TSA wants automated ID verification, ML detection algorithms at security gates
TSA officials requested market research Friday on automated and machine learning capabilities that could update their airport security screening processes. Robert Alexander / Getty Images
A new market research request has the agency exploring automated capabilities for its airport security screening technology.
The Transportation Security Administration is scanning the market to see what automated solutions could update airport security screening.
In a request for information issued Friday, agency officials said they wanted to understand what capabilities could enhance their On-Person Screening and Identity Management systems at airport security checkpoints, including machine-learning algorithms for in-motion security detection and tools to authenticate identity documents.
The TSA designated the RFI strictly for market research purposes, but it focused on capabilities for its Advanced Imaging Technology and the Walk-Through Metal Detector systems, alongside its Credential Authentication Technology systems.
For the AIT and WTMD systems, the agency wants more information on walk-through metallic and non-metallic screening systems, wideband technology to enhance current AIT systems, detection algorithms to help increase in-motion screening performance and synthetic image data generation. Information on automated annotation tools to support detection algorithms is also requested.
Those capabilities would need to operate on an open architecture and be able to help screen anywhere between 150-360 passengers an hour.
TSA officials are also interested in algorithms that can objectively detect ceramics, plastics, paper, leather, carbon fiber and potential explosive material, as well as tools that could reduce the need for passengers to remove light to medium outerwear, headwear and low-threat items like keys and wallets.
For the CAT systems, TSA officials are exploring automating the technology for physical and digital ID scanning to, in part, be able to allocate staff to other security tasks. The market research request also calls for the AutoCAT system to take a live photo of a passenger and compare it with an ID, the back end technology to support updates and the ability for one TSA officer to staff multiple security gates.
Interested stakeholders have until 3 p.m. EST on May 5 to respond.