(photo credit: )
Los Angeles International Airport (LAX) is a major target for terrorists. A computerized system designed by Bar-Ilan University and University of Southern California has made it much harder to break the airport's anti-terror defenses.
Prof. Sarit Kraus, a world-renowned expert in artificial intelligence from the Ramat Gan university's computer science department and a member of BIU's multidisciplinary brain research center, with USC experts recently received a special citation from LA's world airports police division for her contribution to overall airport security at LAX. The system - called ARMOR (Assisted Randomized Motoring Over Routes) - was the brainchild of an international team that included Kraus and was headed by USC Prof. Milind Tambe. Used at the airport since November 2007, ARMOR has resulted in significantly higher success rates for security-related tasks, including the locating and seizure of weapons and narcotics and the identification of suspicious individuals.
The ARMOR system is based on the randomization of countermeasures - those activities through which security forces attempt to foil would-be terrorists. Targeting suspicious activities that may occur well before an attack, the goal of countermeasures is to reveal terrorists' efforts to gather information about airport vulnerabilities and security protocols.
Unlike Israel's Ben-Gurion Airport, LAX is serviced by several major access roads, and there aren't enough resources to stop and check every car, says Kraus. "Similarly, the resources required for sniffing out drugs are also limited - dogs and their handlers can't cover every corner of every terminal at all times. The answer is smart randomization - checkpoints and canine patrols in an unpredictable pattern that provides sufficient protection for the most vulnerable areas while preventing terrorists and drug dealers from predicting where the next inspection will take place," she explains.
One of the advantages of the system is its basis in games theory - an area of science that focuses on how perceived advantage affects human decisionmaking. "ARMOR provides a randomized schedule of where to set up check points and send canine patrols, but security officers also have the ability to override this automated schedule," she explains. "If an officer makes these changes in a non-randomized, predictable pattern - a pattern that could potentially tip off terrorists or drug dealers - a warning is issued by the system. This improves security, and promotes trust of the system by its human users."
In her current research, Prof. Kraus is working with the USC team on scenarios in which a potential attacker has only limited knowledge about the security protocols and vulnerabilities of a particular target. "Our current model focuses on an opponent with complete knowledge about how an airport or other type of site is being protected and acts rationally based on this information. If we can expand this model to include the attacker whose actions are less firmly based on the weighing of rational options, we will be able to improve security still further."
The LA chief of police cited Kraus for having "performed an exceptional service" that facilitates the critical link between the laboratory and the operational world. Thank you for your outstanding contribution to the security of our nation." The performance of the ARMOR system has been noted throughout the homeland security community and it's been presented as a model for emulation to the US Transportation Security Administration, the Joint Chiefs of Staff Level IV Antiterrorism Seminar and the full Congressional Committee on Homeland Security, as well as several other countries.