Tianyu Xu

Tianyu Xu

B.S. Student in Computer Science

Peking University, School of EECS

Class of 2023 ยท Expected 2027

Embodied AI Visual Navigation Spatial Reasoning Robot Learning VLA Models

About

I am an undergraduate student in Computer Science at the School of Electronics Engineering and Computer Science, Peking University. Since April 2025, I have been conducting undergraduate research in embodied AI and robot navigation under the guidance of Prof. He Wang.

My current research focuses on embodied navigation, multi-view vision-language-action models, spatial awareness, reinforcement learning, and sim-to-real validation. I am also interested in long-horizon mobile manipulation, navigation-manipulation coordination, multimodal world models, and robot generalization.

Education

Peking University

2023 - 2027 expected

B.S. in Computer Science, School of Electronics Engineering and Computer Science.

News

  • 2026.06 Personal homepage updated.
  • 2026.03 SPAN-Nav is available on arXiv.
  • 2025.10 MM-Nav is available on arXiv.
  • 2025.04 Started undergraduate research on embodied navigation.

Publications

MM-Nav pipeline overview

MM-Nav: Multi-View VLA Model for Robust Visual Navigation via Multi-Expert Learning

Under Review

Tianyu Xu*, Jiawei Chen*, Jiazhao Zhang*, Wenyao Zhang, Zekun Qi, Minghan Li, Zhizheng Zhang, He Wang

A multi-view VLA navigation model that distills multiple specialized RL experts into a single RGB-based policy for robust visual navigation and obstacle avoidance.

* Equal contribution.

SPAN-Nav teaser overview

SPAN-Nav: Generalized Spatial Awareness for Versatile Vision-Language Navigation

Under Review

Jiahang Liu*, Tianyu Xu*, Jiawei Chen*, Lu Yue*, Jiazhao Zhang*, Zhiyong Wang*, Minghan Li, Qisheng Zhao, Anqi Li, Qi Su, Zhizheng Zhang, He Wang

An end-to-end navigation model that introduces generalized 3D spatial awareness from RGB video streams for robust vision-language navigation across diverse environments.

* Equal contribution.

Contact

Email: 2300013131@stu.pku.edu.cn

Alternative email: xdsyxty@gmail.com

GitHub: https://github.com/xdsyxty