Hoang-Giang Cao

Ming Chi University of Technology, National Yang Ming Chiao Tung University

Papers

Total Citations

H-Index

About

Hoang-Giang Cao is a robotics researcher whose work bridges the critical gap between simulation and real-world deployment, with a focus on deep reinforcement learning, robotic manipulation, and human-robot collaboration. His major contributions include developing a gradient-based regularization method that ensures action smoothness in robotic control, directly addressing the challenge of applying DRL to physical systems where jerky movements can be catastrophic. In manipulation, he pioneered sim-to-real transfer techniques for learning dense object descriptors, enabling robots to visually understand and interact with objects using rich, transferable representations—a foundational step toward ubiquitous robotics. His most recent work introduces mmPrivPose3D, a privacy-compliant RaDAR-based system for 3D pose estimation and gesture command recognition in collaborative manufacturing, offering a safer alternative to RGB-D cameras in dynamic environments. With publications from 2023–2025, Cao’s research is already gaining traction, and his focus on practical, deployable solutions positions him as an emerging leader in making robots safer, smoother, and more intelligent in real-world settings.