CREATE has been awarded $550,000 by the Domestic Nuclear Detection Office (DNDO) of the U.S. Department of Homeland Security (DHS) to provide decision analysis and experiment design support to evaluate algorithms for detecting radiological and nuclear materials. The CREATE team is led by Professor Detlof von Winterfeldt and provides expertise for developing and implementing an evaluation framework to assess nuclear detection systems for the DHS.
“CREATE has been recognized by DHS as a Center of Excellence with expertise in applying risk and decision analysis to study the threats, vulnerabilities and costs of terrorism as well as threat prevention, sensor placement and management/data analytics," explains Dr. Hora, the Center’s director.
The mission of the DHS is to lead a unified national effort to secure the United States. The mission of the DNDO is to develop the global nuclear detection architecture, and acquire and support the development of the domestic detection system to detect and report unauthorized attempts to import or transport a nuclear device or fissile or radiological material.
Radiation detection systems utilize software-based algorithms to alert users of the presence of radioactive material. For some instruments, these algorithms can also identify and classify the type of material that is detected, such as medical isotopes. Algorithms interpret the collected data and determine the response of the detector whether it be to alarm or identify radioactive material.
The effectiveness of algorithms varies among radiation detectors, in particular in terms of capability to extract meaningful information from noisy data (physical as well as synthetically derived spectral information), but all algorithm design involves trade-offs in detection performance. A methodology is needed to more effectively evaluate the capabilities and limitations of algorithms, better tune these algorithms to achieve program objectives, and build upon the methodology to evaluate complete detection systems.
The team assembled, including experiment design experts from Rutgers University, is internationally recognized for their experience and expertise at tackling precisely this type of challenging multi-disciplinary problem. It is this breadth of experience in decision support and data analytics, literally dozens of examples with a multiplicity of approaches, which DNDO plans to draw upon.
Detlof von Winterfeldt is an expert in decision analysis, risk analysis, environmental policy, behavioral decision research, and homeland security. He is Professor at the Viterbi School of Engineering and the Price School of Public Policy at the University of Southern California.