About Me
Hello! I’m Zhaofeng Luo (罗兆丰), a PhD student in the School of Computer Science at Carnegie Mellon University, advised by Prof. Minchen Li.
I develop robust and efficient physics-based simulation methods, with a focus on GPU computation, contact, and deformable materials. My work connects numerical methods and interactive systems: from accelerating contact-rich simulation to modeling adhesive tape for robotic manipulation and elastoplastic materials for hands-on VR modeling.
CV (PDF) · Publications · GitHub · Email
Selected Research
TapeSim: Adhesive Tape for Robotic Manipulation
TapeSim models a dispensing roll with a rigid cluster, an advancing deformable collar, and a flexible, reattachable strip. The simulator supports attachment, dispensing, cutting, and sealing in a continuous teleoperated box-sealing workflow.
At 32 turns, clustering gives 4.5–8.4× faster mean physics steps at comparable roll motion. Evaluation includes real-motion replays and paired real-to-sim manipulation cases.
AGIPC: Adaptive In-Solve Algebraic Coarsening for GPU IPC
Adaptive coarsening inside the Newton solve reduces computational cost without remeshing, with GPU-parallel assembly and contact handling.
Penetration-Free Elastodynamics without Barriers
An augmented Lagrangian approach to contact-rich elastic simulation, addressing the conditioning and collision-detection challenges of barrier-based methods.
Overlapping Schwarz Space-Time Refinement for MPM
Local spatial refinement and temporal sub-cycling concentrate computation where deformation and contact require it, while preserving standard MPM discretizations within each subdomain.
VR-Doh: Hands-on 3D Modeling in Virtual Reality
An interactive modeling system that combines MPM-based elastoplastic simulation and natural hand interaction. I designed, implemented, and tested the full system. Selected as a Top 10 Technical Papers Fast Forward.
Open-Source and Teaching Resources
- Libuipc — contributions to a C++20 framework for GPU incremental potential contact and coupled rigid and deformable simulation. Project
- GPU solid simulation tutorial — tutorial code and course material for the open-source Physics-based Simulation book. Code · Chapter
- Teaching assistant, CMU 15-369/769 — Numerical Methods: Foundations, ML, and Visual Computing, Fall 2026. Course
Outside Research
I enjoy tennis. At Peking University, I served as captain of the School of EECS tennis team and president of the Student Tennis Association, organizing tournaments and classes.