Sensing and Identifying the Improvised Explosive Device Suicide Bombers: People Carrying Wires on their Body

Abstract

We examined the use of radar to detect humans wearing detonation wires as part of a suicide vest in suicide bombings in an effort to stop the bombing. Dogaru et al. (Computer models of the human body signature for sensing through the wall radar applications. Tech. Rpt. ARL-TR-4290, Army Research Laboratory, 2007) used numerical electromagnetic simulations to show ways to use radar backscatter to detect humans carrying weapons behind walls. We developed Numerical Electromagnetic Code (NEC) simulations for the radar cross-section of wire configurations appropriate to the human body and compared them to the radar cross-section simulations for the human body done by Dogaru et al. We also used GunnPlexer Doppler radar at 12.5 GHz to collect laboratory experimental data from standoff distances of 2—8 meters from the following: human subjects, human subjects wearing a wire loop, and human subjects wearing a simulated vest with wire loops. We performed numerous experiments with both horizontal and vertical polarization (HH and VV), analyzing the data after each experimental run. We developed metrics from examining our experimental data of the radar cross-sections that could be used in building models to more accurately find subjects wearing wires. We wanted a metric to provide us with better statistical detection rates. We found several metrics that improved our ability to detect persons wearing wires. We discovered our best metric was the VV/HH ratio of radar cross-section. From our empirical modeling, we found that the ratio for people wearing wires was statistically different from people without wires at a level of significance of α = 0.05. Using this metric, we built a Monte Carlo simulation model that generated a crowd of people and randomly picked those with wires on their person. We used our metric and a threshold value, which we determined experimentally, to distinguish the persons with wires from those without wires. We found from our simulation that our metric provided a success rate of detecting persons wearing wires of approximately 83.4%, based on running 36,000 trial runs in Excel. The rate of false alarms, where the metric in the simulation model picked a subject who was not wearing wires as a suspect wearing wires, was reduced to 28%.

Using the work of Dogaru et al. and our NEC simulations of wire harnesses in the 0.85—1 GHz range, we found that the radar cross-section ratio (body with wires/body without wires) varied from 11 dB for 1—1.15 GHz VV polarization to —5.5 dB for 0.85—1 GHz HH polarization. This frequency range was the best in the range 0.5—9 GHz.

From our research we showed that we can be successful in finding viable metrics for detecting wires on people using radar observations. This preliminary research and the exciting results it produced encourages one to think that suicide bombers can be found prior to their detonation of their bombs and at ranges that are relatively safe.