AI-driven Intelligent Transport System For Enhancing Road Safety In Bangladesh

Bangladesh is an accident-prone country in the South Asian region. Every day, several lives are lost due to road accidents. This research studies the patterns of accidents that occurred between January and June 2024 in the country and proposes an Intelligent Transportation System (ITS) component designed to reduce road accidents by alerting drivers, law enforcement agencies, transport authorities, as well as fire and medical services. For this research, road accident data were collected from a daily newspaper named "Prothom Alo". It was observed that more than 3,278 people died in these six months. Among these incidents, 73.57% were caused by over-speeding, while 26.43% resulted from illegal lane changes. So, this research introduces an AI-based accident prediction and alarm system that collects real-time videos from the traffic monitoring cameras that are an integral part of intelligent transportation systems (ITS). In this research, YOLO V8 is employed as a core AI tool for identifying vehicle types moving on the road. The proposed algorithm estimates the speed of the vehicles and the direction of running as well as extracts the number plate information. The algorithm was tested by the video collected from the national highway-1 (N1), i.e., the Dhaka-Chittagong highway, and found 95% accuracy of number plate detection and more than 85% of vehicle type detection.