Effect of Visual Threshold on Aircraft Control

Abstract

The resolution achieved when displaying such quantities as bank and pitch angle in a flight simulator is rarely as good as the corresponding resolution in flight. A flight simulator experiment was conducted to determine the influence on the closed loop control of an aircraft of poor resolution in bank and heading. Variations in the visual threshold of these quantities were made by varying the size of a dead-space to the pilot's display. The results show clearly that the control accuracy is degraded. Low-frequency limit cycles occur, of a character similar to those which can be observed in ground based simulators and, to a lesser extent, in flight. From these results it is concluded that the control situation is aggravated not only by threshold effects purposely inserted during the simulation, but also by the thresholds inherent in the pilot's own sensors. Thus in certain control situations, the inability of the pilot to detect attitude or position errors visually will result in control difficulties. The application of these results to the landing approach and flare is discussed. Nomenclature g = acceleration due to gravity, fps h = height, ft I/a = lift parameter I/sec p = rate of roll, deg/sec q = rate of pitch, deg/sec r = rate of yaw, deg/sec s = Laplace operator V = aircraft forward speed, fps y = track position, ft d a = aileron angle, deg be = elevator angle, deg = bank angle, deg \f/ = heading, deg con = longitudinal s.p. frequency, rad/sec £n = longitudinal s.p. damping ratio. Problem In tr odu ction M ANY advances have been made in flight simulation techniques in the past decade. These include devices for both visual and kinaesthetic cue simulation: they apply equally well to research and training simulators. Nevertheless, a gap exists between simulator and flight which pilots often find difficult to span. In consequence, the control technique which a pilot uses in a flight simulator will, under certain circumstances, differ from that which he uses in flight. Extrapolation from simulator to flight is then of severely limited value. This difficulty can be circumvented in the training simulator by 'adjustments' to the simulator until the general concensus of pilot opinion is that the simulator behaves like the aircraft. Unfortunately, this technique is not valid for research simulation. However, the research engineer is allowed to be selective in the pilots he uses. Most investigations are therefore assessed by experienced pilots, familiar with the limitations of the simulation, and able to contribute to the interpretation of the results. Much remains to be done in understanding the inability of the simulator to reproduce certain characteristics of the Presented as Paper 70-357 at the AIAA Visual and Motion