Max Robotics — Global Service Robot Directory + US Certification

About

Yifeng Huang’s research lies at the intersection of robotic perception, motion planning, and visual place recognition (VPR), with a sustained focus on enabling robots to operate reliably in large-scale, uncertain environments. In his early work, Huang tackled the fundamental challenge of integrating motion and map uncertainty into exploration, introducing the RRT-SLAM framework (30 citations) that holistically addresses the motion planning subproblem for holonomic robots equipped with laser range sensors. More recently, he has driven the field of VPR toward efficiency and robustness. His LSDNet (12 citations) pioneered a lightweight self-attentional distillation network that balances high-performance encoding with computational efficiency for long-term robot operation. Building on this, his CAHIR framework (2023) introduces co-attentive hierarchical image representations that unify global and local descriptors, achieving robust place recognition under significant appearance changes. Earlier contributions include an adaptive configuration-space and workspace criterion for view planning (2005), which advanced the next-best-view problem for eye-in-hand robot-sensor systems. Across his career, Huang has consistently pushed the boundaries of how robots perceive, plan, and localize—from foundational uncertainty-aware exploration to cutting-edge lightweight deep learning architectures for lifelong visual navigation.