Development of an Intelligent Emergency Vehicle Priority Passage System Based on YOLOv11-Tiny and Edge Computing
Year: 2026
Type: conference-paper
Abstract: The problem of poor emergency vehicle (EV) rescue efficiency is caused by urbanization. This paper proposes a smart priority passage system that integrates NVIDIA Jetson Orin NX and YOLOv11-tiny object detection algorithms. This paper utilizes transfer learning to optimize the model, which significantly improves the mean accuracy (mAP) from 0.58 to 0.92, effectively suppressing training oscillations and overfitting. The system imports the Region of Interest (ROI) determination logic into the simulated intersection and dynamically adjusts traffic signal priorities in real-time. The verification results confirm that the architecture has high reliability and can significantly shorten rescue passage time, providing a practically valuable solution for intelligent transportation systems (ITS).
Institution: National Taipei University of Technology
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7204134672
