Does driving behaviour in the simulator compare to driving behaviour in the real road? The answer to this question provides insight in the validity of the simulator with respect to driving on narrow lanes. In the current experiment driving behaviour in the simulator is compared to driving behaviour (in an instrumented vehicle) on the road. A group of 19 subjects drove on the road on the “plusstrook” of the A27 south of Houten in an instrumented vehicle. Another group drove the same route in the simulator. Apart from drives on the far left lane, the “plusstrook” (2.60 m wide), drives were completed on the directly adjacent middle lane and a standard reference lane (3.50 m wide). In absolute sense the results show that driving in the simulator, in case there is any difference, is in general more riskful in the lateral dimension compared to the real world. Especially small TLCs were found on both left and right side. Line crossings are bit more difficult to interpret. On the right side they do not exist, neither in the simulator, nor in the real world. On the left side there seems to be an outlyer, in the sense that there are on the (very small) left lane a couple of line crossings. Although these can be called very few (2.5 s in about 5 minutes driving time, which is about 0.8 % of the time), they are still significantly different from 0. The simulator shows for most of the behavioural measures at least relative validity. This means that the pattern of effects of the three lanes is reproduced in the simulator. Where this is not the case, or where deviations are large in an absolute way, this can be explained by the fact that not all on-the-road circumstances were reconstructable in the simulator. This accounts especially for the behaviour of the surrounding traffic. (Author/publisher)
Samenvatting