Biqi Yang

Papers

Total Citations

H-Index

About

Biqi Yang is a leading researcher in robotic perception and manipulation, specializing in bridging the sim-to-real gap for industrial automation. Her work focuses on developing deep-learning frameworks that enable robots to robustly recognize, localize, and grasp objects in cluttered, unstructured environments—critical for tasks like bin picking in logistics and manufacturing. Yang’s most cited paper (52 citations) introduces S2R-Pick, a generic sim-to-real framework that achieves fast, accurate object recognition and localization for industrial bin picking, overcoming the limitations of models trained solely on synthetic data. She further advances grasping reliability with her uncertainty-aware suction grasping pipeline (7 citations), which improves generalization to unseen objects and noisy sensor data. Her self-ensembling approach to instance segmentation (SESR) eliminates the need for costly real-world annotations, enabling zero-shot transfer from simulation to real-world auto-store scenarios. Through these contributions, Yang has established herself as a key innovator in practical, scalable robotic grasping, with her work directly impacting the efficiency of automated warehouses and manufacturing lines.