NCJ Number
190134
Date Published
January 2001
Length
50 pages
Annotation
This report describes the activities and outcomes of a project commissioned to develop a breadboard model that can use cost-effective technology to show that a wide bandwidth radar sensor can detect a weapon being carried by an individual in a noisy operational environment, as well as to improve the signal-to-noise ratio of the breadboard detector and integrate the breadboard electronics into a brassboard package suitable for field testing.
Abstract
The major objectives of the first phase of the project were to determine the ability of the detector to discern between a weapon and common nuisance objects such as cellular phones, as well as to determine how the various elements of the detector system contributed to overall detection performance. Major objectives of the second phase of the project were to determine the variability in detector system performance through limited field testing and ascertain the impact of operational constraints. The project objectives were largely met. It developed a brassboard detection system that performed consistently in a field environment. The system detected concealed weapons on individuals at distances of 15 feet and detected weapons being concealed behind the back. Although not hand-held, it is portable, self-contained, and can operate under battery power for 8 hours. Its power output is 100mW, which is well within the range for human safety, and it is likely that the manufacturing cost of a single unit would be between $500 and $1,000. The project found, however, that the false alarm rate with the existing signal classification algorithm was currently too high for operational use. The signals due to weapons were buried in a large "noise" component due to the person, and there are large variations in the signature from person to person. This made relying on matching patterns as a means of classification unreliable. It will be necessary to reduce the effect of the body signal, so that of the weapon is prominent. There are signal processing and additional hardware development techniques that can be used to improve the detection performance. 44 figures