How do distracted and normal driving differ : an analysis of the ACAS naturalistic driving data

Green, Paul E.; Wada, Takahiro; Oberholtzer, J.; Green, P. A.; Schweitzer, J.; Eoh, Hong · 2007 · ROSA P / University of Michigan. Transportation Research Institute

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 differences between normal and distracted driving to support the development of workload managers for in-vehicle systems. Motivated by the need to identify objective metrics for driver distraction, the researchers analyzed naturalistic driving data from the Advanced Collision Avoidance System (ACAS) Field Operational Test. The dataset comprised 136,792 miles of driving from 96 subjects, evenly distributed across men and women in their 20s, 40s, and 60s. Distraction was operationally defined as four successive video frames where the driver’s head was not oriented toward the forward scene. The methodology involved a two-pass coding process of randomly selected video clips to identify secondary tasks and specific activities. The researchers examined descriptive statistics for steering wheel angle, heading angle, throttle opening, and speed, analyzing how these measures varied by road type (limited access, major, minor), driver age, and distraction status. They also tested distribution fits for these measures and evaluated "throttle holds"—periods of minimal throttle change—as potential indicators of distraction. Logistic regression models were employed to determine which variables best discriminated between normal and distracted driving states. The results indicated that distraction had almost no effect on overall driving performance statistics, except for a 36% decrease in mean throttle opening and a 6% decrease in mean speed. No consistent differences were found in the distributions of steering, heading, or speed parameters. However, logistic regression identified specific factors that distinguished distracted driving: turn signal use and age for expressways; gender and lead vehicle range for major roads; and lane width, lane offset, and lead vehicle velocity for minor roads. Regarding throttle holds, a single universal definition performed poorly, often identifying normal driving as distracted. However, when throttle hold parameters (time window and threshold) were tailored to specific combinations of road type and driver age, they effectively identified distracted states. The significance of this work lies in its contribution to the design of adaptive interface technologies. The findings suggest that simple aggregate statistics are insufficient for detecting distraction, but context-specific models using tailored throttle hold definitions and other driving-related variables can successfully discriminate between normal and distracted states. This supports the feasibility of developing workload managers that assess real-time driving demand and suppress non-essential in-vehicle tasks to enhance safety.

Key finding

Tailoring throttle hold parameters to specific road types and driver age groups significantly improved the detection of distracted driving compared to using a single fixed definition.

Methodology

naturalistic

Sample size: 96

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).

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.

• manual
• visual
• temporal
• distraction detection algorithms
• steering pattern
• external distraction

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, observational prevalence
• Theoretical Contribution: conceptual framework