01
Fault-Tolerant Quadruped Locomotion
Learning and safety-filtering methods that help a Unitree Go2 remain stable under weakened, locked, and zero-torque actuator failures.
Jingxi (Blair) Lu
Ph.D. student in robotics at NC State University.
I work with Prof. Jaemin Lee in the Hybrid Intelligent Experimental Robotics Lab. My research focuses on safe learning, fault-tolerant control, and loco-manipulation for legged and humanoid robots.
I received my M.S. in Electrical Engineering from the University of Southern California in 2026. Previously, I worked on quadruped locomotion, bipedal navigation, humanoid control, efficient vision-language model inference, and learning-assisted optimization.
Selected projects
A brief overview
View project →
02
STATE-NAV
Stability-aware traversability estimation and hierarchical planning for bipedal robots navigating rough terrain.
03
Visual-Token Pruning for Vision-Language Models
Training-free methods that reduce inference cost for autoregressive and diffusion vision-language models.