The Behavioral and Geographical Features Based Machine-Learning Method for Alerting the Drivers

Lack of attention due to drowsiness is a leading cause of vehicle accidents. Drowsiness is a significant risk factor in the breakdown of many people’s lives. The most excellent approach to prevent drowsy driving accidents is to alert the drivers in advance. As a result, there is a requirement for a real-time system that can identify drowsiness early and accurately. In this paper, we proposed the behavioral and geographical features-based machine learning method (BGF-MLM) to identify sleepiness at the earlier stage to detect the driver’s drowsiness. A sizable realistic dataset is incorporated throughout three states of consciousness: attentive, decreased vigilance, and drowsiness. The video utilized fall into two extreme categories: awake and drowsy. Computer vision and deep learning methods retrieve behavioral and geographical features. Behavioral features are extracted using an artificial neural network (ANN) and fed into a long short-term memory (LSTM) network. The geographical features are extracted using an LSTM network. The suggested method details the processes involved in developing a drowsiness detection system to estimate a driver’s tiredness level and issue an alarm before the driver poses a significant risk to traffic safety. If the driver is warned in time, they can escape several possible disasters.