Does the testing environment matter? Carsickness across on-road, test-track, and driving simulator conditions

A new study from TU Delft researchers Georgios Papaioannou and Barys Shyrokau reveals critical limitations in using driving simulators to test motion sickness in autonomous vehicles. Published as arXiv:2602.22671, the research directly compares carsickness across on-road, test-track, and simulator environments—the first study to provide multi-environment validation. The findings show simulators produce significantly lower motion sickness scores, creating a validity gap that could impact the development of comfort-focused autonomous driving systems. With the rise of automated vehicles, accurate sickness prediction has become crucial for passenger acceptance and safety.

The research involved 28 participants performing eyes-off-road tasks while reporting sickness using the Misery Scale and Motion Sickness Assessment Questionnaire. The key technical finding is that simulators have a limited working envelope that prevents accurate reproduction of low-frequency motions below 0.5 Hz—precisely the frequencies most provocative for motion sickness. This creates a systematic underestimation problem for automotive researchers and developers who rely on simulators for testing. The study provides both subjective ratings and objective acceleration measurements, offering a framework for more standardized testing across environments as the industry moves toward validating autonomous vehicle comfort.