Physiological Assessment of Driver Trust in Automated Vehicles under Distinct Driving Styles

Li, Yizheng; Hu, Zhilin; Proctor, Karl; Owens, Andrew; Dorn, Lisa; Zhao, Yifan · 2026 · Crossref

DOI: 10.54941/ahfe1007863

archive: archived pipeline: cataloged

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the challenge of assessing driver trust in automated vehicles (AVs) by moving beyond subjective self-report measures, which suffer from low temporal resolution and bias. The research investigates how distinct AV driving styles—specifically aggressive and hesitant behaviors—affect driver trust through objective physiological signals. The motivation stems from the need to understand the dynamic, real-time evolution of trust, as appropriate trust calibration is critical for the safe deployment of Level 3 automated driving systems. The researchers conducted a driving simulation experiment with 13 participants using a high-fidelity simulator built on the BeamNG.tech platform. The experimental design featured three distinct driving scenarios implemented on the same route: a Baseline (smooth, normative driving), an Aggressive condition (characterized by late braking, high-speed cornering, and curb contact), and a Hesitant condition (characterized by prolonged stopping, slow creeping, and premature braking). Participants experienced the Baseline first, followed by the other two in random order. Data collection included synchronized electroencephalography (EEG) with 24 electrodes, eye-tracking at 120Hz, and subjective trust ratings using Jian’s trust scale and the ROSAS scale. EEG data were preprocessed using band-pass filtering (1–40 Hz) and Independent Component Analysis, then analyzed using short-time Fourier transform for time-frequency analysis. Results indicated that aggressive and hesitant driving styles elicited distinguishable subjective, neural, and attentional responses. Subjectively, the Aggressive scenario significantly reduced overall trust and perceived system competence compared to the Baseline, whereas the Hesitant scenario did not produce a statistically significant decrease in trust. However, the Aggressive condition induced substantially higher subjective discomfort. Eye-tracking analysis revealed that aggressive driving led to a contraction of gaze dispersion, evidenced by reduced saccade amplitude and a lower probability of large saccades, suggesting heightened supervisory monitoring. Conversely, hesitant driving was associated with a marked increase in prolonged fixations (>400 ms) and total fixation time, potentially reflecting uncertainty in the system’s decision-making consistency. EEG time-frequency analysis showed that both non-baseline conditions resulted in power reductions in frontal regions across Delta, Theta, Alpha, and Beta bands. The Aggressive condition exhibited broader spectral modulation, including significant power reductions in parietal, occipital, and temporal regions, while the Hesitant condition showed more concentrated frontal suppression, particularly in the left hemisphere. The study concludes that different forms of distrust are manifested through differentiated gaze dynamics and neural spectral patterns. The findings provide preliminary evidence supporting the feasibility of physiology-based approaches for assessing driver trust in automated driving. By demonstrating cross-modal consistency between behavioral attention, neural activity, and subjective trust, the research lays a scalable foundation for developing trust-aware automated driving systems that better align with human cognitive states. Future work will focus on extending the framework with larger sample sizes and joint feature analysis to quantitatively establish functional relationships between neural activity, visual attention, and perceived trust.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.8-27b-gittensor on 2026-08-23; verification: pending re-verification.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• trust calibration
• trust in automation foundations
• automation
• automation surprise
• acceptance adoption
• situational awareness

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: physiological data, self report data
• Methodological Resource: tool software