Visual demand and the introduction of advanced driver information systems into road vehicles
This thesis contains six studies investigating the impact of advanced in-vehicle information systems on the visual demands of the driver. The experiments, while self-contained were conceived to relate together in a cohesive manner. The first study investigated the reliability of visual behaviour assessment. Video tape records from experimental trials were analysed post-hoc. Significant test/retest correlations were obtained. Experiment two considered the visual demands of the driving task without intervention from new technologies. Results from road trials using an instrumented vehicle suggested changes in the subject's visual scanning which could be related to the roadway environment (i.e., rural, urban and motorway driving). In experiment three the effects of the introduction of a driver information system were assessed using a congestion warning device on public roads. System use resulted in significantly greater: subjective mental workload, glance duration and frequency, and percentage time (eyes) away from the forward view; than the in-car entertainment system, or the control (normal driving). Experiment four replicated experiment three in a fixed base driving simulator. It aimed to establish the value of the simulator for the assessment of driver visual demand. The same significant differences presented in the road trial were observed in the simulation study. In the penultimate study, opportunities for the reduction of driver visual demand were investigated. The subjects were presented with: visual, auditory or visual and auditory route guidance information. Results suggest use of auditory information to supplement visual displays significantly reduces visual demand on the driver. The final study considered the effect of information availability on the distribution of visual scanning. Driver control of in-vehicle information presentation enabled self-determination of visual scanning strategies. Information system control of information presentation was found to disrupt the driver's visual checking. The interface design was shown to force the driver to adopt different visual scanning strategies. The contribution of the experimental work to the assessment of driver visual demand is discussed and the relationships between the experiments explored.