A Drowsiness Detection Decision Support System using Self-Organising Map

Machine learning is currently having its large share of attention in terms of application in several fields and unsupervised learning techniques are at the forefront of this trend. This article harnesses the classification properties of Self-Organizing Maps (SOMs) to accurately classify acquired signals from car drivers in order to support decision-making mechanism in drowsiness detection system. The drowsiness detection SOM facilitates easier and more rapid representation of the input vector space which helps in understanding the intertwined relationship between different inputs as well as categorize inputs in terms of relevance and significance to overall drowsiness detection process.