Risk-Based Air Passenger Screening Could Improve Security

(TSA.gov)

A University of Illinois computer scientist has developed algorithms to assess the relative risk of air passengers that can improve the effectiveness of pre-flight security checks. The work of Sheldon Jacobson in Urbana, Illinois, with Springfield researcher Adrian Lee, appears in the December 2011 issue of the journal Transportation Science (paid subscription required).

Jacobson — who is also a member of Illinois’s mathematics, engineering, and medical faculties — and Lee studied the benefit of matching security resources to the potential passenger risk. The current system assigns maximum risk to each passenger, which leads to the situation where every passenger is considered a possible threat and screened accordingly. This approach, notes Jacobson, “makes the system less secure by over-allocating security resources to those in the system that are low on the risk scale relative to others in the system.”

The current environment that favors broad screening soaks up the time and attention of security officers leads to a larger proportion of high-risk passengers designated for too little screening, while too many low-risk passengers are subjected to too much screening. With so many resources devoted to the large number of low-risk passengers, fewer resources are available to identify or address high-risk passengers.

The Transportation Security Administration (TSA) recognizes this problem and has started its Pre-Check program being tested at seven U.S. airports, to pre-screen frequent travelers, making them known quantities to TSA and thus lower risk. To apply this concept to the overall air travel population, however, requires an assessment of risk of each passenger, and using this information appropriately.

Jacobson and Lee developed three algorithms dealing with risk uncertainty in the air travel population. They then ran simulations of the algorithms at work in a risk-based screening method, to estimate the risk in the overall passenger population and ways to mitigate errors in this estimation procedure to reduce the risk to the overall system.

The simulations show that including these uncertainties can result in sensitive estimations of risk. The findings also show the current policy of overestimating the overall population risk level produces a larger number of improperly screened passengers and a lower level of security in comparison to underestimating passenger risk. The authors discuss as well scenarios that terrorists could try to beat a risk-based process and ways that systems can be designed to mitigate the impact of these attempts.

The results suggest that programs like Pre-Check can improve overall air travel security, by freeing up more assets for passengers with a high risk relative to the overall passenger population. “The ideal situation is to create a system that screens passengers commensurate with their risk,” says Jacobson. “Since we know that very few people are a threat to the system, relative risk rather than absolute risk provides valuable information.”

Read more:

• Univ. Developing Prototype Air Passenger Liquid Scanner
• Simple, Inexpensive Sensor Detects Shoe Bomb Chemical

*     *     *