Glideslope perception during aircraft landing

Ideally, when a pilot approaches a runway on their final approach for landing, they must maintain a constant trajectory, or glideslope, of typically 3°-4°. If pilots misperceive their glideslope and alter their flight path accordingly, they are likely to overshoot or undershoot their desired touch down point on the runway. This experiment examined the accuracy of passive glideslope perceptions during simulated fixed-wing aircraft landings. 17 university students were repeatedly exposed to the following four landing scene conditions: (i) a daylight scene of a runway surrounded by buildings and lying on a 100 km deep texture mapped ground plane; (ii) a night scene with only the side runway lights visible; (iii) a night scene with the side, center, near end and far end runway lights visible and a visible horizon line; or (iv) a night scene with a runway outline (instead of discrete lights) and a visible horizon line. Each of these simulations lasted 2 seconds and represented a 130 km/hr landing approach towards a 30 m wide x 1000 m long runway with a glideslope ranging between 1° and 5°. On each experimental trial, participants viewed two simulated aircraft landings (one presented directly after the other): (a) an ideal 3° glideslope landing simulation; and (b) a comparison landing simulation, where the glideslope was either 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, or 5°. Participants simply judged which of the two landing simulations appeared to have the steepest glideslope. As expected, the daylight landing scene simulations were found to produce significantly more accurate glideslope judgments than any of the night landing simulations. However, performance was found to be unacceptably imprecise and biased for all of our landing simulation scenes. Even in daylight conditions, the smallest glideslope difference that could be reliably detected (i.e. resulted in 75% correct levels of performance) exceeded 2º for 11 of our 16 subjects. It is concluded that glideslope differences of up to 2° can not be accurately perceived based on visual information alone, regardless of scene lighting or detail. The additional visual information provided by the ground surface and buildings in the daytime significantly improved performance, however not to a level that would prevent landing incidents.