Xiangyu Chen - Homepage

Xiangyu Chen (陈翔宇)

I am a First-year Mphil student at the Hong Kong University of Science and Technology (Guangzhou). Before that, I was a research intern at TARS Robotics and Tsinghua University(AIR), co-advised by Prof. Yilun Chen and Prof. Wenchao Ding, working closely with PhD Candidate Ruihai Wu and Prof. Xiaoxiao Long. Earlier in my academic journey, I obtained First Class Honour Bachelor degree in Electrical and Electronic Engineering from Liverpool John Moores University in UK.

News

2025-09: My team won The First Place 🏆 in CVPR 2025 Embodied AI Workshop

2025-05: My team won The First Place 🏆 in ICRA 2025 What Bimanuals Can Do (WBCD) Challenge

2025-03: My team won the Champion 🏆 in ManiSkill-ViTac 2025: Challenge on Manipulation Skill Learning With Vision and Tactile Sensing

2024-07: GuassianGrasper was accepted to RA-L

2024-02: Block-Map-Based Localization in Large-Scale Environment was accepted to ICRA 2024

Research Interests

🤖Robotics

Embodied AI

Manipulation

VLA

Tactile

Mapping and Navigation

Education

	the Hong Kong University of Science and Technology (Guangzhou) Master of Philosophy (General) (2025.9 - present)
	Liverpool John Moores University, UK B.E Electrical and Electronic Engineering (2019-2023)

Research Experience

	TARS Technology (Shanghai) Co., LTD. Research Intern, supervised by Prof. Yilun Chen and Prof. Wenchao Ding (2025.3 - present)
	Institute for AI Industry Research (AIR), Tsinghua University Research Intern, supervised by Prof. Yilun Chen and Prof. Guyue Zhou (2021.6 - 2023.12, 2024.12 - present)
	GAIRLAB, City University of Hong Kong (CityU) Research Assistant, supervised by Prof. Peng Yin (2024.3 - 2024.6)
	EncoSmart Technology (Beijing) Co., LTD Research Intern (2023.11 - 2024.2, 2024.7 - 2024.11)

Publications

(^†: corresponding author; ^*: equal contribution)

	RoboGEM: Learning Language-guided Robotic Manipulation via Generalizable and Efficient Feature Distillation Chunzheng Wang, Yuhang Zheng, Xiangyu Chen, Weize Li, Songen Gu, Yupeng Zheng. ACM International Conference on Multimedia (MM) 2025 RoboSoft'25 Workshop (Oral) Best Paper Finalist (Certificate)
	GaussianGrasper: 3D Language Gaussian Splatting for Open-vocabulary Robotic Grasping Yuhang Zheng, Xiangyu Chen, Yupeng Zheng, Songen Gu, Runyi Yang, Bu Jin, Pengfei Li, Chengliang Zhong, Zengmao Wang, Lina Liu, Chao Yang, Dawei Wang, Zhen Chen, Xiaoxiao Long^†, Meiqing Wang^†. *RA-L, 2024* Homepage · 📑Paper · Code We introduced GaussianGrasper, a robot grasping system implemented by a 3D Gaussian field endowed with open-vocabulary semantics and accurate geometry that is capable of rapid updates to support open-world robotic grasping guided by language.
	Block-Map-Based Localization in Large-Scale Environment Yixiao Feng, Zhou Jiang, Yongliang Shi, Yunlong Feng, Xiangyu Chen, Hao Zhao, Guyue Zhou. *ICRA 2024* 🎬Video · 📑Paper · Code we propose a localization system based on Block Maps (BMs) to reduce the computational load caused by maintaining large-scale maps.

Projects

Mobile Manipulation in household tasks

Xiangyu Chen (cooperated with YueXiu Company)

Whole Body Control

VR Teleoperation

Media: Yuexiu Property

Humanoid Unitree G1 RL locomotion

Xiangyu Chen (Independent Project)

Algorithm: Reinforcement Learning (PPO)

Simulation: Isaac Gym, Mujoco

I implemented reinforcement learning (PPO) algorithm using the rsl_rl library and Isaac Gym to train locomotion policy, and use Mujoco Simulation to test the policy

ICRA 2025 What Bimanuals Can Do (WBCD) Challenge

Hierarchical Bimanual Manipulation Policy for Dining Room Service

Weize Li, Zhengxiao Han, Lixin Xu, Xiangyu Chen, Harrison Bounds, Chenrui Zhang, Yifan Xu,

ICRA 2025 🏆 The First Place (Certificate)

Media: 甲子光年 | 36氪 | 机器之心 | 量子位 | ARX方舟无限

	ManiSkill-ViTac 2025: Challenge on Manipulation Skill Learning With Vision and Tactile Sensing Champion 🏆 in Track 2: Tactile-Vision-Fusion Manipulation (Certificate) The First Place 🏆 in CVPR 2025 Embodied AI Workshop(Certificate) TARS-STEIIA Team Team Leader: Ruihai Wu Instructor: Xiaoxiao Long Members Xiang Xiao^, Yuxing Qin^, Yuhang Zheng^, Xiangyu Chen*
	Mobile Manipulation with Imitation Learning GAIRLAB, City University of Hong Kong (CityU) Kaixin Chai, Xiangyu Chen, Yufeng Li, Peng Yin* 🎬Video Imitation Learning(ACT) Data Collection by Teleoperation System Whole-Body Control
	Wheel-legged Robot dynamics obstacle avoidance Xiangyu Chen, Yufeng Li(Hardware Design) 🎬Video Contribution: Lidar-based SLAM (Cartographer) Navigation (A* and TEB)
	Wheel-legged Robot static obstacle avoidance Xiangyu Chen, Yufeng Li(Hardware Design) 🎬Video Contribution: Lidar-based SLAM (Cartographer) Navigation (A* and TEB)
	Navigation in 3D Large-Scale Environment DISCOVER Lab, AIR, Tsinghua University Xiangyu Chen, Yixiao Feng, Yunlong Feng, Yongliang Shi, Guyue Zhou Lidar-based SLAM (LIO-SAM) block-based-localizaton Navigation (A* and TEB)
	Leader-Follower Multi-Agent Swarm (Sim2Real Transfer) Zhengxiao Han, Jingtian Deng, Chenghao Lin, Xiangyu Chen, Tianao Ren Guyue Zhou Lidar-based SLAM (Cartographer) Pure Pursuit algorithm High stability and obstacle avoidance ability
	Dynamics Obstacles Avoidance Indoor DISCOVER Lab, AIR, Tsinghua University Xiangyu Chen, Zhengxiao Han, Yixiao Feng, Guyue Zhou. Lidar-based SLAM (Cartographer), Vision-based SLAM (ORB_SLAM, RTAB_MAP) Extended Kalman Filter (EKF) Navigation (A* and TEB)
	Cyberdog Navigation in indoor environment DISCOVER Lab, AIR, Tsinghua University Xiangyu Chen, Zhengxiao Han, Guyue Zhou 🎬Video Lidar-based SLAM (Cartographer) Extended Kalman Filter (EKF) Navigation (A* and TEB)
	Cyberdog SLAM in indoor environment DISCOVER Lab, AIR, Tsinghua University Xiangyu Chen, Zhengxiao Han, Guyue Zhou 🎬Video Lidar-based SLAM (Cartographer)
	End-to-end Visual Navigation DISCOVER Lab, AIR, Tsinghua University Guan Wang, Haoyi Niu, Xiangyu Chen, Guyue Zhou 🎬Video Reinforcement Learning Sim2Real

Simulation

	Slave Obstacle Avoidance Gazebo Jingtian Deng, Xiangyu Chen, Guyue Zhou Artificial Potential Field Method
	Wheeled-Legged Navigation Isaac Sim Zhengxiao Han, Xiangyu Chen, Guyue Zhou

Industrial Collaboration

Cooperation with iFytek: VLM for Mobile Manipulation

Pengfei Li, Xiangyu Chen,

🎬Video

The industrial compound mobile robot, enhanced by the capabilities of the Antelope Industrial Large Model in understanding computations, generating industrial code, and processing industrial text, can perform task comprehension, decomposition, instruction generation, and other operations through applications such as industrial task splitting, path code planning, and industrial object detection and image segmentation within a natural language interaction model. This further enhances the efficiency and effectiveness of industrial robots.

Baidu Apollo V2X Vehicle-Road Cooperative Technology Model Construction and Validation

Website

Contribution & Results: we developed a safety model based on two scenarios: ring road conflicts and unprotected left turns, and evaluated the safety performances in both individual vehicle intelligence and V2X scenarios. The results indicate that V2X improved safety metrics by 30% to 90%.

Result: In collaboration with Baidu, contributed to the release of the first Chinese white paper on vehicle-road cooperative technology: Key technologies and prospects of vehicle-road collaboration for autonomous driving.

Awards

City Special Prize, "Unbounded·2023 Shanghai International Student (Nationalized University Students) Innovation and Entrepreneurship Competition"

City Top 10, Best Popularity Award, Second and Third Prizes, 2020、2022 China-U.S. Maker Competition (Shanghai Regional)

Third Prize, 2021 Cross-Strait Maker Competition National Finals

City Second and Third Prizes, 2019、2022 Shanghai College Student Maker Competition

First Prize (National Level), 2017 International Youth Innovation Design Competition China Region

Hobbies

Outside of research, I enjoy playing basketball.