Papers
6
Total Citations
73
H-Index
3
About
Xinhu Zheng is at the forefront of embodied AI and autonomous navigation, pioneering systems that fuse multimodal sensing with cutting-edge neural representations. Their most impactful work, "LIV-GaussMap" (2024, 54 citations), introduces a revolutionary LiDAR-Inertial-Visual fusion framework that leverages differentiable Gaussians to create real-time, high-fidelity 3D radiance field maps—a tightly coupled approach that dramatically improves structural accuracy and mapping quality for mobile robots. Zheng has also made significant strides in semantic navigation, developing LFENav, which harnesses large language models for frontier-based exploration, and E²BA, a backtracking agent that overcomes spatial awareness limitations in unknown environments. Their research extends to safety-critical planning, with work on multi-risk aware trajectory generation for highly dynamic spaces. By addressing fundamental challenges in mapless navigation, multi-agent coordination, and real-time environmental understanding, Zheng’s contributions are shaping the next generation of intelligent, autonomous systems capable of operating reliably in complex, unstructured real-world settings.
Research Focus
Key Achievements
Top Papers
- 1
- 2LFENav: LLM-Based Frontiers Exploration for Visual Semantic Navigation6 citations · 2024
- 3
- 4Multi-agent embodied AI: advances and future directions3 citations · 2026
- 5
- 6