The Department of Homeland Security’s (DHS) Science and Technology (S&T) branch is working several parallel paths to improving capabilities of next-generation explosives detection systems (EDS) through software enhancements of the image processing, agency officials tell TR2.
These efforts to explore the separation of EDS hardware and software development are expected to lead to the release of a Broad Agency Announcement (BAA) later this year, says Doug Bauer, program manager with S&T’s Manhattan II project which is aimed at bolstering development of next-generation EDS systems.
The software efforts include a Request for Information (RFI) that went out last month seeking potential third party with innovative algorithms, a workshop in April catered to the academic community involved in image processing, and an ongoing effort taking images from EDS vendors technologies and letting select third party algorithm developers demonstrate that they can work with the images.
White papers from the RFI are due later this week (TR2, Feb. 4). Through the RFI S&T is surveying the market space to find out who can offer algorithms that are at Technical Readiness Levels between two and four “to see if we can tap into higher creativity, more innovation and novelty, allowing third parties to become participants” in the development of the next-generation EDS, says Doug Bauer, S&T’s program manager for the software efforts.
The hardware and software in existing computed tomography (CT)-based EDS machines currently deployed at airports is developed and supplied by the manufacturers, General Electric [GE], L-3 Communications [LLL] and Reveal Imaging Technologies. However, in medical applications, where the CT technology has been in use longer than the security field, there is already a separation between hardware development and the image processing or algorithm development, Bauer says.
“And many third parties, including radiologists themselves, develop the algorithms by which CTX-based technology is used to diagnose breast cancer,” Bauer says.
Academic Workshop
The workshop next month will be held at the DHS Explosive Center of Excellence at Northeastern Univ. in Illinois. There will be a number of academics there who are involved in image processing and this will give S&T an opportunity to find out who they are, what the need, and how much information is necessary for them to be able to do their algorithm development, Bauer says.
The question of how much information the software developers need is a tricky one because of the reluctance of the EDS vendors to release information given their “valid commercial proprietary interests in the hardware that they have developed,” Bauer says.
The third effort underway is ongoing and involves the collection of images from EDS vendors and developers and sending those images to several third party algorithm developers “to see if they’re able to reconstruct response operator characteristic curves and test the idea that it is possible to take information that has come” from an EDS vendor and have it worked on by an algorithm developer, Bauer says. He adds that the EDS vendors’ intellectual property rights are being observed here.
With the forthcoming BAA S&T hopes to incentivize the algorithm developers to come up with images that are transformative in their capability, both with respect to capturing all the physics that’s in the raw data coming from the machine, and with respect to dealing with other issues that obviously come in image processing, including human factors considerations, the interest in a standard display, and the interest in those kinds of enhancements that are best appointed for human beings to use in trying to diagnose real threats from nuisance interference they many encounter,” Bauer says.
Bauer says his initial concern with improving the software algorithms in the next-generation EDS systems is the ability to detect threats and to lessen the false alarm rates. He wants to be able to show his customer, which is the Transportation Security Administration (TSA), what can be expected from the next-generation technology.
TSA is currently working with several contractors on next-generation EDS systems. Last fall TSA awarded contracts to L-3, OSI Systems‘ [OSIS] Rapiscan division, and SureScan Corp. to improve their respective next-generation EDS to meet agency requirements (TR2, Oct. 15, 20028). OSI expects to have its first systems ready this year for testing in a foreign airport.
Next-generation systems could be ready for testing later this year in S&T’s Transportation Security Laboratory. How soon the first systems would actually be fielded is unclear and TSA continues to buy current systems. Late last year the agency said it plans to award several potentially large multi-year contracts to current EDS vendors GE and L-3 this June (TR2, Jan. 7).
Current EDS Systems
Bauer and James Tuttle, who heads the Explosives Division at S&T, say that while the algorithm development is aimed at the next-generation EDS work is ongoing to improve the image processing of current systems.
“The threat has definitely expanded,” Tuttle says of the kinds of potential explosives that can be hidden inside checked luggage. “There are things that need to be looked for immediately.”
To improve the ability of existing EDS systems, S&T has a facility at Tyndall AFB, Fla., where it runs real threats through the current systems and then gives the test results back to the manufacturers so they can improve their algorithm development, Tuttle says.
As part of the EDS algorithm plans, S&T is also working on creating an interface standard that will allow third party algorithm developers to have a format to “uniformly access the data they need to undertake the image processing, the reconstruction that is part of the algorithm development,” Bauer says. S&T is working with the National Electrical Manufacturers Association to adapt the Digital Imaging and Communications in Medicine standard to the Digital Imaging in Communications for Security standard, he says. The standard is the interface between the EDS hardware and the software.
Having third party software developers for the next-generation EDS systems not only increases competition leading to better capabilities, which in turn can lead to lower operating costs, the competition itself may drive down upfront acquisition costs. Bauer says that the open architecture approach to software enhancements for the EDS systems may also mean that the hardware platforms can stay in place longer while only the software is replaced.
One of the goals of the next-generation EDS is to have hardware platforms with fewer moving parts, which would mean less maintenance costs and therefore lower operating expenses.