Long Zeng
Papers
4
Total Citations
25
H-Index
4
About
Long Zeng is a leading researcher in mobile robotics, focusing on dynamic scene understanding, robot navigation, and pedestrian trajectory prediction. His major contributions include the creation of the THUD (Tsinghua University Dynamic) robotic dataset, a large-scale indoor dataset designed to train and evaluate robots in dynamic environments, addressing the limitations of static datasets. This work has garnered 11 citations since 2024. Zeng also advanced robot navigation with a two-stage reinforcement learning approach for long-range indoor dense crowd environments, cited 6 times, and introduced FusedNet, an end-to-end relocalization network using cross-attention to fuse global and local image features for improved accuracy in dynamic scenes, with 4 citations. Additionally, his work on dual-alignment domain adaptation for pedestrian trajectory prediction enhances multi-scene model performance, also cited 4 times. Zeng’s research is pivotal for applications in autonomous driving and service robotics, demonstrating significant impact through innovative datasets and algorithms that push the boundaries of robot mobility in complex, human-inhabited spaces.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Dual-Alignment Domain Adaptation for Pedestrian Trajectory Prediction4 citations · 2024
- 4