PORTLAND, Ore. The National Aeronautics and Space Administration (NASA) sponsored engineers at the Oregon State University (OSU) to design a more efficient air-traffic control system for commercial aircraft. The new system uses smart computer agents to make recommendations to air traffic controllers on how to reduce congestion and avoid bottlenecks, but leaves it up to human experts whether to implement the suggestions.
The stakes are highair traffic control costs the U.S. over $40 billion per year at 5,000 public airports managing 40,000 flights a day. Delays at airports now account for $19 billion in unnecessary operational costs and $12 billion in lost time for passengers. The proposed new system would reportedly reduce those costs by increasing the efficiency with which air traffic controllers avoid delays by 20 percent.
| A new air traffic control concept developed by researchers at Oregon State University and NASA uses computerized "agents" that report positions of aircraft and can help minimize congestion and delays. (Photo courtesy of Oregon State University)|
The current system of shepherding flights to the correct flight paths and runways at the airport would remain in place the same way it is today, but the automated agent would use artificial intelligence to gather information about delays and weather on other flights that could affect an air traffic controller's decisions, making appropriate suggestions to remedy bottlenecks.
For instance, if a destination airport is experiencing delays that would force an outgoing aircraft to circle waiting for runway time when it arrives at the destination, then a suggestion would be made to the air traffic controller that congestion could be decreased by having the departing plane wait a few minutes before taking off. The air traffic controller would then make the decision whether to implement the suggested delay, depending on the overall situation at their airport, of which only they may be totally aware.
Recommendations like this are already being made in individual cases and certain locales, according to professor Kagan Tumer at OSU, but NASA and OSU want to make it standard procedure by using computer systems to continuously accumulate and evaluation this type of information. Agents would bring suggestions to the awareness of traffic controllers as needed, but humans would choose which suggestions to implement and which to ignore. Also, no suggestions would be made regarding the critical aspects of an air traffic controllers job, such as guiding craft during takeoffs and landings.
Funding was provided by the Next Generation Air Transportation Systems Program at NASA and the Cyber Physical Systems Program of the National Science Foundation.