Aggressive Driving Detection and Alert System Leveraging Current Signal Light Infrastructure Brian Li 1 , Emma Li 1 and Hui Li # 1 Mission San Jose High School, Fremont, CA, USA # Advisor ABSTRACT According to the National Highway Traffic Safety Administration, aggressive driving behaviors are responsible for 54 percent of annual traffic accidents from 2007 to 2019. These accidents cost our society billions of dollars econom- ically. Additionally, hundreds of thousands of innocent lives are taken away and a profound psychological impact is placed on survivors. Existing research on aggressive driving detection mostly relies on devices mounted on vehicles for collecting the data when driving. Such approaches, however, will be hard to enforce in real life because of con- sumer privacy concerns. Also, not all existing vehicles can have the monitoring system, thus the overall effectiveness is limited. The approach that we propose here is to leverage the existing infrastructure of traffic lights to enforce the detection of aggressive driving behaviors for all vehicles at all times. The undertaken approach introduces centralized detection and alert systems that can work with an existing traffic light control system and a real digital map system as a real-world application. The detection system utilizes both traffic lights and speed sensors mounted alongside to detect aggressive driving behaviors at intersections. The alert system adopts an efficient specialized graph traversal algorithm to pinpoint impacted traffic lights and alert only the drivers nearby those lights. To demonstrate the ap- proach, we prototyped both a detection system and an alert system running on an Arduino board. We verified their consistent high accuracy and real-time response by applying them to an experimental setup that consists of streets and intersections with traffic lights and infrared sensors. 1. Introduction Aggressive driving has consistently been the leading cause for traffic accidents. The statistics published from the National Highway Traffic Safety Administration shows that over 54 percent of the annual traffic accidents are caused by aggressive driving from 2007 to 2019[1]. Not only do these accidents create a heavy toll onto society, with billions of dollars as an economic cost, but more profoundly they often lead to the loss of innocent people’s lives, and a psychological impact on those who manage to survive. Some interesting research efforts have been devoted to enhancing the techniques of detecting aggressive driving be- haviors. Haike Guan et. al. [2] applied a machine learning method to automatically detect certain aggressive driving behaviors based on the video recording collected from a surveillance camera mounted on taxis or trucks. Volume 10 Issue 4 (2021) ISSN: 2167-1907 www.JSR.org 1