Title :
Potential IED threat system (PITS)
Author :
Li, Hua ; Bramsen, Diane ; Alonso, Rafael
Author_Institution :
SET Corp., Arlington, VA, USA
Abstract :
This paper presents a spatial reasoning approach for predicting IED emplacement in an urban terrain. PITS is a tool that dynamically assesses the potential IED threat (PIT), i.e. the likelihood of emplacing IEDs by insurgents in a geographic area. PIT is determined for each region based on a linear function of relevant features which are systematically derived using heuristics and subject matter expert inputs. The feature weight, indicating the emplacer´s preference, is dynamically adapted using historical and ongoing IED attacks. At anytime, PITS can automatically generate a colored IED threat map predicting the locations of future IED emplacements. We have performed systematic evaluation of PITS and the results show that PITS can predict IED emplacements significantly better than the random predictor. PITS can work as a standalone tool or as a module to be integrated into other IED prediction systems. The IED threat map and the feature weights can be used to aid decision makers in counter-IED efforts.
Keywords :
explosives; military computing; spatial reasoning; colored IED threat map; linear function; potential IED threat system; spatial reasoning; Cities and towns; Data mining; Feature extraction; History; Joining processes; Pattern analysis; Pattern recognition; Performance evaluation; Prediction algorithms; Roads;
Conference_Titel :
Technologies for Homeland Security, 2009. HST '09. IEEE Conference on
Conference_Location :
Boston, MA
Print_ISBN :
978-1-4244-4178-5
DOI :
10.1109/THS.2009.5168041