The effects of vibration on visual performance and display legibility

The research programme described in this thesis examines the theoretical and practical aspects of visual degradation caused by vibration. Initially, the published literature was reviewed in two areas: the effects of vibration on visual performance and the influence of display parameters on legibility under normal and vibration-degraded viewing conditions. A series of eight experiments were then conducted to assess visual performance and display legibility with reading and threshold detection tasks presented on cathode ray tube displays. The first experiment compared the visual degradation caused by sinusoidal and narrow-band random vibration in the frequency range 2.5 to 31.5 Hz. Random vibration produced a significantly smaller reading performance decrement than sinusoidal vibration. This finding was attributed to the greater probability of low image velocity with the random vibration stimuli and, in a follow-up experiment, velocity-based predictions accounted for the visual performance decrements due to laboratory-simulated aircraft vibration. A comparison of the effects of whole-body, display and simultaneous whole-body and display vibration showed that at frequencies below 4 Hz, vibration of the display resulted in the largest performance decrement and simultaneous vibration of observer and display, the least. In a second series of experiments, display parameters (horizontal and vertical character separation, contrast, font and symbol definition) were manipulated during static and vibration-degraded viewing conditions. Vertical separation was found to be a more significant variable than horizontal separation and tentative guidelines for character separation were stated. The relationship between contrast and legibility was shown to be quadratic during static and vibration-degraded viewing conditions. Optimum legibility occurred in the contrast range 60 to 88% . The determination of font and symbol definition requirements showed that for characters of equal subtense, 7x9 dot-matrix characters were significantly more legible than 5x7 characters during whole-body vibration. Differences between two frequently recommended fonts, the Huddleston and Lincoln Mitre, were not apparent at 7x9 definition although at 5x7 definition the Huddleston font was responsible for significantly fewer reading errors. In the final experiment, the spatial filtering characteristics of vibration were examined. In addition to an alphanumeric reading task, contrast thresholds for sinusoidally modulated grating patterns were determined during whole-body vibration. It was shown that the effects of vibration could be conceived as low-pass spatial filtering and the implications of this finding for simulating the effects of vibration were discussed. The overall findings of the experimental programme have been summarised and areas for future research discussed