Machine Learning Techniques in the Detection of Drowsiness in Drivers

Abstract

Drowsiness is one of the major risk factors that contributes to numerous accidents. With a variety of strategies and accurate categorization procedures that identify different drowsy stages and alert the individual at particular periods, this study seeks to lessen its effects. To improve the detection effectiveness of the current methods, Fractional Fourier Transform (FrFT) has been employed for feature extraction, ABC (Artificial Bee Colony) for optimisation, and NN and sparse classifiers for classification. The two solutions exhibit high efficiency when it comes to improving the system's accuracy and other performance measures. The two strategies are also contrasted, with the latter producing better results.