A Fuzzy-Logic Approach to Dynamic Bayesian Severity Level Classification of Driver Distraction Using Image Recognition
DOI: 10.1109/access.2020.2994811
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This paper addresses the critical need for accurate detection and classification of driver distraction severity to enhance Advanced Driver Assistance Systems (ADAS) and prevent road accidents. The authors argue that existing research often focuses on single distraction activities, failing to account for the compounded risk of multi-class distractions (e.g., simultaneous visual, cognitive, and physical distractions). To mitigate this, the study proposes a novel methodology that classifies distraction into three severity levels: safe, careless, and dangerous. This classification aims to trigger control transitions in semi-autonomous vehicles when high-severity distraction is detected. The study utilizes the Naturalistic Driving American University of Cairo (AUC) Distraction Dataset, comprising 14,478 video frames from 44 participants across ten distraction classes, including safe driving, phone usage, texting, and talking to passengers. The authors developed a Discrete Dynamic Bayesian (DDB) model combined with a Mamdani-based fuzzy logic system. The model processes four key inputs: hand position, face orientation, current driver activity, and previous driver distraction history. The fuzzy logic system assigns weights to these inputs based on established risk metrics, prioritizing texting as the most dangerous activity, followed by phone usage and passenger conversation. The DDB component calculates distraction severity by analyzing the likelihood of continuous distraction over sequential frames, incorporating prior evidence from ground-truth labeled data. The results demonstrate that the proposed system effectively detects multi-class distractions and classifies them into appropriate severity levels. The model highlights that driver distraction severity is dynamic, with instances showing rapid transitions from careless to dangerous driving when multiple distraction factors converge. The fuzzy logic approach successfully models the uncertainty inherent in human driving behavior, allowing for a more nuanced assessment than binary detection systems. The study confirms that combining visual cues (face orientation), physical cues (hand position), and cognitive load (activity type) provides a robust framework for severity classification. The significance of this work lies in its contribution to the development of safer ADAS and semi-autonomous vehicle technologies. By providing a method to quantify distraction severity in real-time, the system enables proactive safety interventions, such as handing control to the vehicle during high-risk scenarios. The approach addresses the limitations of previous studies by considering the cumulative impact of simultaneous distractions, thereby offering a more comprehensive tool for accident prevention. The findings support the integration of fuzzy logic and dynamic Bayesian networks in driver monitoring systems to handle the unpredictability of human behavior.
Key finding
The proposed fuzzy-logic and dynamic Bayesian model successfully classifies driver distraction severity into safe, careless, or dangerous levels, demonstrating that multi-class distractions can rapidly transition from careless to dangerous driving states.
Methodology
dataset
Sample size: 44
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 | 2026-05-07 |
| archive | success | canonical_url | — | — | 12 | 2026-06-06 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-04 |
| chunk | success | chunk | — | — | 1 | 2026-06-04 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-04 |
| enrich | success | — | — | — | 1 | 2026-05-07 |
| promote | success | — | — | — | 1 | 2026-05-07 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | partial | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified_with_issues.
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).
- Methodological Resource: tool software
- Theoretical Contribution: conceptual framework, theory or model