Flight-Path Estimation in Passive Low-Altitude Flight by Visual Cues

A series of experiments has been conducted, in which subjects had to estimate the flight path while passively being flown in straight or in curved motion over several types of nominally flat, textured terrain. Three computer-generated terrain types have been investigated: 1) a random "pole" field, 2) a flat field consisting of random rectangular patches, and 3) a field of random parallelepipeds. Experimental parameters were the velocity-to-height (V/h) ratio, the viewing distance, and the terrain type. Furthermore, the effect of obscuring parts of the visual field has been investigated. Assumptions have been made about the basic visual-field information by analyzing the pattern of line-of-sight (LOS) rate vectors (streamers) in the visual field. The experimental results support these assumptions and show that, for both a straight as well as for a curved flight path, the estimation accuracy and estimation times improve, to a certain extent, with the V/h ratio. Error scores for the curved flight path are found to be about 3 deg in visual angle higher than for the straight flight path, and the sensitivity to the V/h ratio is found to be considerably larger. For the straight motion, the flight path could be estimated successfully from local areas in the far field, although at the expense of larger estimation times. However, curved flight-path estimates have to rely on the entire LOS rate pattern rather than on local field estimates.