Identifying Occurrences of Abnormal and Drunk Driving Using Smartphones

Drunk driving is the act of operating a motor vehicle while under the influence of alcohol or other drugs that can impair a person's ability to drive. It is a serious offense that can lead to deadly consequences. Alcohol impairs judgment, coordination and reaction time. When someone is driving under the influence, they are unable to make quick decisions or react to sudden changes in the environment. This puts themselves and other drivers at risk. Drunk driving is a major cause of fatal car crashes, and it is illegal in every state. In initial phase of research, major focus have been given for reviewing existing methods. Overall drunk driving detection problem can be solved by three methods i.e, 1) Dynamic Time Warping, 2) Threshold-based, and 3) Machine Learning methods. Most works in the related architecture follows one or more of these methods. The proposed methodology in the study uses threshold-based and machine learning based methods to detect the drunk driving incidents in details. Based on extensive studies on drunk driving detection, research gaps and problem definition of this research work are highlighted. In the future, a system to overcome the limitations presented in the base paper will be developed.