Characterizing driver speeding behavior when using partial-automation in real-world driving

Haus, Samantha H.; Gershon, Pnina; Mehler, Bruce; Reimer, Bryan · 2022 · Crossref

DOI: 10.1080/15389588.2022.2089664

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 how partial-automation systems influence driver speeding behavior in real-world conditions. Speeding is a significant contributor to traffic fatalities and injuries, yet the emergence of SAE Level 1 and Level 2 automation features, such as Adaptive Cruise Control (ACC) and lane-centering systems, may alter the prevalence and characteristics of this risky behavior. The research aims to identify and describe distinct types of speeding behaviors when drivers use partial-automation compared to manual driving, providing insights necessary for developing effective safety countermeasures. The analysis utilized data from the MIT Advanced Vehicle Technology Naturalistic Driving Study, involving 15 drivers who operated Volvo S90 vehicles equipped with ACC and Pilot Assist (PA) for one month. The dataset comprised 3,413 speeding epochs defined as traveling at least 5 mph over the speed limit for a minimum of 3 seconds on motorways during free-flow conditions. Researchers employed Dynamic Time Warping to characterize speed-exceedance profiles and multivariate modeling to evaluate associations between speeding duration, magnitude, and variability. Finally, they used Gower dissimilarity measures and Partitioning Around Medoids clustering to classify speeding behaviors into distinct groups. The results identified four distinct speeding behaviors in both manual and partially-automated driving: Incidental (short duration, low magnitude), Moderate (short duration, moderate magnitude), Elevated (moderate duration, high magnitude), and Extended (long duration, high magnitude). While the types of behaviors remained consistent across driving modes, their characteristics differed significantly. Incidental and Moderate speeding durations were significantly longer when using partial-automation compared to manual driving. Conversely, Elevated speeding was more prevalent and associated with higher speed magnitudes during manual driving. Although Extended speeding was more common during automation use, it exhibited lower mean and maximum speed magnitudes compared to Extended speeding in manual driving. Overall, speeding with partial-automation tended to have longer durations but lower magnitudes than manual speeding. The findings indicate that partial-automation does not eliminate or introduce new speeding behaviors but rather modifies their characteristics, particularly by extending duration while reducing speed variability and magnitude. This suggests that automation may moderate risk during long-duration speeding events but could pose safety concerns due to increased exposure time and potential driver distraction. The study concludes that safety systems designed to mitigate speeding must account for these divergent behaviors, implementing strategies specific to both manual and automated driving states to address the unique risks associated with each.

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.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

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

• speed choice
• automation surprise
• situational awareness

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: observational prevalence, behavioral performance data
• Methodological Resource: dataset resource