Revealing Images: INEEL Portal Detects Concealed Weapons
A security staff member displays a small sample of the weapons that have been found by the groundskeepers outside the courthouse. INEEL's Concealed Weapons Detector has proven to be a formidable deterrent to people bringing dangerous items into the courtroom.
The first INEEL-designed concealed weapons detector was installed at the Bannock County Courthouse in Pocatello, Idaho, in 1998. Over the years, it has stopped thousands of weapons from entering the courtrooms, everything from penknives to axes. But more interestingly, it has stopped many weapons from even entering the building. Its reputation for detecting hidden items is so good that groundskeepers continue to find an arsenal stashed behind bushes and trash cans, abandoned by owners who would rather risk losing their property than discovery.
Fast-forward a few years to 2003. The next-generation INEEL concealed weapons detector, now commercialized and marketed as SecureScan 2000 by View Systems, Inc., is installed in Washington Irving High School, in New York, N.Y. Students carrying lunches, books and backpacks walk through the portal as if they were walking through a doorway. But some of the students carry more than the paraphernalia of today’s teenager — the MP3s, pagers and cell phones. Some carry weapons.
The New York City Police Department School Safety Division reported an incident last year after an alarm sounded as a student passed through the portal. The monitor indicated that a weapon was concealed in the student’s mouth. The subsequent search revealed that the student had hidden a razor blade in the upper palate area of his jaw.
According to the NYPD, student slashings are a major threat citywide. The police stated that standard metal detectors often resulted in false detections and an increase in man-hours to conduct hand searches. They said jewelry and coins were being detected instead of razor blades and other weapons.
The new INEEL weapons detector is sophisticated enough to discriminate between threat and nonthreat items such as keys and coins. And it is sensitive enough to identify threat items as small as a box cutter or razor blade.How they work
The new INEEL weapons detector is sophisticated enough to discriminate between threat and nonthreat items such as keys and coins. And it is sensitive enough to identify threat items as small as a box cutter or razor blade.
Typical airport metal detectors use an active electromagnetic technology based on pulsed induction. PI systems send pulses of current through coils of wire within the panels of a portal. The pulse creates a magnetic field, which reverses polarity and collapses when the pulse ends. This collapse results in an electrical spike. The spike causes a short — about 30 microseconds — reflected pulse. If there is metal within the magnetic field, it makes the micropulse last just a bit longer.
Historical data have shown that the responses from typical magnetic monitoring sensors have extreme variability and generate numerous false alarms. Many factors impact and alter performance, including the gait of the person walking through the portal, speed of passage and position in the portal. This makes it difficult to discriminate between threat/nonthreat responses and requires labor-intensive skilled supervision to operate.
The INEEL system is a passive device that senses disturbances in the ambient Earth’s magnetic field — disturbances such as that caused by a weapon passing through the aperture of the portal.
The detector uses 16 magnetic gradiometer sensors, arrayed on both sides of the portal aperture. Data are collected from each of the gradiometers, and the change in the magnetic field over ambient background is determined. After the individual sensor responses are computed, the data from all of the sensors are processed as a group to determine the detected object’s location and size.
The system provides a graphical interface to the operator by using freeze-frame video capture technology, and places filled circles — dependent upon the number of items detected — over the video image indicating where suspected weapons may reside on a person. The circle sizes vary in proportion to the strength of the measured signal.
Since the original weapons detector was placed in the county courthouse, INEEL researcher and electrical engineer Dale Kotter has designed new enhancements to increase system sensitivity, further reduce false alarms and recognize evolving weaponry.The Science behind the System
The INEEL detector uses a proprietary method to process and transform thousands of real-time data points from the portal detector array into a signature pattern for analysis.
Threat items, such as guns and knives, and nonthreat items, such as cell phones and pagers, produce unique magnetic signatures, almost like fingerprints. The signature is variable and can be impacted by some of the same factors as those affecting electromagnetic detectors — gait and speed of passage, proximity to center of portal and background clutter. Additionally, the orientation of the weapon can impact the signature.
Kotter employs several signal process methods to analysis the magnetic signature. The data is filtered to remove environmental noise and passed through threshold analysis to discriminate low-level nonthreat alarms. While the simplest way to reduce false alarms is to ignore events close to background, the INEEL system employs other methods so that it doesn’t miss almost baseline threats.
Kotter worked with Quantum Mechanics — the sensor manufacturer — to improve sensor sensitivity by imbedding microprocessors into the sensors. Now it can detect that box cutter heading into an airplane or that razor blade into the schoolroom.
Over time, research showed that operators at the Bannock County Courthouse began to recognize signature patterns and could identify the cause of the weapons detector alarms. Simply stated, this human ‘learned’ response has been optimized and automated.
Kotter has developed advanced signal processing algorithms using statistical analysis tools and a variant of a Probabilistic Neural Net to analyze the magnetic spectrum. The algorithms perform a type of pattern recognition and calculate a probability factor that the collected signatures correlate to a database of weapons and nonweapons.Test Results
Exacting science must stand up to the rigors of the field and the weapons detector is no exception. Any detector system must smoothly and accurately process snaking lines of travelers or students. Kotter continues to test the INEEL detector for accuracy reproducibility. He, along with colleague Lyle Roybal, replicated field conditions and ran various threat/non threat items through the detector.
They tested against the dead zone — the mid-portal area where many standard detectors are unable to identify a weapon — and found that the advanced signal analysis techniques could identify a cell phone, even at levels approaching baseline.
They performed experiments on items that have a high tendency to generate false positive alarms, such as watches, cell phones and PDAs. Proof-of-concept experiments demonstrated a 94.7 percent correct classification between cell phones and weapons and 91 percent classification among nonthreat items.
They verified that a weapon placed at hip, shoulder, mid-body and feet created unique responses that could be classified.
The researchers also conducted experiments on classifying multiple weapons on an individual. In 86 percent of the test cases, the detector correctly classified a razor blade in a pocket, a cell phone on the hip and a knife in the shoe.
Thus far, the National Institute of Justice has funded the INEEL research. Kotter is looking beyond NIJ for future research, actually quite a bit beyond. Kotter traveled to Russia last December and met with scientists from the Russian Institute of Radio Engineering and Electronics. They have developed an active RF technology to detect tumors in bodies. Kotter believes the signal — about the frequency of radio and just as harmless — could be adapted to detect weapons and contraband. He is pursuing a joint project through the Department of Energy’s Initiative for Proliferation Program (see Need to Know July 2002). View Systems, Inc., the commercial provider of the current version of the INEEL’s weapons detector, is also interested in partnering in the research.
Everyone who has ever boarded a plane, entered a federal building or attended high-profile events such as the Salt Lake Winter Olympics wants an efficient weapons detector; one that can spot a tiny, but deadly weapon but one that won’t result in false alarms from under wire bras and hiking boots.
The INEEL is developing a solution that can pinpoint the location of a hidden knife yet allows to pass the average traveler, laden only with cell phones and keys.