Monitoring fatigue and drowsiness in motor vehicle occupants using electrocardiogram and heart rate − A systematic review

Background Fatigue is a complex state resulting in decreased alertness, often accompanied by drowsiness. Driving fatigue has significantly contributed to traffic accidents globally, highlighting the need for effective monitoring techniques. Various technologies exist to enhance driver safety and minimize accident risks, such as fatigue detection systems that alert drivers as drowsiness sets in. In particular, measuring heart rate (HR) patterns may offer valuable insights into the occupant’s physiological condition and level of alertness and may allow them to understand their fatigue levels. This review aims to establish the current state of the art of monitoring strategies for vehicle occupants, specifically focusing on fatigue assessed by HR and heart rate variability (HRV). Methods We performed a systematic literature search in the databases of Web of Science, Scopus, and PubMed, using the terms vehicle, driver, physiologic monitoring, fatigue, sleep, electrocardiogram, heart rate, and heart rate variability. We examined articles published between the 1st of January 2018 and the 31st of January 2023. Results A total of 294 papers were identified from which 71 articles were included in this study. Among the included papers, 57 utilized electrocardiogram (ECG) as the acquired signal for HR measures, with most ECG readings obtained through contact sensors (n = 41), followed by non-intrusive wearable sensors (n = 11). The majority of the works rely on subjective self-reported fatigue ratings (n = 27) and video-based observer ratings (n = 11). Regarding validation, 16 papers do not report validation of any kind. From the included papers, only 13 comprise a fatigue and drowsiness estimation system, 9 with reported validation. Some report acceptable performances, but reduced sample size limits conclusions. Conclusions This review highlights the potential of HR analysis and non-intrusive instrumentation for continuously monitoring driver’s status and detecting sleepiness. One major issue is the lack of sufficient validation and estimation methods for fatigue, contributing to insufficient methods in providing proactive alarm systems. This area shows great promise but is still far from being reliably implemented.