Kaiwen Yuan

Papers

Total Citations

142

H-Index

About

Kaiwen Yuan is a leading researcher in the field of autonomous systems and robotic perception, with a primary focus on sensor fusion, calibration, and inertial navigation. His work addresses critical challenges in integrating LiDAR, camera, and IMU data for reliable real-world deployment. Yuan’s most influential contribution is **RGGNet** (101 citations), a pioneering deep learning framework for tolerance-aware LiDAR-camera online calibration that replaces traditional hand-crafted features with geometric deep learning and generative models—significantly improving scalability and accuracy for autonomous vehicles. He also developed a **simple self-supervised IMU denoising method** (31 citations) that enhances inertial-aided navigation by removing noise from low-cost IMUs without requiring ground-truth data, making robust navigation more accessible. Additionally, his work on **LiCaS3** (10 citations) introduces a self-supervised approach to LiDAR-camera temporal synchronization, solving a practical bottleneck in multi-sensor fusion. Yuan’s research is distinguished by its emphasis on self-supervised learning and geometric reasoning, enabling systems to calibrate and synchronize autonomously. His contributions are vital for advancing reliable perception in autonomous driving, robotics, and mobile platforms, and his methods are increasingly adopted by both academia and industry for their practicality and performance.