Pengcheng Wang

Do research trivial in hindsight, elegant in insight.

prof_pic.jpg

2505 Hearst Ave

Berkeley, CA 94720

I am a Ph.D. student in the Mechanical Systems Control Lab at UC Berkeley, advised by Prof. Masayoshi Tomizuka (Member of the National Academy of Engineering). Prior to that, I received dual bachelor’s degrees in Mechanics and Aerospace from Tsinghua University. During my undergraduate studies, I worked with Prof. Shengbo Eben Li.

I worked on Transferable and Scalable Reinforcement Learning for Robotics. Specifically, I care how RL agent can digest the scaled model size/ data/ training, and can be efficiently generalized across different embodiments/ tasks/ dynamics.

Outside of research, I love gaming with large-scale strategy systems such as Stellaris, Civilization VI, and Victoria 3. I’m also passionate about singing and served as one of the tenor leads in the Choir of Tsinghua University. Check out my favorite performance!

Email: wangpc [AT] berkeley.edu

Feel free to contact me for research collaborations or just to chat!


selected publications

  1. ICLR 2025
    residual-mppi.gif
    Residual-mppi: Online policy customization for continuous control
    Pengcheng Wang, Chenran Li, Catherine Weaver, and 4 more authors
    ICLR 2025
  2. Under Review
    DADP2.png
    DADP: Domain Adaptive Diffusion Policy
    Pengcheng Wang, Qinghang Liu, Haotian Lin, and 4 more authors
    Under Review
  3. ICRA 2025
    RPG.png
    Residual Policy Gradient: A Reward View of KL-regularized Objective
    Pengcheng Wang, Xinghao Zhu, Yuxin Chen, and 3 more authors
    ICRA Safe-VLM Workshop (Spotlight)
  4. L4DC 2026
    TDMPC_square.jpg
    TD-M(PC)²: Improving Temporal Difference MPC Through Policy Constraint
    Haotian Lin, Pengcheng Wang, Jeff Schneider, and 1 more author
    L4DC 2026