It is well known that sleepiness is an important contributor to crashes. The drivers are influenced by circadian low, hours slept or being awake. There is also reason to believe that variations in interior road noise may be a contributing factor. Up to now the road noise models in simulators are not very mature and not validated. The aim of the technical part of the SleepNoise project was to develop the road noise model for the ViP simulation infrastructure, including individual vehicle types driving on individual road types. The goal was a sound model, close enough to the original real world setting as to be approved by vehicle noise experts. The resulting modelled interior sound spectrum should not deviate more than a few decibels from the corresponding recorded sound. The resulting sound model proved to work very well for the intended use, facilitating detailed investigations of the influence of sound on driver behaviour in addition to providing realism and presence in the simulated environment. It allows changing the vehicle cabin sound model independently of the influence of road texture. Thus, more combinations of vehicle cabins and road textures can easily be investigated if needed. The realism of the road noise model was validated by NVH experts from Volvo Car Corporation well acquainted with the modelled vehicles. A driving simulator experiment was carried out and found to support the hypothesis that high levels of low-frequency road noise in the car contribute to increased driver sleepiness and impaired performance during night-time. The unexpected carry-over effect suggests that the onset of conditions such as those found in realistic road noise has a lasting effect, which cannot be neutralized by a short break while still seated in the vehicle. The carry-over effect might have been neutralized by allowing participants to exit the simulator, e.g. to stretch their legs and have a drink of water or so, which might also imply that countermeasures against driver fatigue need to have a proper impact to counter such lasting effects. These issues require further research. (Author/publisher)
Abstract