Design and Implementation of Smart Alert System for Reducing Road Traffic Accidents in Myanmar

The source of most road traffic accidents is related to the drivers' behavior. Therefore, it is urgently needed to identify unsafe driving and to take actions to change this situation. A new method to detect driver behavior is naturalistic driving. In this method, drivers' behavior is traced in daily trajectories by analyzing their moving recorded GPS data through mobile phones. In this paper, a new method for warning and alert drivers about their aggressive behavior is proposed to reduce the risk of accidents due to the drivers in Myanmar. It is also expected to improve drivers' attentiveness about their driving behavior as well as to inform the responsible organizations such as police stations, other transportation companies and insurance companies about their driver's style of driving. There are four stages in this proposed method. They are data collection, trajectory cleaning using DBScan clustering method, classification of motorized and non-motorized vehicles using J48 algorithm from Weka tool, external interference and aggressiveness evaluation and alerting. The main contribution to this research work is the development of smart alert system for driver and his organization.