Machine Learning to Enable Cost-Effective Detectors at Borders
Nations need to protect their citizens from the threat of nuclear terrorism. Nuclear security deters and detects the smuggling of unique nuclear materials -- highly enriched uranium, weapons-grade plutonium or materials that produce a lot of radiation -- across national borders.
The researchers developed an algorithm capable of identifying weak radiation signals, such as might be seen from plutonium encased materials that absorb radiation. It works even in the presence of a high radiation background, including everyday sources such as cosmic rays from space and radon from the rock underfoot.
Based on the research results, our algorithm could improve the ability of radiation portal monitors at national borders to tell the difference between potential smuggling activity and benign radiation sources. For instance, naturally occurring radioactive materials such as ceramics and fertilizers, or radionuclides in recently treated nuclear medicine patients, can set off "nuisance" alarms at radiation scanning facilities.
"There's also the concern that somebody might want to mask a radioactive source, or special nuclear material, by using naturally occurring radioactive materials such as granite or even kitty litter."
As vehicles or boxes are scanned, the data from the detector can be put through these algorithms that unmix the different sources. The algorithms can quickly identify whether special nuclear materials are present.
The team turned to specialists in machine learning, who could use data collected by the team to "train" algorithms to look for the signatures of materials that have the potential to make a nuclear bomb.
The University Defence Research Collaboration (UDRC), led by the University of Edinburgh, develops unprecedented research in signal processing with application to the defence industry, shares knowledge, and promotes communications, guidance, and training.
By bringing together researchers from different aspects of signal processing, this project addresses the research challenges of operating in a networked battlespace. This collaborative centre of excellence for signal processing embraces academia, Research and Technology Organisations, defence manufacturing industries and the Defence Technology Centres.
This collaboration has developed cutting edge signal and data processing capability in the defence industry. In addition to traditional physics-based sensors (such as radar, sonar, and electro-optic, 'human sensors', e.g. from phones), our research will provide valuable signals and information that could advance situational awareness, information superiority, and autonomy. Persistent real-time, multi-sensor, multi-modal surveillance capabilities will be at the core of the future operating environment for the Ministry of Defence and core technology in modern society.
New signal processing techniques are required in a future where a large-scale deployment of multi-modal, multi-source sensors are used across various environments. The UDRC research considers the fundamental questions of scalability, adaptability, and resource management of multi-source data. This research focuses on high-volume, high-velocity data from non-traditional sources and high uncertainty. We also use powerful machine learning techniques, including deep learning, to enable a faster and more robust understanding of new tasks, anomalies, threats, and opportunities relevant to operational security.
The UDRC Phase 3 project, Signal Processing in an Information Age, is an ambitious initiative that brings together internationally leading experts from 5 leading centres for signal processing, data science and machine learning with ten industry partners.
Institute of Digital Communications, University of Edinburgh
School of Informatics, University of Edinburgh
University of Strathclyde
Queen's University Belfast.
The consortium works in defence signal processing research through the MOD's Centre for Defence Enterprise and the US Office of Naval Research. The team have significant experience in technology transfer, including tracking and surveillance (Dstl), advanced radar processing (Leonardo, SEA); broadband beamforming (Thales); automotive Lidar and radar systems (ST Microelectronics, Jaguar Land Rover), and deep learning face recognition for security (AnyVision).