Jian Pu

Papers

Total Citations

H-Index

About

Jian Pu is a researcher whose work sits at the intersection of robotics, computer vision, and artificial intelligence, with a particular focus on advancing autonomous navigation and intelligent control systems. His most significant contribution to date is the development of UL-SLAM, a universal monocular line-based Simultaneous Localization and Mapping system that overcomes the limitations of traditional SLAM approaches. By unifying structural and non-structural line constraints, Pu’s method addresses the triangulation degeneracy problem and tracking instability that previously restricted line-based SLAM to Manhattan-world assumptions, enabling robust performance in more general environments. This work has already garnered 17 citations since its 2024 publication, reflecting its immediate impact on the field. Additionally, Pu has explored sample-efficient deep reinforcement learning for robot control, proposing a demonstration-guided approach that mitigates the sparse reward problem, and has contributed to point cloud processing with GS-Net, a graph neural network-based sampling method. His research demonstrates a clear trajectory toward making autonomous systems more reliable and efficient, with potential applications ranging from service robotics to autonomous driving.