Preventing Crashes in Mixed Traffic With Automated and Human-Driven Vehicles
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 the safety risks associated with mixed traffic environments, specifically focusing on car-following scenarios where human drivers follow autonomous vehicles (AVs). While AVs are projected to reduce crashes by eliminating human error, recent reports indicate that mismatches in braking and acceleration decisions between AVs and human drivers can lead to rear-end near-crashes. The research aims to determine if human drivers exhibit different braking behaviors when following an AV leader compared to a human-modeled leader, thereby identifying potential collision risks in mixed traffic streams. The researchers conducted a driving simulator experiment with 48 participants (aged 18–30) divided into two groups. The simulator featured a straight, two-lane road with eight stop-controlled intersections spaced 500 meters apart, with a speed limit of 30 mph. In the first scenario (AV-HUMAN), 24 participants followed an AV leader programmed with four distinct constant deceleration profiles (1, 2.25, 2.75, and 3.25 m/s²). In the second scenario (HUMAN-HUMAN), a separate group of 24 participants followed a lead vehicle whose speed profiles were modeled after four experienced human drivers. Data collected included vehicle speeds, acceleration/deceleration rates, and clearance distances. The results revealed a significant behavioral mismatch in the AV-HUMAN scenario. Statistical analysis showed a significant difference (p=0.04) between the overall deceleration rates of the human participants and the AV leader. Participants tended to brake later and more abruptly than the AV, particularly when the AV decelerated at higher rates (2.25–3.25 m/s²), whereas they matched the AV’s gradual braking only when the AV decelerated slowly (1 m/s²). In contrast, the HUMAN-HUMAN scenario showed no significant difference in deceleration rates between participants and the human-modeled leader (p=0.85). Participants maintained larger average clearances (46.06 m) behind the human-modeled leader compared to the AV leader (24.64 m), suggesting greater comfort and predictability when following human-like driving patterns. The study concludes that the rigid, constant deceleration patterns typical of current AV algorithms create a mismatch with human driving expectations, increasing the risk of rear-end collisions in mixed traffic. The findings imply that AV braking algorithms should be adjusted to mimic human-like, variable deceleration profiles to improve safety and public acceptance. By aligning AV behavior with human norms, the technology can better integrate into existing traffic streams without causing disruptive interactions.
Key finding
Human drivers exhibited significantly different deceleration rates and braking speeds when following an AV leader compared to a human-modeled leader, with no significant difference observed in the latter scenario.
Methodology
simulator
Sample size: 48
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).
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| 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).
- Empirical Findings: behavioral performance data
- Methodological Resource: dataset resource
- Theoretical Contribution: computational model