Imaging, Data and Communications
Advanced electronic/optoelectronic technologies designed to allow stable, intimate integration with living organisms will accelerate progress in biomedical research; they will also serve as the foundations for new approaches in monitoring and treating diseases.
Research Theme
Aim
To develop a combined radar cross-section and synthetic aperture radar machine learning based classifier.
Objectives
1. Adopt data fusion approaches combining RCS and ISAR information for the studied targets and for the various polarization cases with machine-learning (ML) based methods for target classification.
2. Assess classifier robustness with respect to aspect angle, noise levels, data sampling variations, target positional information, and potential reduced information scenarios.
3. Develop a preliminary data set using simulation, polarization-specific, and measured data of relevant targets while using measurement and modelling facilities within The Scottish Microelectronics Centre (SMC).
Description
There is an increasing defense requirement for classification techniques for airborne, terrestrial, and seaborne targets. Typical radar cross-section (RCS) signatures provides information on the electromagnetic scattering of targets while synthetic aperture radar (SAR) provides a two-dimensional representation of the distribution of scattering features in terms of 2D images. These two forms of information are complementary yet different: RCS signatures and measurements describe the aggregate scattering response under particular observation conditions and polarization states, whereas SAR can provide relative spatial details on the scattering features of targets.
The PhD research work will aim to investigate a machine-learning (ML) framework for classifying a broad range of airborne, terrestrial, and sea-borne targets using both RCS signatures and ISAR representations. The main motivation will be to assess if non-imaging-type RCS signatures along with SAR-derived information, can provide more robust classification of targets, and compared to when these representations are independently examined. Numerical modelling, full-wave simulations using commercial software tools, and measurement information strategies will be adopted while using the existing facilities within the RF lab at UoE.
An engineering or physics background is suitable, but engineering is the preference.
Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree.
This position is fully funded.
Research Theme
Multi-agent Systems and Data Intelligence
Aim
To develop a neuro-algebraic multi-agent digital twin for 6G network subsystems and individual network elements to monitor, analyse, certify, and mitigate safety and cybersecurity threats.
Objectives
- To develop an architecture for interactive multi-agent monitoring and analysis of the states of 6G network subsystems, aimed at detecting, analysing, and predicting hazardous situations, anomaly behaviour, and potential intrusions.
- To develop a technology for the formalisation of telecommunication network elements and subsystems as algebraic specifications, enabling symbolic (algebraic) modelling, verification, and reasoning about their behaviour and interactions.
- To develop methods for analysing and responding to anomalous network activity, with the aim of identifying critical situations and supporting timely mitigation of safety and cybersecurity threats.
- To develop an agent architecture based on the interaction between algebraic and neural components, enabling the combination of formal reasoning, data-driven learning, behavioural prediction, and adaptive decision-making.
Description
The project addresses the development of an architecture of interacting agents that operates as a digital twin of a 6G telecommunication network. The proposed system continuously observes the states of network elements and subsystems in real time. Each agent has access to a virtual representation of the network component it supervises, described as an algebraic digital twin that captures structural, behavioural, and interaction properties. In addition, each agent includes a neural component trained on normal behaviour patterns of the corresponding 6G subsystem or element.
By analysing real-time observations using learned behavioural models and verifying them through the algebraic component, agents can detect deviations, unsafe states, anomalous activity, or potential intrusions. When undesirable or dangerous behaviour is identified, the affected agent communicates with other agents and generates coordinated responses. These responses aim to support timely mitigation, neutralisation of attacks, and assurance of safe and secure operation.
Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree.
This position is funded at the Home fee rate.
Research Themes
Aim
Develop multimodal virtual sensing methods that leverage data-driven models to infer rich sensory feedback from minimal physical inputs, thereby enabling adaptive and reliable robotic manipulation in cost-constrained and hazardous environments.
Objectives
- Develop AI-based virtual sensor models that can approximate multimodal feedback (vision, haptics, force-torque, audio) from limited real sensor inputs.
- Design and evaluate multimodal fusion strategies for learning robust object property and interaction representations that generalize across tasks and environments.
- Investigate transfer learning and domain adaptation methods to enable deployment of virtual sensors trained on simulation or rich offline datasets to real-world robotic platforms.
- Validate virtual sensing for manipulation tasks by benchmarking performance against fully instrumented systems in both controlled and hazardous/constrained scenarios.
Description
The PhD project investigates virtual sensing for robotic manipulation, focusing on the use of data-driven models to approximate multimodal sensory feedback. The core objective is to train AI models on rich sensory datasets (e.g., vision, haptics, force-torque, proprioception, audio) to learn robust representations of object properties and interaction dynamics. At deployment, these models will infer missing modalities from minimal physical sensing, enabling reliable manipulation in cost-constrained or hazardous environments. Key research challenges include multimodal fusion, domain adaptation, and the transfer of representations from simulation or offline data to real-world robotic systems. The project aims to advance theoretical understanding of virtual sensing architectures while delivering practical methods for adaptive, resource-efficient robotic manipulation.
Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree.
This position is fully funded.
Research Theme
Novel Computing and Beyond-CMOS hardware
Aim
This project seeks to develop high performance, non-linear metamaterials by harnessing the atomistic resolution of self-assembling DNA nanostructures.
Objectives
- Further development of DNA origami based metamaterials to deliver efficient frequency shifting structures to convert MWIR and NIR light into the visible region;
- Characterise the performance and robustness of these metamaterials;
- And develop methods to integrate the metamaterials into carrier materials to facilitate manufacturing.
Description
The Mk1 eyeball is a remarkable and delicate organ. Vision is our primary sensing modality in many environments, and eye protection from foreign bodies or high-energy radiation is vital. This provides an opportunity space for the development of technologies to integrate enhanced sensing into eyeware. The ability to be able to perceive radiation in currently invisible regions of the electromagnetic spectrum could create significant advantage in threat identification and target acquisition. Non-linear (up-conversion) metamaterials are able to frequency shift infrared light into the visible region. Integrating such materials into eyeware could enable the passive perception of heat signatures and laser-illuminated objects. This project seeks to build on our ability to fabricate metamaterials from DNA nanotechnology. We will design, build and test metamaterial up-conversion structures to understand their performance and robustness. We shall develop strategies to embed our metamaterials in, or coat them onto, macroscale devices for further assessment as potential materials for low-cost eyeware.
Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree.
This position is fully funded.
Research Themes
Aim
Develop an integrated probabilistic decision-support system for detecting and evaluating the origin of nefarious microbial threats.
Objectives
- Fusion of diverse data sources (e.g. observations, biological, genetic) into an integrated system.
- Modelling and propagation of uncertainties, from quantitative data (e.g. measurement uncertainty), subjective information (e.g. intelligence data) and uncertainties in the possible origin or dissemination of pathogenic microbes.
- Development of a fast, usable and generalisable graphical decision support system for supporting real-time and uncertainty-aware decisions.
Description
The aim of this project is to integrate Bayesian probabilistic/AI graphical modelling approaches with diverse datasets to assess and characterise microbial threats - combining subjective and background information, genomic data, proteomic and toxin data, and microbial growth data to provide a comprehensive and dynamic view of pathogenic microbial encounters. The research will develop probabilistic methods for distinguishing between natural, intentional or accidental outbreaks of pathogenic microbes – this is crucial for understanding the current threat from bioweapons.
In some cases, the question of nefarious intent might be answered almost definitively by a specific type of omics-based science, such as in the case where very clear genetic manipulations are evident in a microbe’s DNA. However, in many other cases, the proposition of nefarious intent will need to be assessed in probabilistic terms and there will be a requirement to incorporate a diverse suite of non-definitive but corroborating information in one model. Bayesian graphical modelling approaches such as Bayesian Networks and Chain Event Graphs provide an appropriate framework for integrating this information and have been used successfully in a forensic science context.
The challenges to be addressed by the project include applying these new approaches to real-world decision-making processes, adapting models for different pathogens in real-time investigative situations, and incorporating complex and highly diverse biological and contextual information.
Some background and interest in statistics or probability, interest in applying mathematical methods to biology.
Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree.
This position is fully funded.
Research Themes
Multi-agent Systems and Data Intelligence
Aim
Develop statistical learning methodology for active sonar that provide robust detection and classification across noisy underwater environments while producing calibrated uncertainty estimates to support reliable defence and security decision-making.
Objectives
The project is divided into an initial exploration stage, followed by one of three possible research avenues, with various possible extensions.
- Exploration. Through a review of existing literature, available data sources, and exploratory numerical experiments, investigate the most promising route for maximising the potential of statistical learning methodology in classifying active sonar data while retaining useful quantification of uncertainty.
- Methodology development. Enable transfer across frequency bands, sensor configurations, environments, and target classes, including adaptation from broad pretraining to specific operational detection tasks. Depending on the exploration stage, this may involve
- the development of a foundation model for active sonar data which can be fine-tuned to specific domains,
- a transfer learning model to exploit knowledge from one active sonar domain with large amounts of available data (such as high-frequency side scan sonar) to one with little labelled data (low frequency sonar),
- an active learning approach in which the model works only on the target domain data, first clustering it and building a statistical classifier based on human prioritisation of the detected clusters.
- Enable practically useful and useable uncertainty quantification, e.g. through Bayesian deep learning approaches or conformal prediction, focusing on uncertainty-aware classification and detection methods in high-clutter situations so that model outputs provide operationally meaningful confidence estimates and prediction sets.
Description
Active sonar is a key sensing technology for underwater defence and security, but reliable detection and classification remain challenging as acoustic returns are strongly affected by propagation conditions, reverberation, environmental variability, target aspect, and low signal-to-noise ratios. Labelled active sonar data are limited and often highly scenario-specific, making it difficult to train models that generalise reliably to new environments or sensor configurations.
Existing machine learning research in the acoustics domain focusses largely on passive monitoring due to the financial costs associated with both employing active sonar and labelling collected data.
This project will investigate statistical learning approaches for active sonar classification. The student will develop machine learning methods that learn transferable representations from large collections of underwater acoustic data, using self-supervised, weakly supervised, or simulation-assisted training where appropriate; the approach will be determined based on initial experiments, a literature review, and available data sources.
A central focus will be uncertainty quantification. The project will study methods such as Bayesian neural networks, deep ensembles, conformal prediction, calibration techniques, and probabilistic classifiers to ensure that the system outputs not only a prediction, but also an operationally meaningful confidence measure.
The probabilistic outputs of the methodology will be used in a downstream task in a practically useful way, such as flagging uncertain detections to a human operator.
A strong mathematical background and programming experience, preferably in Python, are desirable. Prior experience in machine learning, signal processing, Bayesian statistics, acoustics, or underwater sensing would be advantageous but is not essential.
Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree.
This position is fully funded.
In-situ optical spectroscopy during Laser Powder Bed Fusion (LPBF) captures high-resolution signals encoding the physical state of the melt pool. However, this data presents a formidable statistical challenge: observations are strongly non-i.i.d., exhibiting spatiotemporal autocorrelation, directional asymmetry from moving laser paths, and complex thermal dynamics. Because labeled defect data is rarely available at scale, this project bypasses supervised approaches to build principled, physics-informed unsupervised learning frameworks.
This project will build on the supervisors’ research in stochastic modelling, statistical learning, numerical analysis, and mathematically grounded anomaly detection. The research will focus on formulating flexible, label-free statistical machine learning for real-time anomaly detection and defect classification. Depending on candidate interest and project evolution, potential methodological avenues include combining spatial modelling with non-parametric hypothesis testing to separate systematic process variation from distributional shifts. To account for complex spatiotemporal dependencies, you will investigate robust calibration schemes (such as block-permutation or wild-bootstrap methods), while exploring techniques ranging from sparse matrix decomposition to geometric clustering and manifold learning to isolate and classify physical anomaly regimes like porosity, lack of fusion, or keyhole collapse.
Bridging spatial statistics, rough path theory, functional data analysis, and advanced process engineering, this project translates rigorous mathematical methodology into immediate industrial impact for metal additive manufacturing.
The project will be in collaboration with the LPBF team at the National Manufacturing Institute of Scotland (NMIS), and colleagues from the Department of Mathematics and Statistics at Strathclyde University.
Funding may become available on a competitive basis.
- a 1st Class undergraduate degree (or equivalent).
- the University’s English language requirements.
Applications are welcomed from self-funded students, or students who are applying for scholarships from the University of Edinburgh or elsewhere.
Reconstructing 3D gas concentration and temperature profiles within dynamic, turbulent plumes from sparse optical measurements is a classic ill-posed inverse problem. Positioned at the intersection of radiative transfer, fluid dynamics, and statistical physics, this project aims to push the boundaries of computational imaging in extreme, harsh environments.
The research will focus on developing forward models governed by coupled radiative and heat transfer PDEs, while exploring alternative, more informative measurement modalities to significantly boost spatial resolution and noise robustness under severe data scarcity. To address the chaotic nature of dynamic flows, you will construct Bayesian inference frameworks and random media models to parameterise chaos and achieve data-consistent uncertainty quantification. Furthermore, the project offers scope to integrate scientific machine learning, such as neural operators and diffusion models, to efficiently represent complex, swirling fluid structures alongside classical physics-based inversions. Bridging deep mathematical theory with real-world applications in aerospace propulsion, combustion efficiency, and environmental monitoring, this project involves direct collaboration with specialists in gas metrology and aerospace engineering to validate computational models against experimental data.
It is ideal for ambitious candidates in Engineering, Applied Mathematics, Computational Physics, or Signal Processing with a strong foundation in differential equations, linear algebra, inverse problems, or statistical learning.
Funding may be available on a competitive basis.
- a 1st Class undergraduate degree (or equivalent).
- the University’s English language requirements.
Applications are welcomed from self-funded students, or students who are applying for scholarships from the University of Edinburgh or elsewhere.
Research Theme
Aim
This PhD project aims to conduct underpinning research to address problems involving the learning of hierarchical, time-varying multi-dimensional state space models for dynamic objects/phenomena, noisy measurements made in complex backgrounds and/or in the presence of calibration errors.
Description
Dynamical system models have been the main pillar of conventional model-based approaches in control, signal processing and sensor fusion: Sensor signal processing and inference algorithms for applications such as multi-object detection and tracking, robotic simultaneous localisation and tracking (SLAM) and calibration of autonomous networked sensors are designed by combining the known physics and stochastic elements into dynamic system models. Model inaccuracies can be mitigated to achieve significant performance gains in inference and decision-making by leveraging data and model size, following the recent advances in machine learning. However, learning from data in dynamical system models to jointly address the epistemic and aleatoric uncertainties involved remains a challenge stemming from a range of factors such as noisy data and labelling, inhomogeneous sampling, model complexity and the intractability of posterior inference.
The incumbent will have the opportunity to steer the direction of the research in consideration of the impact on engineering problems, including learning models for complex backgrounds in radar detection, learning of birth and trajectory models to improve detection and tracking, or semi-supervised/unsupervised training of sensor data classifiers.
This project is fully funded.
Research theme
Novel Computing and Beyond-CMOS hardware
Aim
Validate quantum computation paradigm for information processing and sensor management
Objectives
- Quantum optimisation for sparse antenna-array design
- Joint quantum sensing and detection for RF sensing with applications in Rydberg atoms RF sensor arrays.
- Real-time Detection in electronic support measure
Description
Quantum computing is a promising new computational paradigm that researchers hope can provide meaningful computational advantage for many different problems of interest. Despite the growing interest in both these fields, a relatively unexplored area of research is how these two quantum technologies, i.e. sensing and computation, can interface to bring meaningful advantage to real world problems, more specifically can quantum computing be leveraged to improve classical and quantum sensing protocols. The aim of this project is to exploit Quantum Information Processing (QIP) and Quantum Optimisation (QO), for classical sensor management and quantum sensor information processing. The scope will be on the near- and long- term quantum computers, such as quantum annealers and quantum gate arrays, for complex optimisation problems and real-time detections, which are problems in the distributed and centralised radio frequency sensing.
This project is fully funded.