Xiuli Yu

Beijing University of Posts and Telecommunications

Papers

Total Citations

H-Index

About

Xiuli Yu is pioneering the intersection of computer vision and robotic manipulation, with a sharp focus on high-precision manufacturing automation. Her research spans robotic grasp detection, object detection for assembly lines, and sub-millimeter-accurate visual servoing. Yu’s most cited work, “Robotic Grasp Detection Based on Transformer” (2022, 14 citations), introduced a novel transformer-based architecture that significantly improves grasp pose estimation, laying a foundation for more adaptive robotic handling. She further advanced industrial vision with “YOLOOD” (2023, 6 citations), an arbitrary-oriented detection method tailored for flexible flat cables in robotic assembly, addressing a notoriously difficult perception challenge. Her most recent contribution, “EA-CTFVS” (2024, 4 citations), tackles the long-standing peg-in-hole assembly problem—a critical bottleneck in manufacturing with strict tolerance demands. Unlike prior work validated only in simulation or under limited conditions, Yu’s environment-agnostic coarse-to-fine visual servoing method achieves sub-millimeter accuracy in real-world settings. With a growing citation footprint and a trajectory toward solving real industrial challenges, Xiuli Yu is establishing herself as a rising force in vision-based robotic automation.