IOT and AI-Based Automatic Crash Detection Systems for Emergency Services: A Review

Abstract

Road traffic collisions are a serious public safety issue that continues to cause high rates of death, injury and economic losses around the world. Reporting systems for accidents are frequently inefficient and slow, especially in low-traffic regions and remote areas, which can delay the response to an accident. With the advent of the Internet of Things (IoT), Artificial Intelligence (AI), computer vision and intelligent transportation systems, automatic crash detection systems have been developed that have ability to monitor in real-time and report the crash quickly. A comprehensive review of the various IoT and AI-based crash detection technologies is provided including the sensing hardware, communication infrastructure, cloud and fog computing, computer vision and machine learning techniques. It reviews the latest studies on using accelerometers, GPS, surveillance cameras, CNN and YOLO-based models to detect accidents, and offers a comparative assessment of their techniques, advantages, obstacles, and suggestions. There are several research gaps identified in the review, such as limited accident severity assessment, limited multimodal sensor fusion, and scalability issues; interoperability problems, privacy concerns, and real-world validation problems in different traffic environments. Based on these findings, future research directions for the paper are discussed, focusing on Explainable AI, Real-time inference at the edge, Secure IoT communication and Scalable intelligent transportation frameworks. The study illustrates how the synergy between AI technologies and IoT-enabled sensing and communication capabilities can significantly improve the capabilities of accident detection, reduce emergency response times and contribute to overall road safety.