Midair collisions: limitations of the see-and-avoid concept in civil aviation

Midair collisions cause loss of life and property and undermine general aviation; they also represent failures of the see-and-avoid concept in air traffic control. This study identifies limitations of the see-and-avoid concept.

Analysis of National Transportation Safety Board data on 159 U.S. civil aviation midair collisions and limitations of the see-and-avoid concept.

On average, there were 15.6 midair collisions annually from 1991 through 2000. At least one aircraft was maneuvering in 88% of collisions, and both in 70%. There were 77% that involved arrival to, departure from, or flight over an airport, with 61% in the traffic pattern. Head-tail collisions were more frequent in the traffic pattern than out (28.3%, 2.8%, p < 0.05). Other horizontal convergence angles were equally frequent in or out of the pattern: head-on (8.3%, 11.0%), obtuse (11.0%, 8.3%), and acute (13.8%, 16.6%).

Because the relative bearing to each aircraft on an unaccelerated collision course is constant, pilots sometimes cannot see converging aircraft when climbing, descending, or level. Even if a converging aircraft is unobstructed, it appears small, motionless, camouflaged, and inconspicuous until imminent impact. A statistical model reveals that the probabilities of seeing and avoiding a converging 40-ft aircraft, for an optimal observer or theoretical pilot scanning 2/3 or 1/3 of the time, respectively, are less than 0.91, 0.60, 0.30 at 200 kn; 0.49, 0.32, 0.16 at 300 kn; 0.28, 0.18, 0.09 at 400 kn; and 0.15, 0.10, 0.05 at 500 kn. The see-and-avoid concept has striking physical and behavioral limitations.