吳璠

吳璠 WU Fan

Postdoc at TUM - Munich Institute of Robotics and Machine Intelligence

Munich Institute of Robotics and Machine Intelligence, Technical University of Munich

Biography

I am currently a research scientist (postdoc) at Munich Institute of Robotics and Machine Intelligence (MIRMI) and the Chair of Robotics and Systems Intelligence (RSI) with Prof. Sami Haddadin. Before joining MIRMI as a PostDoc, I did my PhD at King’s College London, where I focused on energy efficient compliant actuation. Prior to that, I studied Applied Mathematics and worked as a quantitative trader in stock and futures market. I have been a fan of many Sci-Fi works featuring futuristic robot technologies, from giant robots (like Gundam and Pacific Rim) to cyborgs (eg Ghost in the Shell), so I decided to participate in the exiciting journey towards making great robots.

My research interests include robot skill learning, behavior trees, foundation models for robotics, robotic system architecture, and causality-aware autonomous decision making. My current focus is on robot skills development, acquisition and learning. I am leading a team working on large-scale collective robot learning and explore robot skill learning by leveraging scaling in physical robot experiments.

On the practical side, my goal is to build a robot skills and behavior coordination system that can be deployed in real flexible manufacturing use cases. Moreover, my dream is to make robots achieve human-level performance in tool-use and craftsmanship. On the theorectical side, I want to understand how intelligent species conceptualize the world and build their decision making system based on fundamental conpects about objects, actions and causes-and-effects.

Interests
  • Robot skill learning and acquisition for manipulation, tool-use, and craftsmanship
  • behaviour-trees for reactive manipulation skills and task planning
  • Causality-aware autonomous decision making
Education
  • PhD in Robotics, 2016-2020

    King’s College London, UK

  • MSc in Financial Mathematics, 2012-2013

    King's College London, UK

  • BSc in Applied Mathematics, 2008-2012

    University of Science and Technology Beijing, China

Recent News

[Jun. 2023][paper] 5/6 papers I co-authored/mentored are accepted to IROS 2024.

[May. 2024][paper][event] I had a wonderful ICRA week in Yokohama, Japan. Our recent progess on Collectie Learning was reported by Prof. Sami Haddadin in his Plenary Talk: The Great Robot Accelerator: Collective Learning of Optimal Embodied AI.

[Apr. 2024][paper] Our work on behavior tree generation using LLMs has been accepted to be presented at ICRA24 Workshop Co-manipulation. The main author, my master student AO Jicong is finishing his master thesis on this topic.

[Jan. 2024][paper] Three papers I co-authored/mentored got accepted to ICRA 2024 and one RAM paper will be presented there too! Check it here for the papers!

[Oct. 2023][paper] Our work titled Digital Robot Judge (DR.J): Building a Task-Centric Performance Database of Real-World Manipulation with Electronic Task Boards is accepted by IEEE Robotics and Automation Magazine (RAM) and will be published soon.

[Oct. 2023][paper] Our work on Tactile Manipulation Using Unified Force-Impedance Control is accepted and will be presented at CDC 2023 in Singapore.

[Oct. 2023][Event] I am co-organizing the Workshop: Robotics and AI in Future Factory at IROS 2023, find more info here!

[Oct. 2023][paper] Our work on deformable object manipulation is presented in IROS 2023, check it here.

[Sep. 2023][paper] Our colleague Kim Pepper presented our work Individualized Training of Back Muscles Using Iterative Learning Control of a Compliant Balance Board at ICORR 2023 in Singapore.

[Jun. 2023][Event] We successfully organized the 3rd Robothon Grand Challenge in automatica!

[Jun. 2023][Event] We are showing multiple demos at Automatica trade fair.

[Jul. 2023][paper] We have 3 papers presented in IFAC 2023. Check it here for the papers!

Demo Gallery

One block in the collective learning system.
One block in the collective learning system.
Collective Learning - Parallel and Distributed Robot Learning System.
Collective Learning - Parallel and Distributed Robot Learning System.
Safe tactile manipulation using unified force-impedance control.
Safe tactile manipulation using unified force-impedance control.
Robot assembling a gearbox using multiple tools.
Robot assembling a gearbox using multiple tools.
Robotic assembly planning.
Robotic assembly planning.
Cable routing and fixing with contact detection and error handling.
Cable routing and fixing with contact detection and error handling.

Recent Publications

Quickly discover relevant content by filtering publications.
A Scalable Platform for Robot Learning and Physical Skill Data Collection
Accepted to IROS 2024
Elliptical K-Nearest Neighbors - Path Optimization via Coulomb's Law and Invalid Vertices in C-Space Obstacles
Accepted to IROS 2024
Flexible Informed Trees (FIT*): Adaptive Batch-Size Approach in Informed Sampling-Based Path Planning
Accepted to IROS 2024
Ontology Based AI Planning and Scheduling for Robotic Assembly
Accepted to IROS 2024
Learning Task Planning from Multi-Modal Demonstration for Multi-Stage Contact-Rich Manipulation
TacDiffusion: Force-domain Diffusion Policy for Precise Tactile Manipulation
Tactile-Morph Skills: Energy-Based Control Meets Data-Driven Learning
1 kHz Behavior Tree for Self-adaptable Tactile Insertion
2024 IEEE International Conference on Robotics and Automation (ICRA).
1 kHz Behavior Tree for Self-adaptable Tactile Insertion

Recent Posts

Experience

 
 
 
 
 

Senior Scientist at Munich Institute of Robotics and Machine Intelligence (MIRMI) Postdoc at Chair of Robotics and Systems Intelligence (RSI) with Prof. Dr. Ing. Sami Haddadin.

My current focus is robot skill development, acquisition and learning within the KI.Fabrik (AI Factory) project (https://kifabrik.mirmi.tum.de).

I am also supporting management or research in these projects:

 
 
 
 
 
King's College London | Robot Learning Lab
PhD candidate
January 2016 – February 2020 London, UK

Teaching Assistant for Teaching Assistant for several courses including Robotic Systems, Introduction to Robotic, Real-time Systems.

Research Assistant for side projects.

Accomplish­ments

Coursera
Deep Learning Specialization
See certificate

Contact

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nam mi diam, venenatis ut magna et, vehicula efficitur enim.