A newly-developed algorithm has been shown to quickly sort benign from illicit radiation signatures in the same cargo, allowing faster and less expensive detection of weapons-grade nuclear materials at borders.
Angela DiFulvio, the assistant professor of nuclear, plasma and radiological engineering from the University of Illinois hopes that “The findings will be helpful in reducing the false positive alarms at radiation portal monitors, even in scenarios with multiple sources present, and enable the use of cost-effective detectors, such as organic scintillators.”
Highly enriched uranium and weapons-grade plutonium are the key ingredients in the threat of nuclear terrorism, and with more advanced and comparatively easier to implement detection solutions being offered by deep-learning algorithms, it is safe to assume border security mechanisms will be strengthened by this emergent technology currently being explored.
The researchers behind this technology have created and algorithm that focuses on identifying weak radiation signals, seen often from plutonium-encased materials that absorb radiation and subsequently thwart current detection methods. The technology even allows detection to occur in high-radiation environments, amidst even cosmic rays and radon from underground.
Based on the researcher’s extensive testing, they believe that the usage of their algorithmic technology could improve the ability of radiation monitors at national borders, allowing authorities to see the difference between benign radiation sources and potential smuggling of weaponised plutonium and highly enriched radiation used often as catalyst for nuclear terrorism. Naturally occurring radioactive materials (ceramics, fertilizers and even the residual traces of radionuclides found in patients receiving radiology treatments) can be sifted out as false-positives using the deep-learning algorithm developed.
Sorting, or ‘unmixing’ benign sources of radiation, such and the ones seen above, from genuine threats found in weaponised nuclear materials is the crux of the algorithms development and current operational approach. Algorithms have been developed and then trained to identify potential threats hidden in what would have before read as background radiation before, unless the cargo was personally examined by a trained human-eye.
“We crafted an unmixing model that both reflects the basic physics of the problem and was also amenable to fast computation,” stated Alfred Hero, professor of electrical engineering and computer science and professor of engineering. His team is looking forward to further implementing this technology in the coming year.
The team at Tower Technology is forever fascinated with the amazing potential of an algorithmic approach to situations that seemed untenable and difficult to navigate. The right minds and the right code will overcome a lot of seemingly unsurmountable odds.