A Data-Driven Autonomous Driving System for Overtaking Bicyclists
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 research addresses the safety and perceptual challenges of automated car-to-bicycle overtaking, a maneuver associated with high collision risk for vulnerable road users. While Advanced Driver Assistance Systems (ADAS) are prevalent, they are typically designed from a driver-centric perspective, often ignoring bicyclist safety perceptions. The study aimed to develop data-driven models for initiating automated overtaking and evaluate these models from both driver and bicyclist viewpoints to identify discrepancies in satisfaction and perceived risk. The research comprised two studies. Study I utilized naturalistic driving data from the Safety Pilot Model Deployment (SPMD) dataset, involving 102 instrumented vehicles. Researchers extracted 740 car-to-bicycle overtaking events and developed four logistic regression models to predict overtaking initiation based on gap distance, relative velocity, vehicle velocity, and lane position. These models achieved Area Under the Curve (AUC) scores above 0.8, indicating strong classification performance. Study II implemented these models in a CarSim driving simulator to conduct a subjective assessment. Thirty-two participants (16 drivers, 16 bicyclists) evaluated simulated overtaking scenarios. The experimental design manipulated vehicle speed (25 mph vs. 40 mph), lateral offset (50th vs. 75th percentile distances), lane type (shared vs. dedicated bike lane), and the presence of oncoming traffic. Participants rated their satisfaction and perceived risk of collision after each scenario. The findings revealed significant differences in how drivers and bicyclists perceive safety. Both groups reported lower satisfaction and higher perceived risk when overtaking occurred at higher speeds or in the presence of oncoming traffic. However, the negative impact on bicyclists was mitigated by the presence of a dedicated bicycle lane. Crucially, bicyclists sought greater lateral room during overtaking, whereas drivers were satisfied with the current lateral offsets derived from naturalistic data and did not perceive significant risk from the smaller distances. This indicates a mismatch between driver behavior and bicyclist comfort levels. The study concludes that automated overtaking systems designed solely on driver behavior may not ensure bicyclist safety or satisfaction. The results highlight the necessity of adjusting automated overtaking parameters to account for bicyclist perceptions, particularly regarding lateral distance and speed. Stakeholders, including automated feature developers and policymakers, must address the inconsistency between driver and bicyclist perspectives to create safer shared road environments. The research provides a framework for integrating human factors from both road user groups into the design of autonomous driving functions.
Key finding
Higher overtaking speeds and oncoming traffic reduced satisfaction and increased perceived risk for both drivers and bicyclists, but dedicated bicycle lanes mitigated these negative effects for bicyclists, who also preferred greater lateral clearance than drivers typically provided.
Methodology
simulator
Sample size: 32
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, crash risk outcomes
- Theoretical Contribution: computational model