In automated vehicles, passengers will focus on non-driving tasks, causing a mismatch between expected and sensed motion that leads to motion sickness. Driving simulators show clear visual motion but reduced or missing physical motion, which also causes sickness. Both cases highlight the need for improved motion control strategies. Additionally, because motion sickness susceptibility varies amongst individuals, countermeasures must address this variability at a personal level to be effective. This presentation focuses on a recently proposed motion sickness modelling framework that integrates a group-averaged sensory conflict model such as subjective vertical conflict model with an individualized accumulation model based on Oman (1990) to capture differences in susceptibility. A Gaussian mixture model of individualized motion sickness parameters across a population was created using various datasets. This probabilistic model enables accurate prediction of motion sickness responses in unseen datasets. This model framework has been incorporated in a novel Model Predictive Control based Motion Cueing Algorithm (MCA) to control the motion sickness in driving simulators. The sensory conflict term within the cost function helps in optimising the simulator motion to recreate realistic motion perception while reducing motion sickness. Simulated evaluations followed by human-in-the-loop experiments, showed that our novel MCA achieved atleast a 56% reduction in motion sickness relative to benchmarks without statistically significant degradation of perceived realism. These results emphasize the importance of coherent motion cues at reduced magnitudes to balance perceived realism and comfort. This research offers a substantial advancement in the field of driving simulator design and individualized modelling of motion sickness.
A Mon, study studied this question.