Advanced Computing in the Age of AI | Sunday, May 19, 2024

Mass Shootings, Contraband Cargo in Crosshairs of AI, Advanced Scanning 

via Shutterstock

Editor's Note: This article originally appeared last December and has been updated to reflect recent developments at Liberty Defense.

AI may be on the cusp of making inroads against two menaces of modern life: mass shootings and terrorism. Two start-ups have technologies under development designed to spot bad actors and bad cargo before they can cause harm.

Liberty Defense is combining radio frequency scanning with real time image processing to detect weapons carried near a school, church, shopping centers or other public places before the criminal enters the building, setting off alarms, locking doors and putting people inside on alert.

And Vanguard Infrastructures is developing an imaging security system using CT scanning that, combined with analytics handled by a Cray XC supercomputer, promises to deliver an advance over current x-ray scanning (more on this below).

At Vancouver, BC-based Liberty Defense, the company is developing Hexwave, which the company said will offer AI-enhanced, active 3D imaging, processed and analyzed from antennae picking up radio frequency (RF) bits, providing “real-time concealed threat images.” The product is based in part on technology developed starting in 2014 at MIT Lincoln Laboratory, with which Liberty Defense has an exclusive license.

The company said the system enables discrete, modular, layered protection for crowded areas and can be scaled to cover multiple entry points to a building or area. The integrated sensor-AI array can detect metallic and non-metallic firearms, knives and explosives – including plastic weapons made, for example, with 3D printing, according to Bill Riker, CEO of Liberty Defense.

He said the system detects concealed weapons by sending out radio frequency energy, which bounces off the weapon and reflects back to the on-site Hexwave appliance, which extracts information from the RF scan bits, and then uses GPUs to generate 3D images. ML inference algorithms trained to recognize weapons are run on CPUs. As the system becomes more widely used, Riker said, more data will be collected in the cloud and used to train the system to become increasingly adept at weapon recognition. System updates will be handled over the air.

Liberty Defence's Hexwave appliance

Initial controlled field trials were conducted by MIT Lincoln Labs with the Massachusetts Bay Transportation Authority, the public transit system for the Boston area. Aman Bhardwaj, Liberty’s president and COO, told us the trials involved testing the system for degradation in 3D imaging due to the subway environment and interference such as metal clutter and metro station electromagnetic fields. “They were focused on how the radio frequency can get disturbed by external interference,” Bhardwaj said, “and they didn’t see any degradation in 3D images.”

Recent developments at Liberty include its announcement last month of a collaboration agreement with the Virginia Division of Capitol Police to beta test Liberty’s HEXWAVE system and assess HEXWAVE’s integration and communication capabilities with existing security systems in a live environment. In June, Liberty announced a memorandum of understanding with the German soccer team FC Bayern München, which plays its home games at a 75,000-seat stadium, to beta test HEXWARE.

Privately funded Liberty Defense recently announced it closed a $7 million (Canadian) private placement, with Canaccord Genuity Corp., Vancouver, acting as the lead agent. The global market size for weapons detection systems is expected to grow to $11 billion by 2025, according Homeland Security Research Corp. Riker said Liberty plans more field trials and initial customer installations over the next 12 to 18 months, with commercial availability around mid-2020.

Riker told us he foresees a day when detection technology like Liberty’s proliferates within widely used, low cost surveillance systems and helps bring mass shootings (there have been more than 300 annually in the U.S. since 2015) under control.

“Currently, there is a significant gap that must be addressed in urban security threat defense, and this is a very capable analysis and sensor tool,” he said. “We see it across a variety of markets and public venues, like stadiums, places of worship, train stations – even home security detection systems. There are cameras everywhere now, and this could add another layer of security. We see mass proliferation of this in the future.”

Meanwhile, Vanguard Infrastructures has chosen Cray to provide compute and analytics capabilities for data collected by scanners within its Stargate Imaging security system.

Scanning and inspecting large volumes of goods in cargo and across borders is a slow, intrusive process that is often incomplete or creates shipping delays. Current scanning techniques typically utilize x-rays, which Vanguard said produce a limited, 2D view of objects in a container without a way to detect the location of individual objects; in addition, objects’ densities are obscured by surrounding, overlapping materials.

The Stargate systems uses CT scan technology to inspect containers and correlate the materials found inside with their associated atomic numbers, according to Vanguard, to spot threatening, hazardous or illegal content without manually inspecting each container.

Troy Cooper, Vanguard’s co-founder and CEO, told EnterpriseTech that while CT scanning has been used to inspect contents in smaller items, such as suitcases, Vanguard’s has patented the scaling of CT bore size, which is like enlarging the view of a camera, to take in entire pallets, containers and even rail cars. Once the scan is completed, the Cray system processes the CT data against a table of atomic numbers, which determines the container contents automatically.

Vanguard's Troy Cooper

Cooper said the density of the scanner can be adjusted to penetrate thick plate metal and that a container’s contents can be scanned and identified within a matter of minutes. To avoid slowing the shipment process, Cooper said, he anticipates Vanguard’s system used mainly at the point of cargo departure because the loading process is slower – and thus can accommodate inspections without causing delays – than unloading.

Cooper said Vanguard ran a trial project with the U.S. Navy, a blind test in which, unbeknownst to the company, 10 objects were inspected by a Vanguard scanner.

“We ran the test, we found 10 items, and told them we can identify seven of them but not the other three,” Cooper said. “The Navy said: ‘This is an award winning-type product because no one’s ever found all 10, you’re the only ones, and the three you couldn’t identify, that’s because they’re top secret. Those materials aren’t listed anywhere in the atomic number table.’”

Cooper said Vanguard plans to deploy the Stargate system at customer sites by Q4 2019.