PhD studentships

PhD studentships are a type of scholarship for your research. Generally, a PhD studentship will provide at least the full standard UK/EU fees, and will usually include a maintenance stipend as well. The studentship will normally focus on an area of research that is of interest to the sponsoring party. Stay tuned for more PhD studentships in the future.

Faculty of Computing, Engineering and the Built Environment

An integrated approach to improve hydraulic and hydrologic design of constructed wetlands 

Closing date: Wednesday 5 January, 23.59pm

Subject description: The research subject Integrated Approach to Improve Hydraulic and Hydrologic Design of Constructed Wetlands includes the study of processes related to hydraulic criteria and mixing patterns for optimization of constructed wetlands efficiency and prediction of pollution mitigation into downstream waterways through quantification of hydrodynamics and transport processes.

Project Description: We are recruiting a PhD student on optimization of hydraulic criteria and hydrologic design of constructed wetlands. The student is expected to generate a unique database about treatment, mixing and physical characteristics of the systems obtained through field experiments, including tracer studies. Numerical modelling tools are to be applied on the validated obtained datasets of mixing, treatment and physical parameters.

The project will require field visits to collect data, analyses of data and development of a model to optimize the hydraulic design of the systems and the prediction of reduction of pollutant load based on the empirical datasets. It provides an excellent opportunity for the student to establish an international research profile through national and international projects and through our industrial collaborations with our external partners, including The Coal Authority, Constructed Wetland Association, Severn Trent, Thames21, and other partners.

Duties: The project will be conducted by the PhD student, together with supervisors and technical support personnel.

For information on entry requirements and more, please download the accompanying Word document.

Investigation of flow and heat transfer in micro channels using nano-fluids 

Closing date: Sunday 5 January, 23.59pm

The objective of this research work will be to build a general purpose experimental test rig to assess the thermal performance, flow and pressure drop characteristics of the nano-fluid flow in a microchannel heat sink. The experiment will be carried out using selected nano-fluids of various volume concentrations. The effects of the thermophysical properties of the nano-fluid on the thermal performance and flow characteristics will also be investigated. Numerical methods using CFD will be used alongside experiments to investigate innovative designs of the microchannel. It is expected that the outputs of this project will further advance and support the development of the micro channel cooling capability.

For information on entry requirements and more, please download the accompanying Word document.

Developing data assimilation and machine learning approaches to analysing complex multimodal biometric data towards a more personalised and adaptive mental health care

Closing date: Sunday 19 January, 23.59pm 

The PhD research will be to design and implement personalised adaptive computer-based interventions to support mental health issues, for example, in people experiencing anxiety disorders. These interventions may be software-based apps to support mindfulness practices, gamebased systems or virtual / augmented reality experiences. However, to create the adaptive component of these systems, the team will look at the types of emotion, as measured by multimodal biometric and physiological responses, which are typically being experienced by the volunteers. Once profiled, computerised-based interventions will be developed, which will adaptively respond and encourage a shift from negative emotions to positive ones as a result of the implementation of personalised biofeedback loops.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Enhancing audio signal processing chains in virtual environments

Closing date: Sunday 19 January, 23.59pm 

In multisensory systems, such as 3D applications and Virtual Reality, spatial cues describing the environment surrounding the user are carried through sound. Human hearing uses auditory stimuli to perceive space and identify entities that cannot be detected by other senses. Hence, in realistic representations of virtual environments (VEs), rendering spatial hearing enhances interaction and immersion. Spatial information of environments can be conveyed to listeners through real-time convolution of impulse responses in the signal processing chain (Vorländer et al, 2014). Siltanen (2005), in addition, proposes an approach that, based on reducing the geometry of the environment, allows real-time acoustic modelling. In virtual reality, presence is also induced by auditory information through spatialised audio (Larsson et al 2010, Stanney et al, 1995). A study is therefore proposed towards improvements of audio processing chains for virtual or augmented reality.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Innovative virtual and augments simulations for medical education

Closing date: Sunday 19 January, 23.59pm 

The undergraduate medical curriculum relies on students learning a range of essential clinical skills that they will need to put into practise when they become qualified doctors. Many of the skills can be difficult to master and are currently learnt from either textbooks or from tutoring by doctors. Neither approach is ideal because 1) text-based approaches do not reflect the practical nature of the clinical skills and 2) the time that the clinicians have available to them to instruct the students is limited. Therefore, computer-based simulations are playing an increasingly important role in medical education and training.

This project will build upon the existing work of Dr Wilson (Wilson et al., 2017a; Wilson et al., 2017b; Wilson et al., 2018) looking to further develop important computer simulations to support clinical training. The computer-based simulations will be based around augmented, virtual reality and / or serious games to teach undergraduate medical students a range of essential clinical skills. They will be evaluated in the target population for their efficacy, usability and user experience.

The potential candidate would be expected to be familiar with programming computer games technology, virtual and augmented reality as well as game based learning.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Computational analysis of style in traditional fiddle playing 

Closing date: Sunday 19 January, 23.59pm 

The fiddle, dimensionally the same as a violin but played in a folk style, has been historically central to the traditional music of the British Isles and many European countries. Since the middle ages, it has been one of the most popular instruments for accompanying traditional dancing (Ling, 1997). As one of the most widely played folk instruments, musicians who play in a range of folk styles are widely accessible, including those participating in the Royal Birmingham Conservatoire’s Folk Ensemble.

This project is an excellent way to provide a PhD student with a project that comes from a team with a proven track record and a subject area where there is relatively easy access to musicians and associated data. It is also an excellent way to explore an important next step in research that brings together computational, musicological and ethnographic study.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Investigate the use of deep learning to achieve an optimum eastern language learning experience in increased fluency of  productive aural skills

Closing date: Sunday 19 January, 23.59pm 

This research aims to investigate the use of deep learning in speech recognition to enable an improved experience in learning an Eastern language (such as Punjabi). You will investigate the issues currently being faced in the use of speech recognition systems to learn a language. 

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Data-Driven processes design for Real-Time Service Provisions in Urban Computing Environment 

Closing date: Sunday 19 January, 23.59pm 

Urban computing uses ubiquitous computing technologies to gain a better understanding of how to improve our cities (Marciniak and Owoc, 2013). Urban Computing consists of four layers sensing, data management, analytics and service provision. While the sensing and data management layers are well established, it can be argued that to date there is a limited effort in utilising the data analytics layer to inform the design of processes to influence the service provision. It is propositioned that continuous analysis of data in real-time, allows immediate action to be taken when attempting to improve service provisions. Several techniques can be used to bridge this gap by utilising the structured and unstructured data to design processes, referred to as Data-driven Process Design.

Thus, it facilitates identifying process changes in real-time to improve performance and how we can offer service provisions. However, the lack of process layer in urban computing frameworks averts the efforts of sensing, analysing, and managing data to be practically applied for improving service provisions in urban areas. In addition, data-driven approaches can identify complex and nonlinear patterns in data that can be utilised to design processes and process modelling, data management, and process mining. Process mining is useful for conformance checking, performance analysis and predictions, which help diagnose problems and improve processes. Hence, this research will explore some of these data-driven techniques, develop a model including algorithms for mining processes in order to offer a systematic approach for designing processes to support real-time service provisions.

The potential candidate would be expected to be familiar with programming computer games technology, virtual and augmented reality as well as game based learning.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Deep learning approach electricity consumption monitoring

Closing date: Sunday 19 January, 23.59pm 

Many stakeholders can benefit from knowing the energy consumption of different devices within a home. It helps users understand their bills, retailers to plan tariff systems and distributors to plan network expansion. However, placing meters on all devices is expensive. Instead, we will determine device level power consumption based on half-hourly aggregated data available from smart meters. For many devices, validation of power consumption can be detected visually by trained observers. This project will seek “ground truth” use of several device types by visually inspecting the estimates of state-of-the-art disaggregation techniques, and sub-metering a small number of homes.

This will allow the accuracy of the algorithms to be assessed and provide training data for more sophisticated supervised learning techniques.

The main objective of this project is to build NILM model based on Deep Neural Network Architectures. Stemming from this objective, the project aims to improve the accuracy of Non-Intrusive Load Monitoring (NILM) models in the context of Deep Neural Network architecture. Secondly, what reinforcement learning techniques can be used for this specific case of energy disaggregation.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Innovative virtual and augments simulations for medical education

Closing date: Sunday 19 January, 23.59pm 

The undergraduate medical curriculum relies on students learning a range of essential clinical skills that they will need to put into practise when they become qualified doctors. Many of the skills can be difficult to master and are currently learnt from either textbooks or from tutoring by doctors. Neither approach is ideal because 1) text-based approaches do not reflect the practical nature of the clinical skills and 2) the time that the clinicians have available to them to instruct the students is limited. Therefore, computer-based simulations are playing an increasingly important role in medical education and training.

This project will build upon the existing work of Dr Wilson (Wilson et al., 2017a; Wilson et al., 2017b; Wilson et al., 2018) looking to further develop important computer simulations to support clinical training. The computer-based simulations will be based around augmented, virtual reality and / or serious games to teach undergraduate medical students a range of essential clinical skills. They will be evaluated in the target population for their efficacy, usability and user experience.

The potential candidate would be expected to be familiar with programming computer games technology, virtual and augmented reality as well as game based learning.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Attack detection from models using machine learning

Closing date: Sunday 19 January, 23.59pm 

The PhD project tackles runtime detection of cyber attacks. It combines security by design with runtime security checking by using design models for runtime attack detection. Model2Defend proposes research on: (i) graphical design languages with a trace semantics to express security models, and (ii) an attack detector that uses those models and machine learning (ML). Model2Defend aims to give security analysts the means to: (i) security analyse systems at design-time, (ii) represent knowledge about possibly ongoing security threats and (iii) actively defend against threats and attacks.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Characterisation of VBM Algorithms for Processing of Medical MRI Image

Closing date: Sunday 19 January, 23.59pm 

Voxel-Based Morphometry (VBM) is the process of measuring volumetric changes in structural imagery. We propose to characterise different algorithms that measure the accuracy at the scale of gross morphological structures. For images to be studied, the brains have to be labelled first. Before registrations, images are usually skull-stripped, leaving only the brain in the image. Images are then linearly registered using FSL software. Rigid registration is usually using a standard MNI152 template. (Klein et al. 2009) used volume and surface overlap, volume similarity, and distance measures to evaluate how well individual anatomical regions, as well as total brain volumes, register to one another. Metrics for measuring differences in algorithm performance can be average brain volume, grey matter overlap, white matter overlap, correlation of a measure of curvature, local measures of distance and shape between corresponding principal sulci. In conclusion, (Klein et al. 2009) mentioned that the results of comparisons were better or comparable with skull-stripped images.

MR and functional MR image analysis can be a significant portion of the diagnoses of psychological related diseases. One such disease is Autism. When the most significant regions regarding specific condition are identified, appropriate machine learning algorithms can be applied for its analysis.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Transfer learning for Time Series Classification

Closing date: Sunday 19 January, 23.59pm 

During the last two decades, Time Series Classification (TSC) has been considered as one of the most challenging problems in data mining [2]. This is due to its’ large number of practical applications in various domains such as cyber security, medical, activity recognition, energy and transportation. Stock market anomaly detection in business, identifying heartbeat patterns of patients in hospitals and detecting temperature levels in climate science are some of its’ practical examples. Accurate time series classification can increase the business revenue as well as facilitate optimal resource allocation. Notable algorithms have been developed to address the classification problem, while the vast majority of research has focused on developing similarity measures for accurate classification. Significant challenges face time series classification including the diversity of data that is inherited from the diversity of domains from-where data has been collected. Scarcity of labelled data is one of the most common challenges in TSC. The significant difference between deployment and target domains is another challenge. One way to overcome these challenge is to utilise transfer learning.

Transfer learning is the process of first training on a source domain, and then transferring the learnt knowledge to the target domain [4]. Hence, transfer learning can leverage the already existing data of some related task or domain to understand a new domain. This idea has been shown to improve capabilities of machine learning techniques in many tasks such as computer vision and pattern recognition especially with the advance in deep learning approaches. Transfer learning includes many types which include feature transfer, parameter transfer. meta data transfer and model transfer.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Interaction paradigms for intuitive augmented reality

Closing date: Sunday 19 January, 23.59pm 

Augmented Reality (AR) is getting close to real use cases,which is driving the creation of innovative applications and the unprecedented growth of Head-Mounted Display (HMD) devices in consumer availability. However, at present there is a lack of guidelines, common form factors and standard interaction paradigms between devices, which has resulted in each HMD manufacturer and AR developer creating their own specifications.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Pervasive and Ubiquitous Computing

Closing date: Sunday 19 January, 23.59pm 

The proposed research looks the previously identified challenges, particularly standardization and interoperability between networks and devices, and their application to Urban Computing and Industrial IoT applications.

The main question this research posits is:

How to manage the non-uniform distribution of sensors and heterogeneity in Pervasive and Ubiquitous computing applications?

Pervasive and Ubiquitous Computing (PUC) often relies on readily available data sources for the implementation of its services. Notwithstanding, there is heterogeneity both in the type of available data sources and its distribution (physical and contextual location). This proposed project aims to research the impact of said distribution in PUC applications, postulating the sub-research question of how to provide alternatives to mitigate or leverage system heterogeneity towards the implementation of system to manage these data sources and their data with the use of data analytics techniques.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Enhancing Radio Access Network Slicing via Machine Learning Predictions

Closing date: Sunday 19 January, 23.59pm 

Networks are currently required to dynamically adapt to meet the distinct requirements of diverse traffic classes. Emerging classes of traffic such as those generated by autonomous vehicles and various machine-to-machine communications are adding another dimension to the already congested wireless channels [1]. To meet the requirement of this class diversity and high demand, slicing the radio access network (RAN), which complements core and transport slicing, has gained significant popularity, both from academia and industry [2]. In fact, there are now a number of network equipment manufacturers offering different forms of slicing capabilities, but we are yet to see the full potential of this exciting concept. And despite these available options in the market, end-to-end network slicing remain in the development phase and largely still under investigation. However, it is expected to dominate future network access mechanisms, both in the core and the edge of the network.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application

Incorporating privileged information in the context of deep learning for medical imaging 

Closing date: Sunday 19 January, 23.59pm 

In this PhD project we aim to design an algorithm that is specifically designed for incorporating expert medical knowledge (privileged information) into the learning course of image classification deep learning models. It is worth mentioning that the privileged information will only be available during the training phase along with the classification label of each image but will be absent during the testing phase. A similar problem has been investigated in [5] where the privileged information was given in the form of a segmentation mask and was incorporated into the training stage to improve the performance of Convolutional Neural Networks (CNNs). Another approach in [6] used a heteroscedastic dropout approach to incorporate images as privileged information in the learning phase of CNNs and Recurrent Neural Networks (RNNs). Unlike the exiting approaches which are limited to integrating imaging data as privileged information, in this project we aim to incorporate various types and formats of privileged information (such as texts, images and numerical features), depending on the knowledge domain, to improve the performance of image classification deep learning models. The framework that is to be developed in this project will be applied and tested in the context of medical imaging practical problems.

More information on the project, from potential impact to references, can be found on the accompanying PDF.

To apply, please complete the project proposal form and the online application