Vehicle Motion Sickness Profiling Through Interpretable Gradient Boosting: A Framework for Advanced Driver Assistance Systems

Motion sickness poses a significant challenge in vehicle dynamics, impacting passenger comfort and usability. This study presents a robust framework for understanding motion sickness and its origins through advanced modeling and interpretability techniques. The proposed approach leverages both classical and modern algorithms, with CatBoost emerging as the top-performing model after optimization. To address the interpretability of these models, SHAP methods are employed to extract actionable insights, revealing key factors contributing to motion sickness. CatBoost also demonstrates superior interpretability by providing balanced and meaningful feature importance distributions. The combination of advanced performance and interpretability allowed the model to accurately detect diverse motion sickness-inducing scenarios and uncover critical risk patterns.