Improving General Aviation Security

Statistical model shows usual – and unusual – activity

Dr. Justin Chimka, University of Arkansas

 Engineering researchers at the University of Arkansas have developed a statistical model of the typical, day-to-day operations at non-commercial, general aviation airports. Supported by the U.S. Department of Homeland Security and the Mack-Blackwell Rural Transportation Center at the University of Arkansas, the research could help officials detect unusual activity or behavior that might be associated with a security threat.

“We want to understand the variation associated with usual general-aviation activity and operations, so unusual activity can be detected, analyzed and resolved,” said Justin Chimka, associate professor of industrial engineering and a researcher at the rural transportation center. “In other words, how big does a statistical error have to be for it to be considered a threat? Based on historical data, there are basic assumptions or expectations about what should be going on at these airports. Now we have to ask ourselves if reality – what’s really going on – meets these statistical expectations. If not, then we probably should look at it.”

General aviation refers to civilian flying, which includes private and business flights, flight training and crop dusting. The research applied only to these airports, not those that offer scheduled commercial flights.

Since Sept. 11, much attention has been directed toward improving security within commercial aviation, both passenger and freight. General aviation security has not been perceived as a serious threat because general aviation planes carry less fuel and are much smaller than commercial planes. This perception has changed, however, since a suicide attacker crashed a single-engine plane into an IRS building in Austin, Texas, an incident that killed one person and injured 13 others.

Chimka and student Ryan Black analyzed and recreated existing models used to predict growth at general aviation airports. After rendering the models more accurate and efficient, they systematically developed them for security purposes. The models relied on basic demographic information – annual number of landings and takeoffs, total number of planes based at an airport, population within a 100-mile radius, certified pilot schools at an airport, whether an airport has a traffic-control tower, for example – and other detailed data.

The researchers used linear regression models to develop a basic x-y graph demonstrating expected observations of activity associated with errors or deviations. The variables within the models equated the demographic information mentioned above. From the models, Chimka and Black formed a line representing a range of usual or normal activity. Any activity – an extreme spike in the number of landings or takeoffs, for example – that strays too far from the axis of the line is considered dangerous and probably merits attention.

In a subsequent study, Chimka and Black used similar models to analyze border crossings with Canada. Their research could also extend to other modes of transportation, such as highway, maritime systems, mass transit, pipeline systems and rail.

The authors have submitted their study to an aviation studies journal.

Release date: 8/9/2011