Systematic review of distinct modalities and artificial intelligence for distracted driving detection

Distracted driving remains a major contributor to road traffic crashes worldwide despite advances in driver monitoring technologies. Recent developments in machine learning (ML) and deep learning (DL) have enabled increasingly sophisticated detection systems, yet research has been dominated by visual-only approaches that may not fully capture the multidimensional nature of driver distraction. This systematic review synthesizes evidence from 77 peer-reviewed studies published between 2019 and 2024 that applied ML and DL techniques to distracted driving detection across four modality categories: visual, sensor-based, multimodal, and emerging approaches. Visual approaches, particularly convolutional neural networks, temporal architectures, ensemble frameworks, and attention-based models, dominate the literature and frequently report accuracies exceeding 90% under controlled benchmark conditions, yet evidence of cross-dataset robustness and real-world generalizability remains limited. Sensor-based approaches provide complementary insight into vehicle dynamics and physiological states, while emerging paradigms such as radio frequency sensing, neuromorphic computing, and auditory sensing offer privacy-preserving alternatives to camera-based systems. Multimodal approaches often outperform unimodal baselines by integrating complementary visual, physiological, and vehicular information, improving context awareness and detection capability. Across all modalities, a recurring challenge is the reliance on controlled datasets and limited evaluation under real-world conditions. The review identifies a fundamental gap between benchmark performance and deployment readiness and highlights the need for standardized evaluation protocols, cross-dataset validation, multimodal fusion strategies, and lightweight deployable architectures. Addressing these challenges is essential for developing reliable detection systems capable of supporting advanced driver assistance systems and improving road safety outcomes.