Max Robotics — Global Service Robot Directory + US Certification

About

Haoyuan Fu is a rising researcher at the intersection of embodied AI, robotic manipulation, and simulation environments. His work focuses on enabling robots to perceive and interact with the complex, articulated objects that populate human environments—from cabinets and drawers to transparent glassware. Fu’s most impactful contribution is the **AKB-48 knowledge base** (64 citations), a real-world dataset of articulated objects that provides comprehensive annotations of appearance, structure, physics, and semantics, moving beyond synthetic benchmarks to ground robotic understanding in reality. He also pioneered **Category-level Articulation Pose Estimation (CAPE)** (37 citations), tackling the challenging problem of estimating part-level 6D poses across diverse object instances. Recognizing the need for realistic training grounds, Fu co-created **RCareWorld** (35 citations), a human-centric simulation for caregiving robots that incorporates stakeholder input to model care recipients and accessible home environments. His work on **RFUniverse** and **RFTrans** further extends robotic capability to multiphysics phenomena and transparent object manipulation, using refractive flow for accurate surface normal estimation. With over 160 total citations and a clear trajectory from knowledge bases to real-world deployment, Fu is shaping how robots learn to assist humans in daily life.