Improved Contamination Monitoring via Sensor Fusion and Machine Learning for Nuclear Threat Reduction (NTR)

Improved Contamination Monitoring via Sensor Fusion and Machine Learning for Nuclear Threat Reduction (NTR)

 

Supervisors: Dr Charles Kind, Dr Loren Picco

The project is available through the Nuclear Threat Reduction Network (NTR-Net) Centre for Doctoral Partnership. NTR-Net is a new research and innovation network that will enhance the activities of the UK’s NTR Programme and create a pipeline of skilled people to support our national capability and preparedness. The initiative is hosted at the University of Bristol and supported by AWE, the Ministry of Defence and the Home Office.

Project

The secure identification, tracking, and management of radioactive materials are critical components of the UK’s nuclear threat reduction (NTR) strategy. To mitigate risks, it is essential to uniquely identify objects of interest and assess their radioactivity – whether intrinsic or due to contamination. Modern sensor technologies, including alpha, beta, and gamma detectors, allow for precise measurement of nuclear radiation emissions. By integrating this radiation data with additional sensor inputs, such as temperature, we can develop a comprehensive profile of the key physical attributes of radioactive materials.

While current off-the-shelf deep neural network (DNN) solutions have proven effective for object recognition and tracking, particularly in controlled environments, this project seeks to advance this technology further. By fusing high-quality sensor data with DNN methodologies, we aim to enhance the safeguarding and detection capabilities critical to managing the UK’s radioactive inventories. This project will push the boundaries of existing software and hardware, enabling the accurate, rapid, and non-intrusive detection of nuclear materials.

The student will gain hands-on experience in the deployment of advanced sensing devices, encompassing both hardware and software approaches. They will explore early and late data fusion methods to integrate information from these sensors. Relevant courses in radiation risk and safety management will be required for visits to sites with active materials. The student will analyse fused sensor data using techniques from deep neural networks, mathematical modelling, and computer vision.

This PhD will draw upon the expertise of University of Bristol staff involved currently in NTR projects with industrial partners such as AWE and the UKAEA. These projects have already made progress in demonstrating the effectiveness of combining cutting edge sensor fused data and neural network models and this PhD will significantly further push these boundaries.

Candidate requirements

A master’s degree in mathematics, physics, computer-science, or a related field is typically required.

Completing relevant coursework in mathematics, physics and data science is essential. Courses in nuclear physics, particle physics, coding, neural networks, computer vision, sensors and detection techniques, and radiation detection techniques would be advantageous.

Skills:

Essential: Proficiency in experimental techniques, scientific data analysis, and mathematical modelling is essential.

Desirable: Experience with the use of neural networks. Being familiar with relevant software tools and environments for deploying off-the-shelf and custom deep neural networks DNN, such as PyTorch, CUDA, Conda etc. Understanding of the various AI methods for object detection, object recognition and tracking. Experience in sensor fusion techniques particularly early fusion using applied mathematics such as world coordinate systems. Experience in computer vision.

Strong mathematical skills, including proficiency in calculus, differential equations, and statistical methods, are necessary.

A well-written statement of purpose outlining the candidate’s research interests, motivation for pursuing a PhD, and career goals in applying AI and sensor fusion for nuclear threat reduction.

Interview: will be conducted for the shortlisted candidates 

The application process may include a successful interview with faculty members or the admissions committee. This interview may assess the candidate’s research interests, academic background, and suitability for the program. Candidates must be UK nationals, or from a NATO country, and should expect to pass UK government security clearance.

Funding:

This fully funded project covers tuition fees and provides a tax-free stipend (depending on circumstances) based on the UKRI rate (£19,237 for 2024/25). Due to the terms and conditions of the NTR-Net funding, candidates shall generally be required to meet special nationality rules.

We encourage you to make informal enquiries to Dr Charles Kind (charles.kind@bristol.ac.uk) if you have any queries or would like to discuss the project details.

How do I apply?

Please submit the following documents together with your application via https://www.bristol.ac.uk/study/postgraduate/research/physics/

•     Academic CV

•     Your personal statement which introduces yourself and outlines your motivation for PhD research, no longer than two sides.

•     A transcript of any qualifying degrees (completed and/or underway).