Detection and Identification of Concealed Weapons using

ABSTRACT The detection and identi“cation ofc oncealedweapons is an extremelyhar d problem due to the weaksignatureof the targetburied within the much stronger signal from the human body. This paper furthers the automaticdetection and identi“cation of concealed weapons by proposing the use of an eective approach to obtainthe resonant frequencies in a measurement. The tec hnique, based on Matrix Pencil, a scheme for modelbased parameter estimation also provides amplitude information, hence providing a level of con“dence inthe results. Of speci“c interest is the fact that Matri x Pencil is based on a singular value decomposition,making the scheme robust against noise.Keywords: Concealed weapons detectio n, Matrix Pencil, Model-based parameter estimation 1. INTRODUCTION This paper considers the automatic detection and identi“cation of concealed weapons using a model basedparameter estimation (MBPE) approach. Our long-term goal is to combine MBPE with a broadband noiseradar providing adequate coverage of the frequency band of interest while providing covert interrogation of aregion of interest. This region is envisioned to be an indoor environment such as an airport/railwayterminal.This paper presents our initial results in our approach to concealed weapon detection.Concealed weapon detection has b een studied for several years in.