Papers
6
Total Citations
93
H-Index
5
About
Yawen Lu is a computer vision and robotics researcher whose work centers on depth estimation, visual odometry, and autonomous perception systems. Best known for pioneering unsupervised and self-supervised approaches to scene understanding, Lu has made significant contributions to reducing reliance on costly, high-resolution LiDAR sensors in autonomous driving and robotics applications. Lu's most influential work, "An Alternative of LiDAR in Nighttime: Unsupervised Depth Estimation Based on Single Thermal Image" (2021, 48 citations), introduced a novel passive sensing approach that addresses critical limitations of active LiDAR systems, including resolution constraints and potential environmental harm. This work opened new possibilities for nighttime autonomous navigation. Complementing this, Lu's research on simultaneous visual odometry and depth estimation (2019, 24 citations) demonstrated how deep unsupervised learning frameworks can jointly solve two fundamental robotics challenges, advancing camera-based scene reconstruction without ground-truth supervision. Lu has further explored LiDAR-image fusion, self-supervised spectral consistency methods, and most recently, event-based vision transformers for robotic gripper force measurement, reflecting a broadening research trajectory toward tactile and embodied robotics. With nearly 90 total citations across their body of work, Lu represents an emerging voice pushing the boundaries of cost-effective, sensor-agnostic perception for intelligent systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6