artefacts will be developed. The project will be based on the Future Metrology Hub, Centre for Precision Technologies. The selected student will be provided with the training of XCT and simulation software.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Gary Allen

Artificial Intelligence (AI) systems have become increasingly pervasive across various domains, ranging from healthcare to finance, offering remarkable predictive capabilities. However, as AI/ML models deliver accurate predictions, their complexity grows, making it challenging to understand their decision-making processes. The complex nature of these models often hinders their ability to provide comprehensive and understandable insights, limiting their adoption in real-world applications. Consequently, there is a growing demand for innovative approaches that leverage interactive visualisations to enhance the interpretability and transparency of AI systems. Explainable AI (XAI) has emerged as a crucial field aimed at making AI systems transparent and interpretable. Visual analytics (VA) inherently provides a means to represent data or models in a comprehensible manner, particularly for individuals with limited experience in machine learning (ML).

Objectives: • To investigate the state-of-the-art techniques in interactive visual explanations for XAI. • To develop novel methodologies for generating interactive visual explanations that enhance user understanding of AI/ML models. • To evaluate the effectiveness and usability of interactive visual explanations across various domains and user groups. • To contribute to the advancement of XAI through the integration of interactive visualisation techniques.

Entry requirements: A Bachelor’s and/or master's degree (or equivalent) in Computer Science, Data Science/Analytics, Information Visualisation, or a related field, with a focus on visual computing. Proficiency in a programming language is essential (e.g. Python, R, Matlab, Java, C).

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Abirami Gunasekaran
• Dr Pritesh Mistry
• Dr Shamaila Iram

Transitioning to net zero is one of the greatest challenges facing global supply chains and requires radical rethinking of decision making processes across the supply chain. Artificial Intelligence is a significant enabler in this context and research in AI solutions for supply chains has been rapidly expanding. While standard black box machine learning solutions have been considered, the problem of making decisions to meet net zero targets requires deeper understanding of causes of emissions throughout the supply chain to be able to provide justifiable recommendations with clear links from actions to outcomes. This is more likely to be achieved through AI solutions that prioritise explainability and trustworthiness. The purpose of this PhD research is three-fold: investigate the current landscape of intelligent net zero decision making and support for supply chains; propose novel intelligent solutions for net zero supply chain decision support that prioritise explainability and trustworthiness, including, but not limited to, neurosymbolic methods; validate the applicability and likelihood of adoption of these solutions through real-world case studies of UK and global supply chains.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Georgios Bargiannis
• Dr Emmanouil Papadakis

The emergence of big data has stimulated interests in domain specific knowledge extraction. In turn, this can motivate novel approaches to the formulation of strategic decisions and planning. With an ever-increasing decentralised approach to data acquisition, one key challenges lies in the modelling of data that meets the domain requirements of a specific application. For a data analytics pipeline to be domain-specific and explainable, a data model needs to contain semantic descriptors and logical structures that are intrinsic to the domain of interest. To achieve this, the feature space underlying a data model should be tightly coupled with the ontology of a specific knowledge domain, e.g. education, physics and chemistry.

This proposed PhD will focus on developing the theoretical design and implementation of automatically deriving ontologically based features spaces from open data sets. The successful applicant will need to have experience of applied statistics, forecasting and model generation, along with some familiarity of high-level programming languages (such as C, C#, and Python) and software tools (R, Tableau, KNIME).

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Richard Hill
• Dr Minsi Chen
• Dr Pritesh Mistry

The employment of finite element methods in engineering plays a large role in the analysis of structures. With the advancements in computer resources, dynamic analysis applications based on this method are able to analyse large and complex systems. This work aims on the development and implementation of novel finite element modelling methodologies, able to handle the dynamic interaction between the overhead line and the pantograph in railway systems. Focusing on the construction of finite element models of the overhead systems and its dynamic analysis. The newly developed modelling methods are to be incorporated in a state-of-the-art dynamic analysis tool already developed, so its capabilities are augmented. The new methodologies are to be validated using experimental line tests in collaboration with industrial partners of the railway sector.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Pedro Cabaco antunes
• Professor Joao Pombo

Knowledge about the impact of the COVID-19 is evolving and its long-term effect is yet to be fully established. While the efforts to flatten the infection curve is yielding positive results, it is crucial to understand the aftermath, notably from the perspective of recovered people. COVID-19, akin to illnesses such as diabetes, seems to require precautionary measures to observe in managing it. Thus, incorporating diverse information sources from social media will help in offering a useful prevention pathway that is better equipped to tackle resurgences and other eventualities. The main goal of the proposed study is to harness information from online social media to address COVID-19-related challenges along the following dimensions: (1) Systemic Analysis of COVID-19 Journey: While adhering to ethical and GDPR requirements, the goal is to curate relevant datasets that will inform quality standards for research and make available the dataset for future research. This will enrich existing trusted public health information sources. The data will make it possible to extract information on aspects such as lockdown, contact tracing and the public reactions/perceptions. Self-reporting posts will be crucial in offering relevant signals about the users’ pathways to recovery and associated consequences. The approach will rely on time series analysis of related datasets at different stages/periods – during the pandemic (pre-vaccine and post-vaccine) and after the pandemic (recovered people during the pandemic and post-vaccine recovery). Through the analysis of the data, the ‘scarring’ effects of the pandemic and possible societal damages can be mitigated. For instance, during the onset of the pandemic pockets of vandalisation have been reported, which have been triggered by baseless claims, e.g. that 5G causes COVID-19. (2) Behavioural Cues and Predicting Actionable Areas: Triggers for viral infection and transmission within the society can be studied to examine which of the imposed measures are more effective. Insights into these aspects will inform which best strategy to use in communicating public health messages for a wider positive impact on public behaviour. This can be achieved by leveraging advances in A to determine potent actionable areas for the maximum benefit towards combating the virus. (3) Mental Well-being and Community Effort: The infamous pandemic-induced lockdown has its many tolls across various sectors. Crucial to understanding the impact of the lockdown on mental health since people have been locked indoors usually without jobs and momentous apprehension about eventualities. Understanding useful behavioural cues from self-reporting users (primary source or affected users) and secondary sources/third party i.e. from users with first-hand knowledge about such behaviour to study physical and mental health effects such as self-harm, and depression. Also, to study how various actors/stakeholders/organisations/NGOs, civil society, faith groups, influential users play their roles in aiding or abating the fight against the pandemic. To analyse how each stakeholder engages with online followers to help in identifying a set of resourceful organisations supporting in curtailing the negative impact of the pandemic.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Isa Inuwa-dutse
• Dr Sangeet Saha

Mental disorders and diseases affect a large proportion of the whole world population, including children, adolescents and elderly people, causing disabling and life-threatening conditions. They are cause for inequity, social stigma and discrimination, and human rights violations. Care services are limited, costly and often insufficient. In this context non-pharmacological interventions are increasingly important for health and social care. This project offers to students the opportunity to investigate uses of technology to enhance the design, implementation and evaluation of non-pharmacological interventions through addressing one of the following research lines: a) delaying onset and progression of disorders and diseases (e.g. through personalized and adaptive training of key psycho-physical capabilities); b) preserving general function and autonomy (e.g. through augmentation of cognitive, affective and behavioural capabilities, and technology-mediated, personalized activity monitoring and support); c) promoting engagement in daily life activities (e.g. through technology-mediated, personalised activity monitoring, adaptive support and stimulation of activity); d) enhancing rehabilitation (e.g. through personalized and adaptive monitoring and support of rehabilitation activities); e) enhancing social integration (e.g. through social networking technologies to facilitate meaningful social interactions remotely); f) enhancing the personalization, quality and efficiency of care provided by formal and informal caregivers (e.g. through facilitating the monitoring of sufferers’ conditions and activity, the provision of adaptive and personalized activity support, and remote interactions between sufferers and caregivers). To address these lines, students will investigate innovative approaches integrating complexity science, systemic design, human factors and ergonomics, and exploring novel uses of cutting-edge technologies including: assistive technologies; Internet of Things; autonomous agents; affective computing; deep learning; context-aware computing; biometric monitoring; activity tracking; augmented and virtual reality; digital games; social media. Projects will focus on high social impact mental disorders and diseases such as dementia, autism, attention deficit hyperactivity and brain injury.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Grigoris Antoniou
• Dr Georgios Bargiannis

The objective of this research is to leverage generative AI to enhance the assessment and prediction of Energy Performance Certificates (EPC) for residential buildings. Traditional methods of evaluating EPC ratings often rely on static data inputs and manual inspections, which can be time-consuming and prone to inaccuracies. By utilizing generative AI models, such as Deep Learning and Visual Analytics (Interactive Dashboards), this research aims to dynamically contextualize and interpret a wide range of data inputs, including historical energy consumption, architectural features, and real-time environmental conditions. This approach promises to provide more accurate and detailed assessments of a building's energy performance, offering actionable insights that can be tailored to specific contexts and user needs. The proposed study will develop and validate an AI-driven framework capable of generating comprehensive energy performance reports. These reports will not only predict EPC ratings with high precision but also offer contextual recommendations for energy efficiency improvements. The research will involve collecting extensive datasets from various residential buildings, training the AI models to understand the nuanced interactions between different factors affecting energy performance, and testing the system in real-world scenarios to ensure its reliability and practical applicability. By integrating generative AI for contextualization and prediction, this project aims to significantly improve the accuracy, efficiency, and usability of energy performance assessments, ultimately contributing to more sustainable housing practices and informed policy-making.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Shamaila Iram
• Professor Rakesh Mishra

In terms of scale the Boston Consulting Group states that “IoT sensors and devices are expected to exceed mobile devices as the largest category of connect devices in 2018, growing at 23%”. Key markets that will benefit from this technology push are advanced robotics (digital manufacturing), connected and autonomous vehicles, and the emerging health and wellness monitoring market.

These large-scale networks present opportunities and challenges to learn from massive, distributed datasets. Algorithms to perform supervised and unsupervised learning can enable insight to be derived from data. Such data is collected and often analysed locally by individual IoT devices, presenting difficulties when governing the distribution of learning algorithms to arrive at robust conclusions for queries.

This PhD project will develop a governance model for the application and distribution of Machine Learning algorithms within IoT networks. There will be a need to develop proof-of-concept simulations for the purposes of benchmarking and evaluation. Applicants should be comfortable with programming in a high level programming or simulation language (C, Python, R, etc.). A background in mathematics/statistics will be useful, as will experience of planning and workflow scheduling systems.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Muhammad zahid Hussain
• Professor Richard Hill

Hardware-In-the-Loop (HiL) is a novel simulation technique where a physical system interacts within a simulation in realtime. This technique is employed in the development and testing of complex systems. It allows mechanical systems to be tested, avoiding real tests which would otherwise be costly or unfeasible. There are challenges in setting up these types of simulation frameworks. The simulation program is required to be efficient and able to be evaluated in real-time. A robust control system is also necessary to acquire sensor data and control the response of all actuators accordingly. The development of this work is set on the development of a HiL framework for pantograph testing, in interaction with a numerical model of the overhead line. A fully equipped, world class, £ 3.5M pantograph test bench is available to procced with this works. Industrial partners in the transport and railway sector are to be involved in this work.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Pedro Cabaco antunes
• Professor Joao Pombo

Large-scale HPC cluster systems are finding increasing deployment in academic, research, and commercial settings, hence securing the HPC infrastructure is an important task. There are many security threats coming from both the Internet and internal networks. It is crucial to introduce an adequate level of security to the infrastructure, to prevent an unauthorised access to the HPC resources and avoid loss of valuable data. The aim of this project is to investigate the security issues in HPC systems, and devise a framework suitable for securing HPC cluster systems in Higher Education and research institutions. The challenge is to secure internal distributed resources against unauthorised access while permitting easy access by legitimate users, to coordinate security across different node platforms and different specialised function nodes (there is a separation of nodes into 'head nodes', 'compute nodes', 'storage nodes', and 'management nodes'), and to maintain the integrity of all nodes since many nodes share identical configurations. This framework should provide coordination between security domains on campus and between institutions when resources are shared across multiple organisations. It should address the security issues posed by the high-bandwidth connections, extensive computational power, massive storage capacity, and should enable process monitoring, network port scanning and traffic analysis.

Please see our Scholarships page to find out about funding or studentship options available.

• Dr Colin Venters
• Professor Rupert Ward

Research on recognition of Activities of Daily Living (ADL) in a smart home environment has generally followed two distinct directions. On the one hand, knowledge-rich rule-based systems have been developed that rely on rules of the form IF one or more sensor-based events are detected THEN a particular activity or activities are recognised. These systems are easily explainable but require expertise in encoding the rules. On the other hand, the proliferation of machine learning techniques has led to research that relies on learning activity patterns from large sets of sensor data. These approaches can handle uncertain or faulty sensor data but rely mostly on black-box techniques which are not interpretable. The increased demands of activity recognition in real-world case studies require hybrid approaches that combine the strengths and mitigate the weaknesses of rule-based and data-driven approaches. Hybridisation can follow one of these directions: - Handcrafted rules can be used as features in machine learning algorithms. - Results of rule-based reasoning can be used to narrow down machine learning search space. - Rules can be learned from data instead of being hand crafted by experts. - Two-level recognition: machine learning performed directly on raw sensor data, followed by higher-level rule-based reasoning.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Ilias Tachmazidis
• Dr Georgios Bargiannis

The audio mixer in the core technology used to mix popular music. Over the last 50 years the audio mixer has evolved from physical, mainly hardware-based systems to virtual, mainly software-based systems that include greater visual feedback, precision, and processing capabilities. Despite this evolution the user interface of modern audio mixers still harks back to the original channel strip based analogue interfaces that emerged in the 1970s. Whilst this historical metaphorical reference arguably supports experienced users, it lacks relevance to newer/younger users. This presents a barrier to learnability. This research project seeks to bridge the historical design gap in audio mixing interfaces, making them more accessible and intuitive for a new generation of users while retaining functionality for experienced professionals.

Key Objectives:

1) Historical Interface Impact Analysis: Evaluate the impact of retaining the historical channel strip-based interface on the learnability of modern audio mixing systems. Examine how this design choice influences the user experience, particularly for newer and younger users.

2) User-Centric Redesign Strategies: Develop user-centric redesign strategies aimed at modernizing the audio mixing interface while preserving functionality. Integrate contemporary design elements that enhance learnability without compromising the efficiency enjoyed by experienced users.

3) Usability Testing and User Feedback: Conduct extensive usability testing of the redesigned interface with both experienced and novice users. Gather user feedback to refine the redesigned interface, ensuring it effectively addresses the learnability barrier for a diverse user demographic.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Mr Christopher Dewey
• Dr Austin Moore

The Internet of Things (loT) is an exciting development of a collection of emerging technologies that bring together physical objects to exchange data with each other via wireless and physical network infrastructure. One example of particular interest is the potential of Wireless Sensor Networks (WSN), combined with computational power at the 'edge' of a network (otherwise known as 'Edge Computing' or 'Fog Computing'), to provide data that is analysed both at source and along the network path, to an eventual destination. Such data is produced in massive volume and at high velocity, which is generally beyond the limits of current technologies for storage and processing.

This PhD project would investigate approaches to performing real-time analytics upon data from a range of devices (from simple sensors and embedded systems, right through to complex multi-cloud distributed repositories), to enable valuable insight to be delivered across networks with limited bandwidth. There will be a need to develop proof-of-concept prototypes for the purposes of benchmarking and evaluation.

Applicants should be comfortable with high-level programming, as well as being able to specify and configure distributed file system software and analytics platforms. A background in networking (particularly ad-hoc/opportunistic wireless) will be useful, as will familiarity with data from Field Programmable Gate Array (FPGA) devices.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Richard Hill
• Dr Muhammad zahid Hussain

Healthcare offers unique challenges for the deployment of machine learning models where the demands for interpretability and performance in general is much higher as compared to most other domains. Given that the cost of model misclassification is potentially high, explanations with respect to how a machine made conclusion is derived play a significant role informing clinicians making unbiased decisions. Knowledge-based systems aim to represent knowledge explicitly via tools such as if-then rules, which allow such a system to reason about how it reaches a conclusion and to provide explanation of its reasoning to end users. Fuzzy systems have been considered effective in building such rule-based systems with one of the most important advantages lying in their inherent interpretability as they support the explicit formulation of, and inference with, domain knowledge, gaining insights into the complex problems and facilitating the explanation of their solutions.

The aim of this PhD project is to develop fuzzy rule-based systems with a particular focus on scenarios of healthcare systems. At the initial phase, the project will look into a number of existing approaches proposed to address the interpretability issues of medical systems, as well as the recently established fuzzy rule-based models. A core part of the project will involve the design and implementation of a specific fuzzy rule-based model that will work with carefully selected healthcare aspects. The implemented system will be evaluated with respect to simulated bench mark data sets first, followed by a close examination of how such a system may perform in collaboration with medical doctors when applied to a diagnostic problem of realistic complexity.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Tianhua Chen
• Professor Richard Hill

Here we seek to draw great benefit from the greatly increasing role of Low Earth Orbit satellites, which can certainly be also related to systems in operation in geostationary orbits and elsewhere, perhaps even at Lagrange points, of the Earth and Sun. The latter is where the James Webb Space Telescope will be functioning. For Low Earth Orbit, we wish to avail of the growing use of CubeSat missions. Their safe functioning and performance monitoring can be supported by the mapping of space debris and related mapping, carried out by LeoLabs Ltd. (https://www.leolabs.space). This project aims are further supporting such satellites and their functioning, as well as their role, increasingly also of relevance for environmental disaster relief, for construction planning and development, for agriculture, and national statistical authorities (the latter, for data sourcing).

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Fionn Murtagh
• Professor Andrew Ball

The development of domain-independent planners is leading to the use of this technology in a wide range of applications. This is despite the complexity issues inherited in plan generation, which are exacerbated by the separation of planner logic from domain knowledge.

The planning performance of domain-independent engines can be improved by exploiting automatically derived knowledge about the domain or problem structure that is not explicitly given in the input formalisation. The type of knowledge to extract and the way for exploiting it are very interesting topics for the Artificial Intelligence community.

At the state of the art, three main approaches have been investigated: reformulation, configuration and combination. Reformulation techniques focus on changing the way the model is described and provided as input; configuration approaches concentrate on adapting the planner to the specific problem, by changing its internal behaviour. Finally, combination methods improve overall planning performance by combining different planning systems and/or different learning approaches.

This project will investigate innovative techniques for the extraction and exploitation of knowledge in Al planners, in order to improve either the runtime -i.e. the time required for solving a problem - or the quality of generated plans.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Francesco Percassi
• Dr Mauro Vallati

This PhD project aims to integrate IOTA Tangle and fog computing to enable secure, scalable, and efficient distributed intelligence in IoT systems. The research involves designing a comprehensive framework that leverages IOTA's feeless, scalable transaction mechanism within the fog computing paradigm to address data security, integrity, and resource management challenges. Through developing and prototyping this integrated system, the project will enhance security and privacy, optimize resource allocation, and enable real-time distributed intelligence at the edge. Extensive simulations, real-world experiments, and case studies will be conducted to evaluate performance, scalability, and practical applicability, contributing to the advancement of IoT infrastructure

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Tariq Alsboui
• Professor Richard Hill
• Dr Muhammad zahid Hussain

This research aims to integrate Large Language Models (LLMs) for automating the annotation of computer vision datasets in Structural Health Monitoring (SHM). By leveraging the advanced natural language processing capabilities of LLMs, the project seeks to create a framework that generates, validates, and refines annotations, significantly reducing the manual effort and time required. This approach will enhance the accuracy and consistency of annotations, leading to more robust and efficient SHM systems. Additionally, the research will focus on optimizing the LLM-based framework for edge deployment, enabling real-time, on-site data processing and improving the scalability and practicality of SHM applications. This integration promises to revolutionize the dataset preparation process, contributing to the development of more effective machine learning models for structural anomaly detection and maintenance.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Muhammad zahid Hussain
• Dr Tariq Alsboui
• Professor Richard Hill

This ambitious, multidisciplinary PhD project aims to develop novel and wearable assistive technologies. which can transform the lives and well-being of an individual. Benefiting from the supervisory team’s expertise in virtual acoustics, embedded DSP, and interfacing techniques, the PhD project will explore and help to redefine what is possible in user motion tracking, adaptive acoustic sonification, intelligent scene and setting identification and edge-based signal processing for improving the QoL (quality-of-life) experience for users and helping to promote well-being. The project will deliver a step-change in offering the potential for ‘accessible to all’ systems.

The project aims to explore emerging technologies, particularly those identified as having the potential for being assistive and connected (IoT), and to combine these with novel and intelligent application to increase the well-being and foster independent living to its users.

Examples of applications of the project outcomes include (i) Indoor and outdoor navigation solutions for blind/partly sighted individuals, (ii) Intelligent ‘assistive narration’ systems to enable independent living (iii) QoL improvement to social care systems through the integration of intelligent visual systems and early intervention technology.

We live in a society that cherishes independent living but with increased life expectancy, and with sectors of society that suffer from certain health conditions (including, amongst other things, sight loss and low mobility) there is more demand that ever to push research in these assistive technology areas. The project will help to provide better health and care systems.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Steven Fenton
• Professor Hyunkook Lee

This ambitious, multidisciplinary PhD project aims to develop novel and wearable assistive technologies. which can transform the lives and well-being of an individual. Benefiting from the supervisory team’s expertise in virtual acoustics, embedded DSP, and interfacing techniques, the PhD project will explore and help to redefine what is possible in user motion tracking, adaptive acoustic sonification, intelligent scene and setting identification and edge-based signal processing for improving the QoL (quality-of-life) experience for users and helping to promote well-being. The project will deliver a step-change in offering the potential for ‘accessible to all’ systems.

The project aims to explore emerging technologies, particularly those identified as having the potential for being assistive and connected (IoT), and to combine these with novel and intelligent application to increase the well-being and foster independent living to its users.

Examples of applications of the project outcomes include (i) Indoor and outdoor navigation solutions for blind/partly sighted individuals, (ii) Intelligent ‘assistive narration’ systems to enable independent living (iii) QoL improvement to social care systems through the integration of intelligent visual systems and early intervention technology.

We live in a society that cherishes independent living but with increased life expectancy, and with sectors of society that suffer from certain health conditions (including, amongst other things, sight loss and low mobility) there is more demand that ever to push research in these assistive technology areas. The project will help to provide better health and care systems.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Steven Fenton
• Professor Hyunkook Lee

Learning Analytics is an increasingly important area of research which has applications both within education and more broadly within organisations and society. In order to develop effective analytical algorithms it is important to better understand personalised learning. What makes a learning experience “personalised”, in the eye of the learner? To what extent and how can perception of personalisation be ascertained and monitored throughout the learning experience? What can we learn from self-regulation about all this? To address these questions, this study will investigate personalisation in learning and learning analytics in relation to this from a self-regulated learning perspective. Through a human factors and ergonomics approach, prototypical self-regulated learning systems will be analysed and compared, and the mechanics that define personalisation of self-regulated learning will be modelled. These mechanics will then be analysed to identify measurable indicators of personalisation and self-regulation of the learning experience. A learning analytics framework will finally be formulated based on the identified indicators, and tested through appropriate mixed methods approaches.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Rupert Ward
• Professor Grigoris Antoniou
• Professor Richard Hill

Modern scientific and engineering research is highly dependent on software. Its importance in driving forward advances in research in the field of computational science and engineering (CSE) has resulted in calls for it to be classified as a first-class, experimental scientific instrument. However, software as a research instrument has not reached a level of maturity compared with the conventional tools of empirical and theoretical science. Research software is principally developed by end-user developers who have a limited understanding and application of fundamental software engineering concepts, principles, and techniques, combined with a "code-first" approach to development, which is in part driven by the perceived complexity and uncertainty of the problem. This results in research software with suboptimal software design, if any, accidental complexity, technical debt, code smells, and an increase in the risk of software entropy. The consequence of this approach is a pathway to stagnation, decay, and the long-term decline of essential research software investment. It has been widely recognised that the future of scientific and engineering enterprise requires a resilient eco-system of software. The aim of this project is to investigate the development of a software testing framework for measuring the sustainability of scientific and engineering software.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Gary Allen

Current Autonomous Agents with cognitive capabilities based on an internal domain model can perform human like reasoning, and explain their reasoned decision on the basis of their domain model and their situational awareness. Unfortunately, state of the art systems have little or no learning / adaptation capabilities in order to maintain their understanding of the world, and hence maintain their reasoning and explanation capability over time. This project will be aimed at techniques for the knowledge acquisition and maintenance of the domain models, and use the capability of automated planning over a long term horizon as a method to measure progress.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Mauro Vallati
• Professor Thomas Mccluskey

Machine learning from sensors and other hard data streams is often used for manufacturing applications in academia and is generating great interest in industry. Great additional benefit could be derived from encoding human knowledge in a suitable way for processing in tandem with “hard data”.

This project will use the machine tool maintenance domain, where descriptions of the conditions can be highly subjective, as a focus for methods of converting extracted human knowledge into a usable data format, including a measure of confidence in the translation as part of the algorithm. This will require the candidate to learn about the issues associate with human perception and its impact upon the fuzzy nature of the extracted data.

The candidate should have a strong background in computer science and a desire to understand the links with human factors and applied cognitive psychology. The project will be supported by experts from the psychology and engineering departments. .

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Simon Parkinson
• Professor Andrew Longstaff

This project will look at developing mathematical models of brewing coffee with the aim of understanding the recent surprising result that in espresso brewing there is a peak in extraction considered as a function of grind size [1]. This model will look at extending existing models of coffee brewing [1,2] to include inhomogeneous and multiphase flow. Methods used will include analytic techniques such as nondimensionalisation and asymptotic reduction and numerical techniques such as finite difference approximations implemented in, for example, matlab.

[1] Cameron, M.I., Morisco, D., Hofstetter, D., Uman, E., Wilkinson, J., Kennedy, Z.C., Fontenot, S.A., Lee, W.T., Hendon, C.H. and Foster, J.M., 2020. Systematically improving espresso: insights from mathematical modeling and experiment. Matter. [2] Moroney, K.M., Lee, W.T., Brien, S.O., Suijver, F. and Marra, J., 2016. Asymptotic analysis of the dominant mechanisms in the coffee extraction process. SIAM Journal on Applied Mathematics, 76(6), pp.2196-2217.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Ann Smith
• Professor William Lee

This project will focus on understanding a specific example of two phase flow instability: the waves of sinking bubbles seen in a settling pint of stout beer. The partial differential equations of dispersed two phase flows will be used to model this system. These equations will be simplified and reduced by nondimensionalisation and asymptotic analysis. Linear stability analysis will be used to model the onset and nature of the instability: this step will probably require numerical methods such as finite difference methods implemented, for instance, in matlab.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Ann Smith
• Professor William Lee

Modern scientific and engineering research is highly dependent on software. Its importance in driving forward advances in research in the field of computational science and engineering (CSE) has resulted in calls for it to be classified as a firstclass, experimental scientific instrument. However, software as a research instrument has not reached a level of maturity compared with the conventional tools of empirical and theoretical science. Research software is principally developed by end-user developers who have a limited understanding and application of fundamental software engineering concepts, principles, and techniques, combined with a "code-first" approach to development, which is in part driven by the perceived complexity and uncertainty of the problem. This results in research software with suboptimal software design, if any, accidental complexity, technical debt, code smells, and an increase in the risk of software entropy. The consequence of this approach is a pathway to stagnation, decay, and the long-term decline of essential research software investment. It has been widely recognised that the future of scientific and engineering enterprise requires a resilient eco-system of software. Research Software Engineering (RSE) aims to facilitate the creation of well-designed, reliable, efficient software to solve research problems. However, there is little empirical evidence to demonstrate that research software is well designed, if at all, understandable, maintainable and extensible. The aim of this project is to investigate the sustainability of essential research software investment.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Gary Allen

Microservices are a new architectural style in which distributed applications are broken up into small independently deployable services, each running in its own process and communicating via lightweight mechanisms. One of the key aspects of a microservices application is that its complexity is pushed from the components (i.e., services) to the integration (i.e., architectural) level. The challenge of getting insights regarding the performance relationship between such services remains open. The aim of this research is to investigate new approaches for detecting and measuring performance degradation events in a microservice architectures.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Gary Allen

Due to NetZero 2050 target, decarbonizing of the power generation and the transport sector is highly demanded in the UK and other countries and one solution to achieve the target is the usage of distributed energy resources (DERs) including renewable energy generation such as PV, wind energy generator, energy storage system, and electric vehicle connecting to the local grid (microgrid) via power electronic converters. The microgrid could be implemented in islanding mode or grid-supporting mode. As both wind and solar energies are not dispatchable, the adoption of renewable energy into the power system with high penetration of electric vehicle charging/discharging will present significant challenges to optimum control of the microgrid as well as stability of the grid system. Therefore, accuracy modeling and optimization control of such a microgrid system for both islanding mode and grid-supporting mode is highly essential. By way of example, due to the uncertainty and variation of renewable energy sources (RESs), optimization of the energy storage system performance in the microgrid is essential to mitigate the RES fluctuation effects in both islanding and grid supporting mode.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Khoa dang Hoang
• Professor Nigel Schofield

Traditional monolithic architectures no longer meet the needs of scalability and rapid development life cycles. Microservices present a new architectural style in which distributed applications are broken up into small independently deployable services, each running in its own process and communicating via lightweight mechanisms. However, the migration process is not trivial. Migrating monolithic software systems into microservices requires the application of decomposition techniques to find and select appropriate service boundaries. These techniques are often based on domain knowledge, static code analysis, and non-functional requirements such as maintainability. This research aims to develop new migration and decomposition techniques to detangle a monolithic system in order to migrate it to a microservice architecture.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Gary Allen

Recognition of Activities of Daily Living (ADL) and indoor localisation in a smart home environment have attracted considerable interest within the wider research field of Ambient Intelligence. One research strand addresses these tasks through knowledge-rich rule-based systems that rely on rules of the form IF one or more sensor-based events are detected THEN a particular activity or activities are recognised or an inhabitant is located. The complexity of expressing such rules in a multi-modal sensor setup often requires the assumption that, at any given time, there is only one individual within the house. This precludes their applicability in multi-inhabitant scenarios, such as house sharing, families or live-in carers. To remove this assumption, a number of open issues need to be addressed, such as: - Indoor localisation needs to be more fine-grained than room-level and disassociated with activity recognition (a person detected in a room may not be the person carrying out the recognised activity). - Person (re-)identification based on video or infrared sensors needs to take into account simultaneous presence of more than one individual. - Increased need for conflict detection and resolution in the recognition and localisation processes.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Georgios Bargiannis
• Dr Ilias Tachmazidis

Interest in the Internet of Things (IoT) is increasing the demand for new design approaches that can assist the specification of resilient, distributed architectures. A market analysis report by Boston Consulting Group (BCG) “Winning in IoT: It’s All About the Business Processes, https://www.bcg.com/perspectives/218353” predicts that $267 billion will be spent on IoT technologies by 2020. IoT devices already produce massive volumes of data, which places significant demands upon network infrastructure. Mobile devices, that move in, out and between networks, place extraordinary demands upon network management services, which need to keep track of individual device identifiers, as well as knowing which devices to trust.

This PhD project will investigate Multiagent Systems approaches to the design, modelling and evaluation of IoT architectures, exploring the use of goal directed behaviour abstractions to model wired and wireless IoT nodes as part of a larger network. There will be a need to develop proof-of-concept simulations for the purposes of benchmarking and evaluation.

Applicants should be comfortable with programming in a high level programming or simulation language (C, Python, R, etc.). A background in mathematics/statistics will be useful. Experience of predicate/modal logic would be beneficial though not essential.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Richard Hill
• Dr Muhammad zahid Hussain

Multibody dynamics methods have established the grounds for advanced dynamic analysis applications, able to simulate mechanical systems. Multibody models are generally composed by a set of interconnected, rigid or flexible, bodies which undergo large translational and rotational displacements. Hence, large and complex mechanical systems are able to be analysed and studied in a computer-aided environment. The aim of this work involves the development and employment of multibody methodologies to produce realistic and accurate railway pantograph models. The pantograph is today a critical mechanical system in the operation of electric traction trains, both at conventional and hight speeds. The models developed are to be validated with experimental data obtained from line tests and/or test bench tests. The work here developed will allow to produce more accurate, realistic, and robust pantograph models, and better understand its mechanical behaviour when interacting the electrified overhead line.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Pedro Cabaco antunes
• Professor Joao Pombo

The utility of AI-driven systems transcends expert users, but non-expert users are heavily relying on them for many reasons. In the quest to democratise AI-powered systems(akin to the GDPR’s right of the user to know), this project will investigate how explainable artificial intelligence will be useful in improving non-expert users’ engagements with AI-powered systems, which are generally opaque, through an interactive explainable artificial framework. The end goal is to make decisions from such systems to be intuitive and discernable through the prism of rigorous theoretical and empirical analysis. The study will improve accessibility, trust and end-user's expertise to fully harness the power of the systems across various application domains.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Isa Inuwa-dutse
• Dr Sangeet Saha

Traditional education systems often adopt a one-size-fits-all approach, which may not effectively cater to the diverse learning needs and preferences of students. In recent years, the integration of artificial intelligence (AI) into educational settings has opened new avenues for personalised learning experiences. AI technologies offer the potential to analyse vast amounts of data on student performance, preferences, and behaviours to develop personalised learning pathways. Tailoring educational content and strategies to individual student needs has been shown to significantly enhance learning outcomes and engagement. By harnessing AI algorithms, educators can gain insights into each student's strengths, weaknesses, and learning styles, enabling the creation of tailored learning experiences that maximise student success. This research proposal aims to explore the application of AI algorithms in personalising student learning experiences, with a focus on optimising educational content delivery, adapting teaching methodologies, and providing timely interventions to support student progression. Objectives: • To investigate existing AI algorithms and techniques for personalised learning in educational settings. • To develop and implement AI-based systems capable of analysing student data to generate personalised learning recommendations. • To evaluate the effectiveness of personalised learning interventions facilitated by AI algorithms in improving student learning outcomes, engagement, and satisfaction. • To explore ethical considerations and challenges associated with the use of AI in personalised education.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Abirami Gunasekaran
• Dr Pritesh Mistry
• Dr Minsi Chen

It is anticipated that the number of individuals living with dementia in the United Kingdom will surpass 1.6 million by 2050, with an estimated 944,000 individuals affected as of 2024. The dementia statistics hub estimates that the economic impact is considerable, with costs estimated at $1.3 trillion in 2019 and a potential increase to $2.8 trillion by 2030. The complex nature of dementia, which encompasses a variety of subtypes such as Alzheimer's disease, vascular dementia, and Lewy body dementia, requires precise diagnostic methods that ensure the efficiency of treatment and care. Incorrect identification of genetic variations linked to dementia can lead to the annual expenditure of millions of pounds on diagnostic tests and treatments that are ineffective and do not address the underlying conditions. The multifaceted nature of the disease may not be fully captured by current diagnostic methods, which frequently rely on clinical assessments and neuroimaging. This proposal outlines a project aimed at developing predictive multi-modal models to classify dementia variants in at-risk older adults living in community settings using the frameworks and methodologies that we have previously published. This approach will integrate diverse data types that include gait patterns of the individuals as well as EEG signals along with demographic, genetic and lifestyle information to enhance the accuracy and early detection of dementia subtypes. Advanced Machine Learning techniques will be used to develop predictive models to distinguish different variants of dementia.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Pritesh Mistry
• Dr Shamaila Iram
• Dr Abirami Gunasekaran

Railway networks play a crucial role in fostering economic development and promoting public welfare. However, they are frequently subject to disruption by events like accidents, natural disasters, or infrastructure failures. These disruptions can lead to significant congestion and economic losses. Understanding and improving the resilience of these networks is therefore essential. Resilience in this context refers to the network's ability to maintain functionality and recover quickly from disruptions. This research aims to identify key features contributing to the resilience of the network railway network and develop a comprehensive framework for assessing this resilience. The study will apply the framework to a case study of the network railway network, demonstrating its applicability and effectiveness. Employing a mixed-methods approach, the study begins with a literature review on resilience in transport systems, identifying key concepts and frameworks. Subsequently, through qualitative and quantitative analyses, the project will identify critical resilience factors within the railway network and develop a tailored resilience framework.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Chamindi Malalgoda
• Dr Rawia El rashidy

Quantum computers leverage the quantum mechanical phenomena of superposition and entanglement to create states that scale exponentially with number of quantum bits. However, the software to be run on a quantum computer is very different from software for classical computers resulting in solutions are hand-crafted in an ad-hoc manner. A solid foundation for building and integrating quantum software is missing. The aim of this research would be to investigate the architectural styles and tactics of quantum software to achieve reliable and robust software architectures for quantum computing.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Gary Allen

Social robots are autonomous agents that interact, collaborate, and communicate with humans. However, designing and implementing social robotic software architectures is a time-intensive activity with recurrent software design issues such as ensuring interoperability, reusability, customizability, extensibility, and the overall sustainability of software components. It is suggested that many of the challenges may be mitigated at design time by choosing relevant architectural styles and architectural patterns, however, there is little evidence to support this. Open research questions include: How to guarantee the syntactic and semantic interoperability of data exchanged by software artefacts running on a robotic platform? How to integrate and ease the deployment of software modules in robotic architecture? How to ease the customization of robot’s behavior? How to enhance the reusability of software components in robot’s architectures? Reference Architectures have been used for the aggregation of knowledge in a range of specific domains, promoting the reuse of design expertise and facilitating the development, standardization, and evolution of software systems. The overall aim of this research is to advance software architectural-level reasoning through the development of a reference architecture for social robots.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Gary Allen

In terms of scale the Boston Consulting Group states that “IoT sensors and devices are expected to exceed mobile devices as the largest category of connect devices in 2018, growing at 23%”. Key markets that will benefit from this technology push are advanced robotics (digital manufacturing), connected and autonomous vehicles, and the emerging health and wellness monitoring market.

IoT networks are expected to expand and contract dynamically, and there is an assumption that wireless networking will be an essential part of any emergent network architecture. This presents a considerable challenge for network security, especially in terms of how hardware (such as sensor networks and actuators) and mobile devices can be trusted sufficiently to join existing networks.

Current research in this area has identified multiparty authentication as a key challenge for the IoT. Whilst multiparty models have been trialled in multi-cloud environments, the additional complication of authenticating constrained hardware devices presents an interesting and worthy area of investigation that has the potential to make a substantial contribution to the research community.

This PhD project will develop a multiparty authentication model for distributed IoT networks. There will be a need to develop proof-of-concept simulations for the purposes of benchmarking and evaluation, as well as a hardware-in-the-loop demonstrator. Applicants should be comfortable with programming in a high level programming or simulation language (C, Python, R, etc.). A background in mathematics/statistics and network security will be useful, as will experience of interfacing embedded systems and sensors.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Richard Hill
• Dr Hussain Al-aqrabi
• Dr Muhammad zahid Hussain

There is a necessity to consider aspects of security when designing software applications. A contributing factor is the increasing legal and financial pressures software developers are under should their software be left vulnerable. This is further exacerbated by the ease of performing large-scale automated software attacks. There are many best practice guides and standardised design patterns that can be followed to ensure a high-level of security is maintained, but such advice is often provided by a subject expert. This research aims to investigate whether learning what an adversary looks for to determine whether a system is vulnerable, as well as how they attack a software system, can be used to build-in simple deterrents that may ultimately increase security with very little software development effort. This project aims to leverage the fundamental philosophy from ‘secure-by-design’ research within crime prevention and construction sectors. It is foreseen that this project will require the input from cyber criminals to determine what they look for within a software system to determine if it is worth attacking. This will then inform a phase of research into establishing key recommendations to consider during software design to prevent the likelihood of being attacked. Case studies will then be performed to evaluate the developed approach.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Simon Parkinson
• Dr Gary Allen

As the Internet of Things (IoT) continues to expand, there are more Internet-connected IoT heterogeneous devices and applications to be exploited. IoT security and privacy has proven to be a major challenge in the IoT industry. There are different types of attacks and threats that may diffuse the IoT architecture, such as spoofing attacks, distributed denial of service attack, botnets, malicious code injection, replication attacks, and eavesdropping. To resist malicious behaviour, the IoT requires a more holistic architecture for addressing trust, confidentiality and privacy issues. This PhD project will identify security risks in IoT networks and focus on developing new, smarter, and more agile security approaches in IoT. There will be a need to develop proof-of-concept simulations for the purposes of benchmarking and evaluation, as well as a hardware-in-the-loop demonstrator. Applicants should be comfortable with programming in a high level programming or simulation language (C, Python, R, etc.). A background in mathematics/statistics and network security will be useful, as will experience of interfacing embedded systems and sensors.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Muhammad zahid Hussain
• Dr Hussain Al-aqrabi
• Professor Richard Hill

The notion of the Digital Twin, which is a digital replica of a physical entity, has become prevalent in various domains including manufacturing, urban planning, healthcare and transport. Digital Twins can be considered as a mirror reflection of the physical object, maintaining and providing an up to date status of their physical counterpart. Such functionality abstracts access to the physical device, which normally has limited resources, supporting enhanced capabilities such as queries by application agents, status visualisation, simulation of alternative scenarios and advanced diagnostics. This highlights the need for well-defined descriptions of Digital Twins and their physical counterparts in order to enable seamless communication, interoperability and data integration from heterogeneous sources. Such challenges can be addressed through the use of semantic technologies. Thus, Semantic Digital Twins become prevalent for overcoming future challenges and driving innovation across a wide range of associated fields.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Grigoris Antoniou
• Dr Ilias Tachmazidis

The Internet of Things (IoT) is the vision of a network of physical objects (“things”) The Internet of Things (IoT) is the vision of a network of physical objects (“things”) equipped with sensors, software and networking capabilities which enable these objects to collect and exchange data. The PhD project would investigate approaches and develop novel methods for (a) enriching IoT data by linking them to ontologies and other data and information sources, and (b) providing reasoning services for processing IoT data at a high level of abstraction. Addressing (a) would be a major step towards achieving a Web of Things which would be siting on top of IoT functionalities (just like the WWW is residing on top of the Internet). Addressing (b) would enable the intelligent processing of huge amounts of IoT data, and requires to overcome major challenges in terms of the size and dynamicity of IoT data, among others. The project is suitable for a PhD student who has already acquired significant knowledge on semantic and knowledge technologies, e.g. in the areas of semantic web, linked data management, knowledge representation and reasoning, or logic programming. equipped with sensors, software and networking capabilities which enable these objects to collect and exchange data. The PhD project would investigate approaches and develop novel methods for (a) enriching IoT data by linking them to ontologies and other data and information sources, and (b) providing reasoning services for processing IoT data at a high level of abstraction. Addressing (a) would be a major step towards achieving a Web of Things which would be siting on top of IoT functionalities (just like the WWW is residing on top of the Internet). Addressing (b) would enable the intelligent processing of huge amounts of IoT data, and requires to overcome major challenges in terms of the size and dynamicity of IoT data, among others. The project is suitable for a PhD student who has already acquired significant knowledge on semantic and knowledge technologies, e.g. in the areas of semantic web, linked data management, knowledge representation and reasoning, or logic programming.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Professor Grigoris Antoniou
• Dr Ilias Tachmazidis

In practice, software architectures are often documented after implementation. Architectural Reconstruction and Recovery (ARR) aims to reverse engineer the software architecture from system artefacts in order to facilitate analysis and understanding. ARR approaches are either based on behaviour-based clustering techniques such as weight or program slicing or based on filtering methods, which recover partial descriptions of the architecture. However, current techniques are limited to extracting code-level information, not architecturally significant abstractions. In addition, the approaches are context-dependent, and no method focuses on those aspects of the architecture which address software qualities. New approaches are required to recover relevant views of the software architecture, including architectural design decisions, that map to architectural concepts, i.e. patterns and tactics. The aim of this project would be to investigate the efficacy and limitations of existing architecture recovery and reconstruction approaches in recovering relevant architectural views, including architectural design decisions, leading to the design of new, enhanced or hybrid techniques.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Gary Allen

Architectural Consistency (AC) aims to measure the architectural drift or erosion of the implemented architecture against the designed architecture. Several AC techniques have been proposed, which can be classified based on how they extract the implemented architecture using static or dynamic code analysis. However, these approaches have several limitations, including being driven by the codebase, not the architecture, assessing the impact of architectural flaws on the recovered architecture, and the difficulty in quantifying architectural inconsistency effects. The aim of this project would be to investigate the effectiveness of existing architecture consistency approaches in quantifying the impact of architectural inconsistency effects leading to the design of new, enhanced or hybrid techniques.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Gary Allen

Architectural-Level Metrics (ALM) aim to provide a quantitative indicator of the quality of software architecture by indicating hot-spots and architectural flaws in the design. While there are a number of methods and techniques for measuring and understanding code complexity issues, identifying code smells and technical debt, their applicability at the architecture level are inconsistent and not proven. By addressing software sustainability at the architectural level, it allows the inhibiting or enabling of systems quality attributes, reasoning about and managing change as the system evolves, predicting system qualities, as well as measuring architecturally significant requirements. The aim of this project is to investigate the fitness of architectural-level metrics in identifying architectural flaws and quantifying software architecture sustainability leading to the design of new or improved measures and metrics.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Gary Allen

In collaboration with the Western Norway University of Applied Sciences, this PhD project is associated with the recently established SmartOcean centre for research-based innovation funded by the Norwegian Research Council and several industry partners (https://sfismartocean.no). The overall vision of the SmartOcean centre is to create a flexible, robust, energy efficient, and cost-effective smart sensor network platform for marine measurement and data handling. An important part of this vision is to develop a multi-tiered architectural software system framework enabling innovation and cost-effective development of robust software applications and data services. The base of the framework will be internet-of-things, cloud computing services, and the emerging web-of-things paradigm. The goal of the PhD project will be the initial design, prototyping, and validation of an architectural framework in collaboration with industry and research partner institutes in the ocean domain. Aspects of the PhD work may also involve evaluation of existing research-based architectural frameworks (e.g. SysML, DANSE, COMPASS, and AMADEOS) for engineering and validation of system-of-system-based architectures within the context of a smart ocean.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Colin Venters
• Dr Gary Allen

Artificial Intelligence (AI) systems are increasingly integrated into various aspects of our lives, from healthcare to finance, influencing critical decisions. However, the black-box nature of many AI algorithms presents challenges in understanding how decisions are made, raising concerns about bias, fairness, and accountability. Trustworthy AI systems are essential for ensuring user confidence and societal acceptance of AI technologies. Therefore, there is a pressing need to develop transparent and explainable AI models that enable users to comprehend and trust AI-driven decisions. The XAI literature generally acknowledges two forms of explainability: ante-hoc explainability, which is integrated within the operation of AI algorithms, leading to explainability by design. This applies to self-explanatory AI algorithms that yield less precise results. On the other hand, post-hoc explainability occurs after the implementation of AI algorithms. It is utilised for more complex AI algorithms (e.g. neural networks) that yield more accurate results.

This proposed research aims to prioritise explainability by embedding both explanation techniques within the operation of AI algorithms, thus producing interpretable predictive models, rather than black box models.

Objectives: • Investigate existing AI models and algorithms to identify limitations in transparency and explainability. • Develop novel methodologies and techniques to enhance the transparency and explainability of AI systems. • Evaluate the effectiveness of transparent and explainable AI models in fostering trust and acceptance among users. • Provide guidelines and best practices for the development and deployment of trustworthy AI systems.

Entry requirements: A bachelor’s and/or master's degree (or equivalent) in Computer Science or a related field, with a focus on statistical analysis and algorithmic design. Proficiency in a programming language commonly used in AI/ML, such as Python, R, or Java is essential.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Dr Abirami Gunasekaran
• Dr Minsi Chen
• Dr Pritesh Mistry

Overview: This research project focuses on music therapy in the digital era. Given recent advances in web-based technologies and the widespread proliferation of mobile computing devices, there exists the potential for powerful and affordable music therapy systems to be realised. Despite this potential, music therapists have yet to fully embrace technology in their practice (Magee & Burland, 2008). The primary focus of this research is to explore the needs of music therapists to generate a series of design guidelines that can be used to inform the creation of web-based music therapy systems.

Key Objectives:

1) User-Centred Exploration: Undertake interviews, surveys, and participatory design sessions to gain a thorough understanding of the perspectives and requirements of music therapists.

2) Theme-based Guideline Formulation: Analyse the outcomes of the user-centred exploration to identify recurring themes and patterns. From these themes, you will systematically develop a series of design guidelines that address the specific wants and needs of music therapists in the digital landscape. Simultaneously, you will identify avenues for future research.

3) Validation: Devise and evaluate a prototype music therapy system that is directly informed be the outcome of the first two objectives.

Magee, W. L., & Burland, K. (2008). Using Electronic Music Technologies in Music Therapy: Opportunities, Limitations and Clinical Indicators. British Journal of Music Therapy, 22(1), 3–15.

Please see our Research Scholarships page to find out about funding or studentship options available.

Our standard University deadlines apply. Please see our Deadlines for Applications page to find out more.

• Mr Christopher Dewey
• Dr Duke Gledhill