Incorporation of Physiological Features in Drowsiness Detection Using Deep Neural Network Approach

Abstract

The rapid growth in the number of vehicles on the road has exacerbated traffic congestion and the likelihood of more road accidents. Implementing a smart accident prevention system in the subsequent years will be necessary since the number of fatalities increases exponentially. Drowsiness is a feeling that occurs just before falling asleep or the desire to sleep that is very strong for an unusually long period. Therefore, it is indispensable to assess the physical and psychological factors that may impact a driver’s reflexes, resulting in decreased reaction times. One of the primary causes of vehicle accidents is driver fatigue and weariness. When operating a vehicle, driving a car, one must be focused and attentive and careful. This paper proposes a drowsiness detection method that integrates machine learning and physiological approaches such as heart rate and blood oxygen level. We have presented an efficient system to deal with real-time driver drowsiness detection using Convolutional Neural Network and other human biological features, including the blood oxygen level and cardiac rate.