The effect of visual properties of the simulated environment on simulator sickness and driver behaviour

This study aimed to explore the idea of visual complexity in the TRL driving simulator in terms of how it might be quantified and systematically reduced and the effect this would have on simulator sickness levels and driver performance. Three novel simulator environments were designed to explore several visual factors: the presence of textures; the realism of textures; and the presence of vertical relative size cues. These environments were compared to a standard realistic simulation of an urban street. Twenty female participants drove each of the four simulations and completed a set of standardized tasks including natural speed choice, distance judgement and speed maintenance with and without use of the speedometer. Simulator sickness scores were measured before and after each condition. Results indicated no benefit of reduced complexity in terms of simulator sickness. Participants tended to choose a higher speed in all of the reduced complexity environments compared to the standard realistic simulation. When judging speeds without the speedometer participants tended to over estimate their speed, travelling slower than required when attempting to maintain 20 mph but when attempting to maintain 50 mph they tended to under estimate their speed and travel faster than the required 50 mph in all conditions. Overall, the results of this study suggest that a reduced complexity abstract environment might be suitable for some research applications, particularly where the performance of abstract tasks is of interest rather than real-world behaviour and where development resource is an issue.