Driver Drowsiness and Fatigue Detection: An Investigation of Techniques and Challenges

Abstract

Fatigue driving stands as a critical contributing factor to road traffic accidents. To enhance safety, fatigue driving detection technology has evolved into a prominent research focus. This paper first introduces the concept of fatigue driving, then classifies detection methods into four categories based on different input sources. It provides a comprehensive review of studies on methods rooted in physiological features, facial features, vehicle behavior features, and multi-feature fusion—exploring their underlying principles and effectiveness, while revealing the technical bottlenecks of each method in practical applications. The paper aims to offer a theoretical reference for the optimization and innovation of this technology, furnish solid theoretical and data support for method refinement and technological innovation in subsequent research, and help readers in this field gain insights into future development directions. Ultimately, by summarizing the strengths and limitations of current detection approaches, this review highlights emerging trends such as deep learning-based multimodal fusion and real-time embedded detection, pointing toward more accurate, robust, and user-friendly fatigue monitoring systems in the future.