Designing the Non-linearity of Human Motion Perception for Dynamic Driving Simulators
Résumé
Driving simulators allow the exploration of research fields that are difficult to reach in normal conditions, like testing driving assistance systems during dangerous situations. However, there are some perceptive aspects of the simulations that still need to be improved. This is due to their mechanical limitations that do not allow us to produce accelerations or braking at their physical value. As a consequence, driving simulators' motion algorithm is using tilt in combination with linear translations to produce a sensation of linear acceleration. The present study show that motion perception depends on the manner tilt and translation are used together to provide a unified percept of linear acceleration. Our results show that there is a large variability on how humans perceive accelerations and decelerations and it is not advisable to use motion algorithms without taking into account non-linearity of human perception.