The new TRL driving simulator is based on a medium sized car which is surrounded by screens on which are projected computer generated images. The images are produced by Silicon Graphics computers which use Multigen databases of the road system being modelled. The main objective of this study was to validate the use of the simulator to model events which could distinguish between measures of driving performance for subjects who were impaired through different dose levels of alcohol. A secondary objective was to investigate the relationship between `real' road driving tasks and similar simulated tasks. A trial was conducted with a placebo dose and a low and high dose of alcohol using a cross-over design where subjects were their own controls. Baseline measures were taken on some of the tasks. Subjects were required to take simulator and non-simulator tasks which included a real driving task on the TRL small loop. The results indicate that different levels of alcohol can be detected on both simulator tasks and non-simulator tasks, the high alcohol dose of 80 mg/litre resulted in the greater deterioration of driving performance. The relationship between the `real' driving task on the TRL test track correlates positively with the results from the similar simulated task thus demonstrating that the simulator is indicative of what happens in `real' driving. (A)
Abstract