Driver distraction accounts for 20% of motor vehicle crashes (Treat et al., 1977). The increasing prevalence of in-vehicle information systems (IVIS) may exacerbate these effects. In order to design IVIS such that they do not evoke driver distraction, it is paramount to identify whether and to what extent the modality in which IVIS information is presented influences driving performance. This required the quantification of the specific resources used while driving, which few studies have documented. Therefore, the objectives of the current research were to quantify the effects of information modality on driver distraction, on tracking performance in a driving simulator. This, in turn, allowed for an indirect measure of resource allocation, and thus the identification of a preferable information modality for IVIS design. In addition to the primary tracking task in a driving simulator, twenty-four participants performed secondary comprehension tasks posing cognitive distraction. These were performed in three perceptual modalities (auditory, central visual, and peripheral visual), at two difficulties (low and high), over two trials each, which were then compared to pre and post control trials. The distinction of difficulties for each modality was as follows. Auditory low: information was presented aurally with no noise overlaid. Auditory high: information was presented with signal-to-noise difference of 10dB. Central visual low: information was presented at 0° with no manipulation of text presentation. Central visual low: character spacing increased to 4pt for high difficulty. Peripheral visual: information presented at 10° and 20° eccentricities for low and high difficulties, respectively. It was found that the tracking performances were dependant on the modality. This indicates that there is a significant effect (p<0.05, F= 71.4479 (2,22)) of modality on tracking performance. While all significantly different to each other, the auditory modality reflected the smallest decrement in performance with 81.12% of baseline, followed by the central visual condition with 68.47% of baseline, followed by the peripheral condition with 45.50% of baseline. These differences can be attributed to the superior time-sharing capabilities of the auditory-visual dichotomy, as predicted by multiple resource theory. The results therefore illustrate how the modality of IVIS is likely to differentially influence resource demands, which, to varying degrees, would lead to an increased crash risk. Finally, the added factor of difficulty elicited significant overall results (p<0.05, F= 38.9480 (1, 44)), which suggests that these results are applicable across a wide range of difficulties. Further, this indicates that even at low difficulties, the auditory modality is preferable in IVIS design. (Author/publisher)
Abstract