Driver Behavior Classification: A Systematic Literature Review
DOI: 10.1109/access.2023.3243865
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 presents a systematic literature review (SLR) addressing the critical problem of road safety, where human behavior is identified as the primary factor in traffic accidents. Motivated by the high global mortality rate from traffic incidents and the limitations of existing Advanced Driver Assistance Systems (ADAS) in predicting danger, the study aims to synthesize existing research on automatic driver behavior classification. The authors seek to provide a comprehensive overview of classification types, data sources, features, preprocessing techniques, and artificial intelligence algorithms used in the field, thereby offering a guide for future research and practical implementation. The methodology follows a structured SLR approach, analyzing academic papers published between 2015 and 2022 retrieved from five digital databases, including ScienceDirect. The review is organized around six research questions covering behavior objectives, study types, data sources, preprocessing, feature selection, and model performance. The authors categorize driver behavior into five distinct types: abnormal behavior (unsafe/risky), aggressive driving, line deviation, vehicle stopping behavior, and driver status (distraction, drowsiness, etc.). The analysis also considers the impact of contextual factors such as environmental conditions, vehicle variables, and driver characteristics. The findings reveal a wide variety of machine learning (ML) and deep learning (DL) techniques applied to classify these behaviors. For abnormal and risky driving, algorithms such as Random Forest, Support Vector Machines (SVM), and Long Short-Term Memory (LSTM) networks achieved accuracies ranging from 70% to over 99%, depending on the dataset and specific task. Aggressive driving detection utilized CNNs, Gated Recurrent Units (GRU), and ensemble methods, with some models reaching 95% accuracy. Line deviation and stopping behaviors were effectively classified using SVM, Neural Networks, and Logistic Regression, often leveraging data from smartphone sensors, GPS, accelerometers, and video surveillance. Driver status detection, particularly for distraction and drowsiness, showed high performance using CNNs and RF algorithms, with some studies reporting accuracies above 99%. The review highlights that data sources vary significantly, including simulator data, real-world sensor data, and public datasets like NGSIM and SHRP 2. The significance of this work lies in its provision of a new taxonomy for driver behavior classification and a detailed analysis of the state-of-the-art algorithms and datasets. By identifying key contributions and challenges, such as the need for robust feature extraction and the handling of unbalanced data, the paper offers valuable insights for researchers and practitioners. It underscores the potential of AI-driven systems to enhance road safety by accurately detecting and classifying dangerous driving behaviors, thereby facilitating timely corrective measures and improving the overall efficiency of Intelligent Transportation Systems.
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 scout_discovery on 2026-05-08.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | partial | scout | — | — | 2 | 2026-05-08 |
| archive | success | canonical_url | — | — | 16 | 2026-06-09 |
| extract | success | cached | — | — | 2 | 2026-06-09 |
| 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 | semantic_scholar | — | — | 2 | 2026-06-04 |
| promote | success | — | — | — | 1 | 2026-06-04 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-09 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 1 | 2026-06-09 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-09; 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: observational prevalence
- Methodological Resource: dataset resource
- Theoretical Contribution: computational model