Characterizing Heterogeneous Car-Following Behaviors of Human Drivers in Mixed Traffic
DOI: 10.1109/ichms59971.2024.10555751
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This study investigates the heterogeneous car-following behaviors of human-driven vehicles (HVs) in mixed traffic environments containing autonomous vehicles (AVs). While previous research indicated that HVs generally drive more smoothly when following AVs, existing analyses often failed to account for the heterogeneity of individual driving styles or relied on small sample sizes. To address this gap, the authors utilized the Lyft Level-5 dataset, which contains over 42,000 car-following events, to classify driving styles and quantify their impact on traffic safety and stability. The researchers processed trajectory data from 20 AVs operating in Palo Alto, California, identifying 19,754 HV-following-AV and 22,840 HV-following-HV events. They extracted metrics including speed volatility, time headway, and reciprocal time-to-collision (reTTC) to characterize behavior. Agglomerative hierarchical clustering was employed to categorize drivers into four distinct styles: HiVel-LSens-Agg (high-speed, low-sensitivity, aggressive), LoVel-Con (low-speed, conservative), HiVel-HSens-Agg (high-speed, high-sensitivity, aggressive), and LoVel-Agg (low-speed, aggressive). Multinomial logistic regression and multiple linear regression models were then used to analyze how lead vehicle type (AV vs. HV) influenced these styles and subsequent safety outcomes. The results revealed that drivers exhibited different car-following styles depending on whether they were following an AV or an HV. Specifically, drivers were less likely to adopt aggressive or conservative styles when following AVs compared to HVs. Regarding safety outcomes, the study found that leading AVs generally resulted in larger reTTC values (indicating higher collision risk) compared to leading HVs, regardless of the following driver’s style. This was attributed to the less predictable speed profiles of AVs, which reduced driver responsiveness. However, speed volatility varied by context: at low speeds, following AVs led to lower speed volatility (more stable traffic), whereas at high speeds, drivers insensitive to lead vehicle changes exhibited higher speed volatility when following AVs. Additionally, aggressive driving styles were associated with greater danger at low speeds, while style differences were less pronounced at high speeds. The findings highlight the complexity of human-AV interactions, demonstrating that while AVs may induce smoother driving in some contexts, they also introduce specific safety risks due to unpredictability. The study concludes that future AV control algorithms must account for the heterogeneous styles of following human drivers to optimize traffic stability and improve safety in mixed traffic scenarios.
Key finding
Human drivers exhibited four distinct car-following styles in mixed traffic, and while autonomous vehicles generally led to higher collision risk indicators than human-driven vehicles, they resulted in lower speed volatility at low speeds, indicating more stable traffic conditions.
Methodology
dataset
Sample size: 42594
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.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | — | — | — | 1 | 2026-05-28 |
| archive | success | canonical_url | — | — | 1 | 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 | success | semantic_scholar | — | — | 4 | 2026-06-15 |
| 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.
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