Takeover scenario semantic salience: a cognitive framework for linking scenario difficulty to driver performance and neurophysiological state.

Evaluating driver takeover performance (TOP) in conditionally automated vehicles remains an important safety challenge. In this study, we examine Takeover Scenario Semantic Salience (TSSS) as a provisional conceptual framework for describing takeover scenario difficulty in terms of two candidate dimensions: execution clarity and information availability. To examine this framework under controlled conditions, we conducted a driving simulator study with 41 participants under four simplified scenarios designed to instantiate contrasts intended to reflect these dimensions. We also developed the Comprehensive Takeover Performance Index (CTOPI) to summarize takeover-related vehicle-control performance using objective indicators. Eye-tracking and EEG measures were further used to explore associated neurophysiological states. The results suggest that scenarios characterized by reduced execution clarity or reduced information availability were associated with poorer takeover performance under attentive-driver conditions, although the pattern differed across scenario pairs and rounds. The physiological results showed different patterns across the two scenario contrasts, while both scenarios designated as lower TSSS were accompanied by larger pupil diameter; these patterns may reflect differences in cognitive demand. Overall, the findings provide preliminary evidence that TSSS-aligned scenario contrasts may offer a useful descriptive lens for takeover difficulty in controlled settings, while broader validation under more realistic Level 3 takeover conditions remains necessary.