A Review Paper on Drunk Driving Detection System using IOT & ML Techniques

Nowadays, there is a significant increase in drunken-driving accidents from the past few years. Drunk driving has emerged as a significant problem in recent times. With the help of different technological implementations, many preventive measures are taken to date, such as alcohol detection sensors, detection using speech, detection using driving patterns, detection using IOT & Machine Learning techniques etc. But each system has its limitations, such as usability, complexity, scalability, burdensome implementation. This paper describes the various approaches to detect drunken driving with the advantages and disadvantages of each. Out of these approaches, the article focuses on the detailed literature review to discuss existing machine learning algorithms and IOT methodologies to improve the accuracy of the system which can be used in detecting drunken driving. At last, the comparative review is done for all these approaches and we found that Random forest classifier with two stage model is most efficient, having high accuracy.