Xintong (Ian) Yang
Research Associate in robotic manipulation at Cardiff University.
Xintong (Ian) Yang
Research Associate (Postdoc) · Robotic Manipulation · Guided Agentic Engineering · Cardiff University
I develop robotic systems that manipulate rigid, deformable, and granular objects using learning-based and physics-based methods. My research combines guided agentic engineering, differentiable simulation, system identification, reinforcement learning, and affordance modelling to build capable and trustworthy robots for real-world settings.
Research Projects & Papers · Google Scholar · LinkedIn · Zhihu · Email
About
I am a research associate in the School of Engineering at Cardiff University. My current work spans robotic manipulation of deformable and granular materials, differentiable physics-based modelling, guided agentic engineering, and automation for robotic laboratory systems.
I completed my PhD at Cardiff University in October 2023, where I studied hierarchical reinforcement learning and affordance learning for robotic manipulation. Before that, I received my Bachelor’s and Master’s degrees in Mechanical and Industrial Engineering from Guangdong University of Technology, China.
I am also committed to open-source research software and reproducible robotics.
Research Interests
- Guided agentic engineering for real-world robot autonomy
- Deformable and granular material manipulation
- Differentiable physics and system identification
- Reinforcement learning for long-horizon manipulation
- Affordance-guided perception and action
- Reliable real-world robotic systems for human-facing applications
News
- [2025 Nov.] DDBot was published in IEEE Transactions on Robotics. Preprint · Video
- [2025 Oct.] Celebi’s Choice was published at IROS 2025. DOI
- [2025 Feb.] Our work on differentiable physics-based system identification for elastoplastic materials was accepted by IJRR. Preprint · Video · Project page
- [2025 Jan.] I joined a new BBSRC-funded project at Cardiff University on robotic and automated biology laboratories. Grant record
- [2025 Jan.] The AutomaChef preprint became available online. Preprint
- [2024 Jan.] The GAM preprint and video demo became available online. Preprint · Video
Research Grant/Involvements
- [2023-2025] PHYDL: Physics-Guided Differentiable Learning for Robotic Manipulation of Deformable and Granular Materials
- Role: Research Associate
- PI: Ze Ji
- Funder: EPSRC New Horizon
- [2025-now] Almond: Autonomous Laboratory for Molecular Discovery
- Role: Research Associate
- PI: Oliver Castell; Co-I: Ze Ji, Jin Li
- Funder: BBSRC
Selected Publications
- DDBot: Differentiable Physics-Based Digging Robot for Unknown Granular Materials. IEEE Transactions on Robotics, 2025. Preprint · Video
- Differentiable Physics-based System Identification for Robotic Manipulation of Elastoplastic Materials. The International Journal of Robotics Research, 2025. Preprint · Video · Project Page
- Celebi’s Choice: Causality-Guided Skill Optimisation for Granular Manipulation via Differentiable Simulation. IROS, 2025. DOI
- GAM: General Affordance-Based Manipulation for Contact-Rich Object Disentangling Tasks. Neurocomputing, 2024. Preprint · Video
- Recent Advances of Deep Robotic Affordance Learning: A Reinforcement Learning Perspective. IEEE Transactions on Cognitive and Developmental Systems, 2023. Preprint
- Hierarchical Reinforcement Learning with Universal Policies for Multi-Step Robotic Manipulation. IEEE Transactions on Neural Networks and Learning Systems, 2021. Paper · Video · Code
For a fuller project-oriented overview, visit the Research Projects & Papers page. For the complete publication list, see Google Scholar.
Open-Source & Resources
- DDBot-IEEE-TRO-2025 — official code and assets for DDBot
- SI4RP-data — project page and resources for differentiable physics-based system identification of elastoplastic materials
- pybullet_multigoal_gym — open-source robotic manipulation benchmarks on PyBullet
- GAM-paper-codes — code for GAM (Neurocomputing 2024)
- A-2-paper-code — code for A²E (ICAC 2022)
- UOF-paper-code — code accompanying hierarchical reinforcement learning for multi-step manipulation
- DRL_Implementation — deep RL algorithm implementations
- Kuka_Zivid_ros_melodic — KUKA + Zivid ROS setup (Melodic) for Vision-Based Grasping and Manipulation
- UR5e_Zivid_ros_noetic — UR5e + Zivid ROS setup (Noetic) for Vision-Based Grasping and Manipulation
Collaborators
- Ze Ji - Reader, Cardiff University
- Collaboration, Mentor, PhD Supervisor, Co-supervision, Project PI (PHYDL), Project Co-I (Almond)
- Robotics and Autonomous Systems, SLAM, AI for Robotics, Computer Vision
- Yu-Kun Lai — Professor, Cardiff University
- Collaboration, PhD Supervisor, Co-supervision
- 3D Geometry, Computer Graphics, AI
- Oliver Castell - Reader, Cardiff University
- Collaboration, Project PI (Almond)
- Synthetic biology, Single Molecule Science, Artificial Cells
- Jin Li - Lecturer, Cardiff University
- Collaboration, Project Co-I (Almond)
- Microfluidics and nanofluidics, Synthetic biology
- Jing Wu — Collaboration, PhD Supervisor, Cardiff University
- Collaboration, PhD Supervisor, Co-supervision
- Computer Vision, AI
- Seyed Amir Tafrishi — Lecturer, Cardiff University
- Collaboration
- Control engineering, mechatronics and robotics
Student Supervision
- Minglun Wei — Co-supervised, PhD student, Cardiff University
- On-going co-supervision
- Deformable Object Manipulation
- Boliang Cai — Co-supervised, Reserach Associate, Loughborough University
- Past co-supervision, Paper collaboration
- Autonomous Vihecles