TIANXIN HU

TIANXIN HU

I am Tianxin Hu, recently graduated with an M.Sc. in Computer Control and Automation from Nanyang Technological University (NTU), Singapore, under the supervision of Prof. Lihua Xie. My background is in optimization-based planning and control for autonomous systems, with practical experience in ROS/ROS2, C++/Python, and simulation-driven validation.

I am currently transitioning toward reinforcement learning for robot locomotion, with ongoing work spanning quadruped locomotion and humanoid control. More broadly, I am interested in integrating structured dynamical priors into learning-based control to achieve robust, interpretable, and natural robot locomotion. Selected ongoing work is listed below under Ongoing Explorations.

πŸ’Ό Current Position

  • Jun 2026 – Present, Nanyang Technological University (NTU), Singapore
    Research Assistant, Prof. Lihua Xie’s Group
    Working on a humanoid robotics project with Chery Robotics, focusing on reinforcement learning for humanoid locomotion in embodied task execution scenarios.

πŸ“– Education

  • Aug 2024 – Jan 2026, Nanyang Technological University, Singapore
    M.Sc. in Computer Control and Automation, School of Electrical and Electronic Engineering
    GPA: 95/100

  • Sep 2020 – Jun 2024, Wuhan University of Science and Technology, China
    B.Eng. in Vehicle Engineering (Industrial Planning), School of Automobile and Traffic Engineering
    GPA: 88.3/100

πŸ“ Publications

βœ… Accepted & Published

IROS 2026
Overview of the multi-agent off-world exploration framework

Multi-Agent Off-World Exploration for Sparse Evidence Discovery via Gaussian Belief Mapping and Dual-Domain Coverage

Accepted by IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , September 2026

Co-author

Zhuoran Qiao, Tianxin Hu, Thien-Minh Nguyen, Shenghai Yuan

We propose a multi-agent informative path planning framework for sparse evidence discovery in hazardous off-world environments. It combines Gaussian-process interest and risk belief maps, trajectory-intent coordination, and dual-domain coverage to improve robustness under biased areas of interest and limited communication.
ICARCV 2026 Oral
Unitree Go2 robot platform and its interactive object-tracking result

Interactive Object Tracking on Gait-Parameterized Locomotion Policy

Accepted as an oral presentation at the International Conference on Control, Automation, Robotics and Vision (ICARCV) , November 2026

Co-author

Gia-Hung Huynh, Tianxin Hu, Tien-Dat Nguyen, Shenghai Yuan, Thien-Minh Nguyen

We integrate a gait-parameterized locomotion policy as an inner-loop controller with lightweight vision-based ball detection and classical feedback control. The resulting Unitree Go2 system tracks objects in real time while expressing its operational state through trotting, pronking, pacing, and bounding gaits.
ICCA 2026 Oral
sym

Energy-Constrained Navigation for Planetary Rovers With Singular Internal Power Source

Accepted as an oral presentation at the IEEE International Conference on Control & Automation (ICCA), April 2026

Tianxin Hu*, Weixiang Guo*, Rui Qian, Jiaye Jin, Haoran Zhao, Shenghai Yuan, Lihua Xie

We present an energy-constrained trajectory planning framework for planetary rovers powered by an RTG. The method ensures smooth and dynamically feasible trajectories while strictly satisfying instantaneous power and cumulative energy limits.
ICRA 2025
sym

Swept volume-aware trajectory planning and mpc tracking for multi-axle swerve-drive amrs

Accepted by IEEE International Conference on Robotics and Automation (ICRA) , May 2025

First author

Tianxin Hu, Shenghai Yuan, Ruofei Bai, Xinhang Xu, Yuwen Liao, Fen Liu, Lihua Xie

We propose a swept volume-aware trajectory planning and MPC tracking framework for multi-axle AMRs, enabling real-time path optimization that minimizes swept volume and enhances maneuverability in confined environments.
JAI 2025
sym

Tire wear aware trajectory tracking control for Multi-axle Swerve-drive Autonomous Mobile Robots

Accepted by Journal of Automation and Intelligence (JAI) , June 2025

first author

Tianxin Hu, Xinhang Xu, Thien-Minh Nguyen, Fen Liu, Shenghai Yuan, Lihua Xie

We present a hierarchical model predictive control framework for multi-axle swerve-drive robots that integrates tire-wear modeling into the optimization objective, achieving up to 65% reduction in tire wear while maintaining high trajectory-tracking accuracy.
CVPR 2026
sym sym

TGSFormer: Scalable Temporal Gaussian Splatting for Embodied Semantic Scene Completion

Accepted by IEEE Conference on Computer Vision and Pattern Recognition(CVPR) , March 2026

Co-author

Rui Qian*, Haozhi Cao*, Tianchen Deng*, Tianxin Hu, Weixiang Guo, Shenghai Yuan, Lihua Xie

We propose TGSFormer, a scalable Temporal Gaussian Splatting framework that maintains persistent Gaussian memory for temporal prediction with dual temporal encoding and confidence-aware voxel fusion. It achieves state-of-the-art results on embodied SSC benchmarks using significantly fewer primitives.

πŸ€– Ongoing Explorations

Go2 Extreme Parkour Extension β€” Faster Training and More Stable, Natural Locomotion

Building on Extreme Parkour [1], I extended the locomotion training pipeline for the Unitree Go2 with a focus on stability, training efficiency, and more natural gait behaviour.

β‘  More robust adaptation

Expanded the RMA latent representation to retain richer terrain and dynamics context, improving policy stability during challenging parkour manoeuvres.

β‘‘ 3.5Γ— faster training

Replaced complex mesh-based training terrain with lightweight box primitives, increasing training throughput by approximately 3.5Γ—.

Parkour policy evaluation in Isaac Sim

Sim-to-real deployment on the Unitree Go2

Together, these changes make the training pipeline substantially more efficient while improving the stability and motion quality of the resulting parkour policy in both simulation and real-world deployment.

Reference

[1] Extreme Parkour with Legged Robots

Xuxin Cheng, Kexin Shi, Ananye Agarwal & Deepak Pathak Β· IEEE International Conference on Robotics and Automation (ICRA), 2024

πŸ“œ Patents

[P1] (Granted) Multi-axle all-wheel steering vehicle vector-division correction method and system Sep 2025
Tianxian Hu, Xiangrong You, Dongmei Wang, Xinhe Long, Xinyu Wang, Shuang Liu, Yutong Fu, Xinyao Wang, Lefeng Wei, Baizhen Zhang
Patent No.: CN116534005B

[P2] (Granted) A system and method for automatically correcting the deviation of a beam transport vehicle in a tunnel Jul 2023
Xiangrong You, Enze Jin, Dongmei Wang, Yuran Huang, Tianxian Hu, Shuai Zhu, Beijia Zhang
Patent No.: CN114633801B

[P3] (Pending) Energy-saving system and method for beam transport vehicle based on traveling pressure control Feb 2024
Wei Li, Xiangrong You, Dongmei Wang, Xinhe Wang, Peng Cheng, Yuming Liu, Shuang Hu, Tianxian Hu, Xinyao Wang, Lefeng Wei, Baizhen Zhang
Application No.: CN117508238A

[P4] (Pending) Method, system and equipment for planning obstacle avoidance path of multi-shaft all-wheel steering beam transporting vehicle Dec 2023
Xiangrong You, Yucheng Yin, Liwei Wang, Dongmei Wang, Tianxian Hu, Shuang Liu, Xinyao Wang, Lefeng Wei, Baizhen Zhang
Application No.: CN117168481A

πŸš€ Projects

πŸ’Ό Professional Projects

AIBOT Technology Pte. Ltd. β€” Autonomous Navigation System Development

Supervisor: Professor Lihua Xie, Dr. Shenghai Yuan
Position: Navigation Module Lead

Jul 2025 - Jun 2026

  • Responsibilities: Developed the full-stack navigation module for the company’s autonomous mobile robot, covering cross-floor navigation, and dynamic obstacle avoidance in residential and commercial environments.
  • Achievements: Achieved fully autonomous navigation and delivery within multi-story residential complexes.

Xin Chi Technology β€” Research and Development of Autonomous Driving System for Multi-Axle Heavy-Duty Transport Vehicles

Supervisor: Senior Engineer Xiangrong You
Position: Technical Director & Corporate Legal Representative

Jun 2022 - Jun 2024

  • Responsibilities: Led R&D of autonomous driving systems for heavy-duty vehicles, covering mathematical modeling, microcontroller development, planning and control.
  • Achievements: Contributed to multiple patents [P1-P4] and received several national awards [A2-A6].

WUST-CHIJI Formula Racing Team Development

Supervisors: Associate Professor Chao Deng, Dr. Junyi Zou
Position: Leader of the Unmanned Group & Member of Electronic System Group

Feb 2021 - Jun 2024

  • Responsibilities: Led the design and implementation of the autonomous system for the Chiji Formula Racing Team, covering detection, mapping, planning, and tracking.
  • Achievements: Received the FSAC Third Prize, Best Rookie Award [A7].

🎨 Toy Projects

Three-Stage Coilgun PCB Diagram

Three-Stage Coilgun Demonstration

Autonomous Racing Simulation

Spinal Curvature Measurement Program

Robotic Arm Model-Based Control Simulation

Material Handling Demonstration

πŸŽ– Honors and Awards

[A1] (International Level) Best Poster Award β€” ICRA 2026 Workshop Jun 2026
Robots Meet Prior Maps: Leveraging Geometric and Semantic Priors for Real-World Robot Autonomy Β· Vienna, Austria

ICRA 2026 Workshop Best Poster Award certificate

[A2] (International Level) Bronze Award of the 9th China International College Students Innovation Competition Apr 2024
Organizing Committee of China College Students "Internet+" Competition

[A3] (National Level) The First Prize of the 18th "Challenge Cup" National Undergraduate Curricular Academic Science and Technology Works Oct 2023
China Association for Science and Technology, Ministry of Education of the PRC, Chinese Academy of Social Sciences

[A4] (International Level) Bronze Award of the 8th China College Students "Internet+" Innovation and Entrepreneurship Competition Apr 2023
Organizing Committee of China College Students "Internet+" Competition

[A5] (National Level) Second Prize of the 15th China College Student Computer Design Competition Jul 2022
Organizing Committee of Chinese College Student Computer Design Competition

[A6] (National Level) Second Prize of the 15th Energy Saving and Emission Reduction Competition Aug 2022
China College Student Energy Conservation and Emission Reduction Social Practice and Technology Competition Committee

[A7] (National Level) Third Prize and Best Rookie Award of Formula Student Competition Oct 2022
China Society of Automotive Engineers, Organizing Committee of FSAC

πŸ—ΊοΈ Visitor Map

Click to view visitor map 🌍