Early identification and detection of driver drowsiness using machine learning

Abstract

Driver drowsiness is a major cause of road accidents, leading to severe injuries and fatalities. Early detection of drowsiness can significantly enhance road safety by preventing potential accidents. This study presents a hybrid machine learning approach to identify and detect driver drowsiness in real time. The proposed system integrates both physiological and behavioural features, including eye closure duration, yawning frequency, and head position, combined with vehicle-based parameters such as steering pattern and lane deviation. A hybrid model, combining convolutional neural networks and long short-term memory networks, is employed to process real-time video streams and sensor data, enhancing the accuracy of drowsiness detection. The model is trained on diverse datasets to ensure robustness across different driving conditions. The system provides timely alerts to drivers when signs of drowsiness are detected, improving response time and reducing accident risks. Experimental results demonstrate high accuracy and low false detection rates, outperforming traditional machine learning models. This research highlights the effectiveness of hybrid machine learning techniques in enhancing driver safety and provides a foundation for further advancements in intelligent transportation systems. The proposed solution can be integrated into modern vehicles to promote safer driving environments and reduce drowsiness-related accidents.