I-95 Florida Highway Dataset

Despite the considerable efforts to enhance road infrastructure and enforce stricter driving regulations to ensure road safety, the number of accidents worldwide remains alarmingly high, driven by factors such as distracted driving, speeding, and impaired driving. For instance, in the United States, fatal accidents increased by 16% from 2018 to 2022, with the number of fatalities rising from 36,835 in 2018 to 42,795 in 2022. This highlights the pressing need for innovative solutions to mitigate traffic incidents and enhance road safety. Machine learning (ML) can significantly improve traffic safety and reduce traffic- related fatalities. This paper presents a novel ML-based real-time highway traffic incident forecasting system named VehiCast. VehiCast utilizes Vehicular Ad-hoc Networks (VANETs) and Federated Learning (FL) to collect real-time traffic data, such as average delay, average speed, and the total number of vehicles across several highway zones, to enhance traffic incident prediction accuracy in real-time. Our comprehensive experimental results reveal that VehiCast achieves high prediction accuracy, ranging from 96.00% to 97.00%. VehiCast can also efficiently assess the current traffic conditions across various highway zones, enabling more accurate real-time estimations of traffic accidents, latency, and arrival times, ultimately promoting safer driving behaviors and reducing the incidence of traffic-related fatalities.