Nudging human drivers via implicit communication by automated vehicles: Empirical evidence and computational cognitive modeling
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 automated vehicles (AVs) can proactively influence human driver behavior through implicit communication, specifically in scenarios where the AV holds the right of way. While existing research often focuses on AVs yielding to humans, this work addresses the gap in understanding how AVs might "nudge" human decisions during unprotected left turns. The authors hypothesized that subtle longitudinal maneuvers, such as brief deceleration or acceleration, could alter a human driver’s probability of accepting a traffic gap and their decision timing. To test this, the researchers conducted a driving simulator experiment with 19 participants. The study employed a 2 × 5 within-subject design, manipulating the initial time-to-arrival (TTA) of an oncoming AV (4.5 s or 5.5 s) and its acceleration profile. The five acceleration conditions included a constant speed baseline, a "deceleration nudge" (brief deceleration followed by acceleration), an "acceleration nudge" (brief acceleration followed by deceleration), and long-duration acceleration or deceleration maneuvers. Participants were tasked with deciding whether to turn left before or after the AV passed. The study measured decision outcomes (Go or Stay), response times, and subjective negative ratings of the AV’s behavior. Additionally, the authors fitted multiple variants of a drift-diffusion model, a cognitive framework for evidence accumulation, to the empirical data to understand the underlying decision-making process. The results demonstrated that human drivers were sensitive to deceleration nudges but not acceleration nudges. Specifically, the probability of a "Go" decision significantly increased when the AV performed a deceleration nudge compared to the constant speed baseline, whereas acceleration nudges had no significant effect. Long deceleration also increased the likelihood of a Go decision, while long acceleration decreased it. Regarding response times, Go decisions were generally faster than Stay decisions. Notably, long deceleration induced significantly longer response times for Go decisions, particularly when the initial TTA was larger. Subjectively, participants rated long deceleration maneuvers as significantly more negative and confusing than other conditions. The best-fitting cognitive model suggested that humans integrate noisy dynamic information regarding time-to-arrival and distance to a fixed decision boundary, with an initial bias toward the Go decision. The significance of this work lies in its dual contribution to empirical understanding and computational modeling. It provides evidence that AVs can effectively influence human behavior through subtle longitudinal maneuvers, specifically deceleration, without requiring explicit communication. Furthermore, the validated drift-diffusion model offers a parsimonious framework that not only explains observed human behavior but can also predict responses to arbitrary AV maneuvers. This model can be integrated into interaction-aware controllers for AVs, enabling them to plan motions that proactively manage interactions with human drivers, thereby improving safety and traffic flow in mixed-autonomy environments.
Key finding
Human drivers were more likely to proceed with an unprotected left turn when an oncoming automated vehicle performed a brief deceleration nudge, whereas acceleration nudges had no significant effect on their decision-making.
Methodology
simulator
Sample size: 19
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 author_sweep_intake on 2026-05-27.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | author_sweep | — | — | 2 | 2026-05-27 |
| archive | success | canonical_url | — | — | 5 | 2026-06-06 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-07 |
| chunk | success | chunk | — | — | 1 | 2026-06-07 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-07 |
| enrich | skipped | — | — | — | 4 | 2026-07-02 |
| promote | success | — | — | — | 1 | 2026-06-04 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 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.
- situational awareness
- automation surprise
- driver vru interaction
- ehmi external hmi
- mental model of traffic
- acceptance adoption
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
- Theoretical Contribution: computational model, theory or model