<![CDATA[Extended Reality (XR) technology represents a promising advancement for intelligent transportation systems by enhancing traffic monitoring, situational awareness, and operator interaction. Despite progress, conventional vehicle tracking systems continue to exhibit limitations, such as inaccurate trajectory estimation, restricted risk prediction capabilities, and non-intuitive visualization interfaces. This study introduces XR-VITS (Extended Reality Vehicle Intelligent Tracking System), an integrated framework that unifies YOLO-based vehicle detection, Kalman filter-based multi-object tracking, homography-based real-world coordinate mapping, trajectory prediction, collision risk assessment, and immersive XR visualization within a single traffic monitoring architecture. The framework was evaluated on diverse traffic datasets, including urban, highway, adverse weather, and low-light scenarios. Experimental results indicate that XR-VITS achieved a Multiple Object Tracking Accuracy (MOTA) of 89.3% and an Identification F1 Score (IDF1) of 86.2%, surpassing several state-of-the-art tracking methods, such as DeepSORT, StrongSORT, ByteTrack, and QDTrack. Additionally, the system attained a collision risk prediction F1-score of 86.4% while maintaining real-time processing at 25.1 FPS. The XR visualization module further enhanced operator situational awareness and reduced response times compared to conventional 2D monitoring systems. These results demonstrate that XR-VITS provides an effective and scalable solution for next-generation intelligent transportation systems requiring predictive intelligence, immersive visualization, and real-time traffic monitoring.]]>
