I am Yuchen Fan (樊宇琛), a Robotics Engineering undergraduate at Beijing Institute of Technology. I work across algorithms, software, and hardware to turn robotics research into reliable systems, with an emphasis on real-robot deployment and reproducible results.
Research interests: embodied AI, autonomous systems, 3D scene intelligence (3D Gaussian Splatting, SLAM), and robot learning.
🔥 News
- 2026.09: My paper When Does KV Cache Compression Pay? (KVLedger) is accepted as a regular paper at GPC 2026.
- 2026.07: SinD 2.0, a multi-city UAV dataset for SOTIF-oriented safety validation at signalized intersections, is released on arXiv.
📝 Publications
Yuchen Fan
International Conference on Green, Pervasive, and Cloud Computing (GPC), 2026, Springer LNCS, to appear · Project · Reviews (Chinese)
- KVLedger accounts for codec, metadata, transfer, and consumer costs to decide, per link and per model, whether KV cache compression pays off in disaggregated LLM serving.
Yunwei Li, Shengjie Fu, Chunrong Chen, Chengxiang Zhao, Yuchen Fan, Mingyu Zhu, Yanchao Xu, Jiahui Xu, Anran Wang, Huanan Wang, Yuxin Zhang, Lan Yang, Chuzhao Li, Jie Ji, Yi He, Abhijit Sarkar, Akash Sonth, Hong Wang, Jun Li
arXiv preprint, 2026 · arXiv · PDF · Dataset
- My contribution: data annotation and the 3DGS-based visual simulation extension (Sec. VI-G).
🔬 Research
- 3D Scene Intelligence. 3D representations for navigation, visual localization, and editable scene generation.
📖 Education
- 2023 – 2027, Undergraduate in Robotics Engineering, School of Automation, Beijing Institute of Technology.