What we don’t (yet) know about self-driving carsickness
conference paper
Of all passengers, 2/3 does suffer from carsickness to some extent, while drivers suffer considerably less and comprise about 2/3 of all car occupants. Whilst this renders carsickness a minority concern today, in a future of automated vehicles in which all occupants are passengers, the majority of occupants will suffer, thereby making carsickness a game changer regarding the way vehicle motion should be controlled optimally. This paradigm hence requires knowledge about carsickness, explicated by means of numerical models that generate valid predictions for the syndrome as a whole. Although the ISO 2631-1:1997 is still the most widely acknowledged numerical model to predict motion sickness, in this paper we discuss a number of limitations why ISO may not be the appropriate model for predicting carsickness in particular. As a consequence, we define several outstanding questions that should be answered for an optimal prediction of carsickness. These questions concern temporal aspects such as accumulation, habituation, recovery and retention, the effects of angular motion, predictability of motion and visual effects. The latter not only concerns carsickness, but simulator sickness in particular. In addition, we address susceptibility to sickness, i.e., individual and demographic effects, as well as methodological issues regarding the quantification of sickness that currently hamper the progress of our understanding of carsickness. To offer passengers of automated vehicles comfortable rides time after time, answers to these outstanding questions should be known not only regarding vehicle motion control, but regarding routing and driving simulation as well.
TNO Identifier
1024323
Source title
Proceedings of the Driving Simulation Conference, DSC 2022 Europe VR, Driving Simulation Association, 15-16 September, Strasbourg, France
Pages
37-42