Papers
6
Total Citations
119
H-Index
5
About
Shitong Du is a robotics and autonomous systems researcher whose work spans simultaneous localization and mapping (SLAM), LiDAR-based perception, and reinforcement learning for robotic control. His most influential contribution, "Integrate Point-Cloud Segmentation with 3D LiDAR Scan-Matching for Mobile Robot Localization and Mapping" (2019, 62 citations), advanced the field by combining semantic point-cloud segmentation with Iterative Closest Point (ICP) scan-matching, moving beyond purely geometric approaches to enable more robust mobile robot localization. Building on this foundation, Du has consistently pushed the boundaries of LiDAR odometry, incorporating semantic information and GNSS assistance to improve SLAM performance in challenging outdoor and urban environments. His 2021 work on semantically enriched LiDAR odometry and mapping (20 citations) further demonstrated his commitment to bridging perception and navigation. Notably, Du also contributed to reinforcement learning through his Actor-Dueling-Critic method (21 citations), offering an improved value-function framework for robotic control. His curvefusion trajectory-combination approach and vision-based instance segmentation work reflect a broad, interdisciplinary perspective on spatial intelligence. Collectively, Du's research addresses critical challenges in enabling autonomous robots to reliably perceive, localize, and navigate complex real-world environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2The Actor-Dueling-Critic Method for Reinforcement Learning21 citations · 2019
- 3
- 4GNSS-Assisted LiDAR Odometry and Mapping for Urban Environment9 citations · 2023
- 5
- 6