Detecting driver distraction

Liang, Yulan · 2018 · Crossref

DOI: 10.17077/etd.20x5g8oi

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This dissertation addresses the critical safety challenge of detecting driver distraction in real time to enable adaptive distraction mitigation systems. The increasing prevalence of in-vehicle information systems (IVISs) introduces visual ("eye-off-road") and cognitive ("mind-off-road") distractions, which significantly contribute to motor vehicle crashes. While cooperative automation can adapt vehicle functions to driver state, existing research lacked accurate algorithms to detect visual, cognitive, or combined distractions, particularly regarding their interactions. The study aimed to bridge these gaps by developing quantitative detection methods using data mining and naturalistic driving data. The research comprised three primary objectives. First, it refined the detection of cognitive distraction by developing a layered algorithm combining supervised clustering and Dynamic Bayesian Networks (DBNs). Second, it established real-time estimation algorithms for visual distraction using data from the 100-Car Naturalistic Driving Study, linking eye-glance patterns to crash risk. Third, it investigated the interaction between visual and cognitive distractions through a controlled driving experiment, developing a sequential strategy to identify combined distraction states. The results demonstrated that the layered algorithm significantly improved the detection of cognitive distraction compared to previous methods. Analysis of naturalistic data revealed a strong relationship between estimated visual distraction—quantified by cumulative glance duration, history, and eccentricity—and increased crash risk. The controlled experiment found that visual distraction dominated performance impairments, while cognitive distraction reduced the overall impairment effects when combined with visual distraction. Consequently, the study proposed a sequential detection strategy: if visual distraction is detected, cognitive distraction detection is unnecessary, as visual distraction is the primary driver of performance degradation. The significance of this work lies in providing a robust framework for real-time distraction detection using performance indicators and data mining techniques. The findings suggest that visual distraction is the primary factor to monitor for safety mitigation, simplifying the requirements for adaptive systems. These methods are generalizable to other performance impairments, such as driver fatigue, offering a pathway for developing cooperative automation systems that enhance driving safety by dynamically adapting to driver states.

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-prismaquant on 2026-06-10; verification: verified.

Topics

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

• visual
• distraction detection algorithms
• temporal
• cognitive
• external distraction
• visual manual

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: measurement protocol
• Theoretical Contribution: conceptual framework