Abstract
INTRODUCTION: With the increasing prevalence of electric vehicles, motion sickness has emerged as a critical factor impairing passenger comfort. Current studies relying on simulated driving face limitations in replicating real-road conditions. METHODS: We conducted real-vehicle experiments across six roadway scenarios: one-way left turn (R1), linear acceleration/deceleration (R2), sudden arrest-activation (R3), uphill S-curve (R4), downhill S-curve (R5), and one-way right turn (R6). A synchronized system (BioRadio + vehicle gyroscopes) captured subjective ratings from participants (n = 10) and objective data. RESULTS: Significant changes occurred in mean values of GSRmean , HRmean , RMSSD, and RESPmean during motion sickness (p < 0.05), while standard deviations ( GSRSD , RESPSD ) showed no significance. Motion sickness severity ranked as: R4 (8.4) > R5 (7.7) > R3 (6.3) > R2 (4.4) > R1 (2.0) > R6 (1.4), confirming S-curves as the primary trigger. DISCUSSION: The logistic regression model achieved 81.25% accuracy in predicting motion sickness states. This study provides empirical evidence for optimizing vehicle motion control and road design to enhance passenger comfort.