Papers
2
Total Citations
21
H-Index
2
About
Sitong Mao is a rising roboticist whose work pushes the boundaries of autonomous navigation and large-scale environmental mapping. His primary research focuses on neural implicit representations for LiDAR mapping and deep reinforcement learning for visual navigation. In his landmark 2023 paper, "NF-Atlas: Multi-Volume Neural Feature Fields for Large Scale LiDAR Mapping," Mao introduced a novel framework that bridges neural feature volumes with pose graph optimization, enabling robots to construct continuous, high-fidelity maps of expansive environments—a persistent challenge in field robotics. This work has already garnered 18 citations, signaling its impact on the SLAM community. More recently, Mao tackled the problem of real-world robot deployment with "SCALE: Self-Correcting Visual Navigation for Mobile Robots via Anti-Novelty Estimation" (2024), where he developed a method that allows robots to learn from offline datasets and correct their own navigation errors by detecting environmental novelty. This approach promises to make visual navigation more robust and generalizable without costly online training. With a clear trajectory from foundational mapping techniques to practical, self-correcting navigation systems, Sitong Mao is establishing himself as a key innovator in bringing neural representations to real-world robotic autonomy.
Research Focus
Key Achievements
Top Papers
- 1NF-Atlas: Multi-Volume Neural Feature Fields for Large Scale LiDAR Mapping18 citations · 2023
- 2