Yimin Lin

Cloud Computing Center, Renmin University of China

Papers

Total Citations

H-Index

About

Yimin Lin is a researcher advancing the frontiers of robot perception and autonomous navigation, with a core focus on visual localization, odometry, and Simultaneous Localization and Mapping (SLAM). Lin’s major contributions lie in developing deep learning architectures that fuse visual and inertial data to achieve robust, long-term 6-Degrees-of-Freedom (6-DoF) pose estimation. Notably, Lin proposed the Deep Global-Relative Networks, an end-to-end framework that mitigates the persistent drift problem in visual odometry by integrating global and relative pose constraints, a work that has garnered 16 citations. Complementing this, Lin introduced a novel feedback mechanism-based stereo visual-inertial SLAM system, which balances real-time performance with accuracy—a critical challenge in the field, earning 10 citations. This work demonstrates a sophisticated understanding of sensor fusion, combining camera and IMU data to enhance localization reliability in complex environments. Through these innovations, Lin has contributed to making autonomous systems more resilient for long-term deployment, with cumulative citation counts reflecting the community’s recognition of these practical solutions. Lin’s research is particularly valuable for students and engineers tackling drift and efficiency in robot navigation, offering principled, data-driven approaches to a fundamental robotics problem.