A simulation study of driving performance in long tunnel

The purpose of this paper is to understand the problems of cross section design and of methods of providing traffic information by analyzing driving performance and drivers' stated preference in a driving simulation of the very long tunnel (total length approximately 27 km) that is envisaged as one of the future projects under Uchiura Bay in Hokkaido. Twenty subjects participated in the study by using the driving simulator at the Civil Engineering Research Institute (The CERI driving simulator). Trials totaled 160. There was no clear difference in the average driving speed and the average lateral position for different cross section types, partly because visual recognition of the lane markings was good. However, when the right shoulder was narrow (width=0.75 m), more than 70% of the subjects reported that the median wall caused them to experience a negative psychological state. Therefore, to reduce this psychological stress, it is desirable to provide a wider right shoulder in the tunnel. The recognition of longitudinal distance of the subjects while driving through the tunnel seems rougher than for an ordinal tunnel. Thus, it is necessary to place large-sized distance signs at every 5 or 10 km, when one places distance indicators in the tunnel. There is also an increasing need to provide distance information through ITS.