Detection of Abnormal Driving Behavior Detection Using ADBDConvolutional Neural Networks

Abstract

Monitoring anomalous driving behaviours in real time is a critical component of increasing vehicle safety. To improve driver behaviour and driving practisesin order to avoid car accidents. The use of vision-based anomalous driving behaviour detection is growing in popularity because it is fundamental to the safety of drivers and passengers in cars and is a crucial step toward attaining automated driving at this level at this time. This difficult detection task can be greatly aided by recent advancements in deep learning approaches, such as advanced deep learning models' remarkable generalisation power and the large volumes of video clips required for completely training these data-driven deep learning models. To wrap off the research work, novel deep learning-based models, inspired by the newly developed and widely used fully connected convolutional network named the Abnormal Driving Behavior Detection (ADBD Net), are presented.