A Survey and Tutorial of EEG-Based Brain Monitoring for Driver State Analysis

Zhang, Ce; Eskandarian, Azim · 2021 · Crossref

DOI: 10.1109/jas.2020.1003450

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 effective Driver Monitoring Systems (DMS) to mitigate human error, which accounts for 94% of fatal crashes. While fully autonomous vehicles remain distant, DMS integrated into Advanced Driver Assistance Systems (ADAS) offer a near-term solution by monitoring driver cognitive and physiological states. The authors focus on Electroencephalography (EEG) as a primary sensing modality due to its high temporal resolution and direct measurement of brain activity, unlike camera-based systems that only detect physical behaviors. The paper serves as a comprehensive survey and tutorial, reviewing three decades of research on EEG-based driver state analysis, including system setups, signal processing, and classification algorithms. The study categorizes EEG-based research into three primary domains: driver distraction/inattention, fatigue/drowsiness, and brake intention. Experimental setups typically involve driving simulators or real-world vehicles, with data collection systems varying by electrode type (wet vs. dry) and connectivity (wired vs. wireless). Wet electrodes offer higher signal quality suitable for lab environments, while dry electrodes provide convenience for real-world applications. Connectivity-wise, wired systems are preferred for accuracy and stability, whereas wireless systems face challenges with noise and signal loss. The paper details the standard workflow for EEG analysis: signal preprocessing for artifact removal, feature extraction, and classification using machine learning. A significant portion of the review focuses on EEG signal preprocessing, specifically artifact removal techniques. The authors analyze four main methods: Independent Component Analysis (ICA), Canonical Correlation Analysis (CCA), Wavelet Transform (WT), and Regression Analysis. ICA is effective for separating ocular artifacts but suffers from high computational load and manual detection requirements. CCA offers lower computational costs and is suitable for real-time muscle artifact removal. WT provides time-frequency analysis but risks over-filtering useful EEG signals. Regression analysis is computationally efficient but requires reference channels. The paper highlights hybrid methods, such as combining ICA with WT or regression, to mitigate individual algorithm limitations, though these combinations increase computational complexity. The authors conclude that while current EEG-based algorithms show promise for safety applications, significant improvements are required in artifact reduction, real-time processing capabilities, and between-subject classification accuracy. The review underscores the evolution from simple time-frequency analysis to complex, data-driven machine learning models. Future development must address the trade-offs between signal fidelity, computational efficiency, and practical deployment in real-world driving conditions to effectively integrate EEG monitoring into commercial ADAS.

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_with_issues.

Topics

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

• drowsiness detection algorithms
• neuro workload indices
• drowsiness
• distraction detection algorithms
• gaze based attention detection

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: physiological data
• Methodological Resource: tool software, validation psychometrics