Zijia Li

Beijing Jiaotong University, Chinese Academy of Sciences, The University of Tokyo

Papers

Total Citations

H-Index

About

Zijia Li is a robotics researcher whose work spans reconfigurable mechanism design, kinematic analysis, and robot learning from visual data. Li’s most cited paper introduces a foldable, reconfigurable multi-terrain vehicle with a variable wheelbase, designed to traverse challenging environments like crevices and pipes—a contribution that has already garnered 10 citations for its practical impact on rescue and detection robotics. In parallel, Li has advanced the theoretical understanding of redundant serial manipulators through kinematic redundancy analysis of (2n+1)R circular manipulators, offering insights into optimizing joint configurations for dexterous tasks. A third line of work tackles the challenge of enabling robots to acquire mechanical knowledge directly from 3D point clouds using deep learning, a novel approach that allows machines to infer object properties and use them in unfamiliar scenarios. Earlier, Li explored robotic self-action recognition, developing methods for robots to understand and verbalize their own behaviors from first-person perspectives. With a portfolio that bridges hardware innovation, theoretical kinematics, and perception-driven learning, Li’s research demonstrates a commitment to creating robots that are both physically adaptable and cognitively aware, laying groundwork for more autonomous systems in unstructured environments.