Development of an AI-powered AR glasses system for real-time first aid guidance in emergency situations

Abstract

AI-powered augmented reality (AR) systems provide real-time, hands-free guidance, enabling untrained individuals to respond effectively in emergencies. By combining AI decision-making with AR visuals, they enhance awareness, reduce stress, and help bridge the gap before professional help arrives, especially in critical or underserved settings. This paper introduces an AI-powered augmented reality (AIAR) platform designed to assist untrained bystanders in delivering effective emergency response. By combining real-time object detection (YOLOv5), edge computing (Jetson Nano), and multimodal guidance through AR overlays and text-to-speech, the system targets four critical scenarios—bleeding, burns, fainting, and CPR—using specialized detection models built on domain-specific datasets. The decision-making and instructional flows align with medically validated protocols to ensure accuracy and relevance. Testing demonstrated the system’s high usability and responsiveness in delivering hands-free, context-aware guidance, though limitations related to hardware, model performance, and real-world conditions remain. Future developments include more compact AI hardware, wearable AR enhancements, adaptive interaction, clinical trials, and the integration of advanced AI, such as generative models and natural language interaction.