Detecting concealed small weapons carried by people has received significant interest from law enforcement agencies as well as the military, most frequently for application in controlling checkpoints (in airports, border crossings, public spaces, etc.). Imaging systems for concealed weapons based on radar or other sensor technologies have been recently developed and tested. Most of the existing electromagnetic (EM) sensors suitable for this application operate at very high frequencies, usually in the millimeter or terahertz frequency bands and produce high-resolution images. Although these EM waves can penetrate through clothing (textile materials), they have very poor penetration properties through many common construction materials (such as brick or concrete). Most through-the-wall radars must operate at much lower frequencies, usually below 4 GHz, in order to “see” targets behind walls. However, at those low frequencies, the image resolution is degraded, so small weapons carried by humans may be difficult to detect directly in the image domain.
This work presented polarimetric radar approaches that could be used in detecting small weapons concealed behind walls. The techniques rely on the fact that the radar signature of targets with a large shape factor is strongly polarization-dependent. This property was exploited for weapons such as rifles or rocket-propelled grenades (RPGs). The emphasis was on weapons carried by humans. In this context, the problem could be reformulated from detecting a weapon to discriminating a human carrying a weapon from a human without a weapon. It is important to emphasize that the approaches presented do not apply to small weapons (such as handguns) that do not satisfy the large shape factor requirements. Also essential is that the weapons are passively carried by humans (as, for instance, in the port-arm position), such that the radar illuminates the target mostly at broadside. These techniques do not apply to a rifle or RPG held in the shooting position, where the radar illumination is straight along the barrel.
Four types of targets were considered that are frequently present in a sensing-through-the-wall (STTW) scenario: walls, furniture objects, humans, and rifle-like weapons. The interest in detecting these targets usually increases from walls and furniture objects (which are considered clutter) to humans, and eventually weapons carried by humans. The cross-polarization signature of these targets also increases in the same order (from walls to weapons), whereas the co-polarization signature strengths follow the exactly opposite order. This suggests the use of the cross-polarization to co-polarization ratio as a good way to enhance the response from the most interesting targets (humans and weapons), while at the same time rejecting the clutter (walls and furniture in the scene). This metric can also be used as a discrimination tool in STTW imaging radar or Doppler radar.
A signature optimization procedure was introduced based on polarization direction transformations. Computer models showed that this technique works as expected on an isolated rifle in free space, but breaks down in a more complex scenario where the rifle is carried by a human. A more robust approach is based on the strong cross-polarization response of a tilted rifle. In order to maximize the signature difference between a human carrying a weapon and a human without a weapon, the cross-to co-polarization return ratio was used. By simulating multiple STTW radar imaging scenarios, it was demonstrated that this ratio is enhanced by at least 8 dB in the presence of an AK-47 rifle.
The major advantage of using the ratio of two different signatures of the same target for discrimination is that this procedure is largely independent of target size (in other words, by scaling up the target dimensions, the cross-to co-polarization ratio is not changed). A weapon detection algorithm was developed based on synthetic aperture radar (SAR) images, and was successfully applied to the simulated image of a complex room containing personnel with or without rifles. The idea also was applied to a Doppler radar scenario that looks at a walking human, where the discrimination is based on magnitude differences in the Doppler spectrum.
This work was done by Traian Dogaru and Calvin Le of the Army Research Laboratory. For more information, download the Technical Support Package (free white paper) at www.defensetechbriefs.com/tsp under the Physical Sciences category. ARL-0093
This Brief includes a Technical Support Package (TSP).
Through-the-Wall Small Weapon Detection Based on Polarimetric Radar Techniques
(reference ARL-0093) is currently available for download from the TSP library.
Don't have an account? Sign up here.