Insights for the Future of Car Rental and Ridesharing: Driving Behavior Across Different Levels of Automation

Meda, Pranav; Contreras, Aubrey Victoria; Lo, Wei-Hsiang; Huang, Gaojian; Luo, Yue · 2025 · ROSA P / Mineta Transportation Institute

archive: archived pipeline: cataloged verified

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

Summary

This study investigates how human driving behavior changes across different levels of vehicle automation, aiming to inform the future integration of autonomous vehicles into car rental and ridesharing industries. As autonomous technology matures, there is potential for a unified on-demand transportation model; however, understanding driver adaptation to varying automation levels is critical for safety and system design. The research specifically addresses two objectives: identifying behavioral differences among drivers operating vehicles at Level 0 (manual), Level 3 (conditional automation), and Level 5 (full automation), and exploring how these behaviors vary between "risky" and "conservative" assistance feature styles during lane-keeping and lane-changing tasks. The researchers conducted human-subject experiments with twelve participants (aged 21–29, holding valid driver’s licenses) using a medium-fidelity driving simulator (MiniSim). The experimental design included six driving trials, with two trials conducted at each automation level. To capture comprehensive behavioral data, the study utilized wearable instrumentation, including a motion capture system (Movella/Xsens) to monitor body posture and joint angles, and eye-tracking glasses (Pupil Core) to record gaze patterns and pupil diameter. Participants performed driving tasks involving lane keeping and lane changing under both risky (late intervention) and conservative (early intervention) assistance modes. Data collection also included pre- and post-study questionnaires assessing driving experience, trust, and acceptance of autonomous systems, as well as NASA-TLX questionnaires to measure cognitive load after each trial. The findings revealed distinct behavioral patterns associated with each automation level. In terms of driving performance, drivers exhibited stable speed and steering control during Level 0 and Level 5 operations. However, at Level 3, participants showed a noticeable decrease in speed and a significant increase in steering variability, particularly following takeover requests. Regarding body posture, drivers maintained a tense posture during manual driving (Level 0), whereas Level 3 required specific postural preparations to facilitate takeover actions. Eye movement analysis indicated that Level 0 driving involved active scanning and continuous visual control, while Levels 3 and 5 were characterized by notable shifts in attention and gaze coordination. These results suggest that the transition to conditional automation introduces specific biomechanical and cognitive challenges that differ from both manual and fully autonomous driving. The study concludes that the unique behavioral shifts observed at Level 3 highlight the need for improved vehicle design and driver training to manage the complexities of conditional automation. The authors recommend implementing multimodal interfaces and alarm systems, enhancing vehicle ergonomics to support posture transitions, and developing training programs to increase driver awareness. These measures are intended to mitigate the short-term behavioral changes identified in the study, thereby improving safety and user experience in future autonomous rental and ridesharing fleets. The research underscores the importance of understanding human factors in mixed-automation environments to ensure the successful deployment of autonomous vehicle technologies.

Key finding

Drivers exhibited stable speed and steering control at Levels 0 and 5, but experienced decreased speed and increased steering variability during Level 3 conditional automation.

Methodology

simulator

Sample size: 12

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. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

StageOutcomeToolModelPromptAttemptsCompleted
discover success rosap 2 2026-05-23
archive success 1 2026-05-23
extract success cached 2 2026-06-10
clean success 1 2026-06-01
chunk success 1 2026-06-01
embed success 1 2026-06-02
enrich success 1 2026-05-23
promote success 1 2026-05-23
summarize success llm qwen3.6-27b-prismaquant summ-v5 3 2026-06-10
tag success vector_similarity 19 2026-06-11
verify success 2 2026-06-10

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

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

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).