🙌Hi! I’m Xinyi Yuan, currently a Ph.D. candidate in Harada Lab, Osaka University, supervised by Prof. Weiwei Wan. Previously, I got my MPhil degree at the Intelligent Transportation (INTR) Thrust in the Hong Kong University of Science and Technology (Guangzhou campus), supervised by Prof. Meixin Zhu. I earned my Bachelor’s in Robotics Engineering from Southern University of Science and Technology, exploring human-robot interaction strategies for upper-limb rehabilitation in Brain-Robot Rehabilitation Technology Lab. From Feb. 2024 to Sep. 2024, I remotely engaged in fascinating legged robot research at the Shanghai AI laboratory supervised by Dr. Chenjia Bai.

✨ News

• 2025.05:  📝 We’ve released a new paper on Conservative Soft Actor-Critic algorithm.
• 2025.04:  📌 I started my Ph.D. in Harada Lab at Osaka University.
• 2025.03:  📝 We’ve released a new paper revisiting inverse kinematics with a refined numerical solver.
• 2024.10:  📌 I joined Harada Lab at Osaka University as a research student.
• 2024.10:  📝 Please view our new work on diffusion-based locomotion policy on project website.
• 2024.06:  📅 My MPhil graduate examination will be arranged on 2nd, July.
• 2024.04:  🎉 Our paper is accepted by the IEEE Transactions on Intelligent Vehicles.
• 2024.03:  🎉 Our paper is accepted by the 35th IEEE Intelligent Vehicles Symposiumk (IV 2024).

📝Publications

* Co-first author, ✉️ Corresponding author.

• IKSel: Selecting Good Seed Joint Values for Fast Numerical Inverse Kinematics Iterations
  Xinyi Yuan, Weiwei Wan✉️, Kensuke Harada.
  Preprint. (Submitted to IEEE Transactions on Automation Science and Engineering)
• Preference Aligned Diffusion Planner for Quadrupedal Locomotion Control
  Xinyi Yuan*, Zhiwei Shang*, Zifan Wang, Chenkai Wang, Zhao Shan, Meixin Zhu✉️, Chenjia Bai✉️, Xuelong Li.
  arXiv preprint arXiv:2410.13586
• Effective Reinforcement Learning Control using Conservative Soft Actor-Critic.
  Xinyi Yuan*, Zhiwei Shang*, Wenjun Huang, Meixin Zhu✉️, Chenjia Bai, Di Chen.
  Preprint. (Submitted to IEEE Transactions on Automation Science and Engineering)
• AggFollower: Aggressiveness Informed Car-Following Modeling.
  Xianda Chen, Xinyi Yuan, Meixin Zhu✉️, Xinhu Zheng, Shaojie Shen, Xuesong Wang, Yinhai Wang, Fei-Yue Wang.
  IEEE Transactions on Intelligent Vehicles (IF=14, JCR Q1)
• Learning Realistic and Reactive Traffic Agents.
  Meixin Zhu✉️, Di Chen, Xinyi Yuan, Zhiwei Shang, Chenxi Liu.
  35th IEEE Intelligent Vehicles Symposium (IV 2024)

👩‍🎓 Educations

• Sep. 2022 - Jul. 2024
  MPhil(Master of Philosophy), Intelligent Transportation Thrust, System Hub
  Hong Kong University of Science and Technology (Guangzhou campus).
  
• Sep. 2018 - Jun. 2022
  Bachelor’s Degree, Robotics Engineering, Department of Mechanical and Energy Engineering
  Southern University of Science and Technology.
  

👩‍💻 Research Experience

• Harada lab (Research Student)
  Oct. 2024 - Apr. 2025
  Supervised by Prof. Weiwei Wan
  

• Shang Hai AI laboratory (Remote Research)
  Feb. 2024 - present
  Supervised by Dr. Chenjia Bai
  Motivation: Improve the performance of the diffusion policy on high-frequency locomotion policy on legged robots.
  Main approaches: Propose the two-stage learning framework that combines the offline diffusion planner with online preference alignment.
  

• Drive AI laboratory, HKUST (GZ)
  Dec. 2022 - present
  Supervised by Prof. Meixin Zhu
  Motivation: Incorporate driving aggressiveness into data-driven models, and realize safe and human-like driving behavior modeling.
  Main approaches: Embed aggressiveness tokens into Physics-Informed LSTM architecture, test on the Lyft and HighD datasets.
  

• Brain-Robot Rehabilitation Technology Lab, SUSTech
  Sep. 2020 - Jun. 2022
  Supervised by Prof. Mingming Zhang
  Motivation: Utilize online metabolic cost estimation to build a human-in-the-loop rehabilitation strategy.
  Main approaches: Use multimodal physiological signals to estimate metabolism by multiple linear regression, and design interaction strategies.
  

• Shenzhen Subway WebGIS Qualitative and Quantitative Hybrid Research
  Sep. 2021 - Aug. 2022
  Supervised by Prof. Lili Wang
  Field research, literature review, and geographic information visualization on Shenzhen’s cultural landscape development.
  

👔 Internships

• Gongzhiyi Technology Co., Ltd., Shenzhen
  Jun.2022 - Aug. 2022
  User research, user education, and product design for the Series A+ funded blueprint management software “Evercraft.”
  
• Guoxin Hongsheng Equity Investment Fund Management Co., Ltd., Shenzhen
  Mar. 2021 - Jun. 2021
  Research intern, Private Equity Investment Department (IV)