Radiological Source Localisation
Abstract
The increasing threat of chemical, biological, and radiological (CBR) attacks has resulted in a significant interest in research on countering such attacks. This research focuses on countering radiological attacks, which may, for example, be carried out using radiological dispersion devices or dirty bombs. The ability to rapidly localize a radiological source can assist emergency responders to disable, isolate, or safely remove such a device. This report describes some preliminary work the authors have carried out in the area of radiological source modeling and localization. This work concerned localization of a single fixed gamma radiation source of unknown activity level. The accuracy with which the source location and the activity could be estimated was studied using the Cramer-Rao bound analysis. A simple deterministic analytical approach as well as several probabilistic estimation techniques were investigated using simulated and real measurement data. The inverse square law-based deterministic solution developed in this work uses radiation measurements collected at four arbitrary points to estimate the source position and activity. This algorithm was able to provide reasonable source estimates based on real data collected using the Low Cost Advanced Airborne Radiological Survey (LCAARS) system developed by DSTO. The maximum likelihood estimator and a nonlinear least squares estimator yielded quite accurate estimates. While the maximum likelihood is an asymptotically efficient estimator, it is a batch algorithm and is unattractive for operational use. An inexact recursive least squares algorithm was developed and it produced good estimates when applied to real data. Unscented Kalman filter (UKF) and extended Kalman filter (EKF) algorithms also were investigated. The UKF approach performed well but the EKF was divergent and could not return acceptable source estimates.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 2007
- Accession Number
- ADA471550
Entities
People
- Ajith Gunatilaka
- Branko Ristic
- Ralph Gailis
Organizations
- Defence Science and Technology Group