Detection of nuclear material at border crossings using motion correlation

This paper considers the problem that arises when a vehicle carrying nuclear material is detected approaching a border crossing. As quickly as possible, and with automation we wish to identify which vehicle among all those in the area is likely to be carrying the source. We show that if the border crossing area has technology for tracking the position of vehicles, we can correlate observed movements with observed changes in levels of detected radiation---for as the vehicle carrying the material gets closer to the detector, the stronger will be the detected radiation. We use a simulation model that captures the stop-and-go dynamics of a border crossing area to evaluate our ideas, and find a highly successful technique that tracks which vehicles move just when detected radiation changes, coupled with fitting radiation intensity/distance observations to an inverse-square law. This method almost always isolates the sought vehicle just as soon as the minimum number of data observations is obtained.