Robotic assembly is crucial across many industries, from automotive and aerospace to electronics, medical devices, and even everyday items like furniture and appliances. However, it remains one of the toughest challenges in robot learning due to the combined challenges of accurate perception, multi-modal sensing, affordance reasoning, long-horizon task and motion planning, grasping, contact-rich and dexterous manipulation, physical simulation, sim2real transfer, and generalization over arbitrary objects with complex shapes. Our workshop aims to gather experts to discuss how robot learning can tackle these challenges. Key topics include:

• Assembly planning: Learning to plan and optimize for safe, efficient, and feasible assembly sequences.
• Assembly skill learning: Learning skills such as grasping, in-hand manipulation, insertion, screwing, and tool usage to solve contact-rich and high-accuracy assembly tasks.
• Foundation models: Leveraging foundation models for generalizable perception, planning, and control for assembly.
• Multi-modal sensing: Integrating vision, force, and tactile sensing for improved precision and reliability for assembly.
• Datasets and benchmarks: New datasets or benchmarks on assembly tasks featuring task and object diversity, real-world complexity, multiple modalities and robot embodiments.
• Human-robot interaction: Learning assembly from demonstrations or language instructions, or collaborating with humans.
• Industry-level assembly: Filling the gap of employing robot learning to solve real-factory assembly problems with scale.

• Dynamics-Guided Diffusion Model for Sensor-less Robot Manipulator Design [Paper] [Video]

  Xiaomeng Xu, Huy Ha, Shuran Song

• Local Policies Enable Zero-shot Long-horizon Manipulation [Paper] [Video]

  Murtaza Dalal, Min Liu, Walter Talbott, Chen Chen, Deepak Pathak, Jian Zhang, Russ Salakhutdinov

• From Imitation to Refinement - Residual RL for Precise Visual Assembly [Paper]

  Lars Lien Ankile, Anthony Simeonov, Idan Shenfeld, Marcel Torne Villasevil, Pulkit Agrawal

• Force-Aware ProDMP with Event-Based Replanning [Paper] [Video]

  Paul Werner Loedige, Maximilian Xiling Li, Rudolf Lioutikov

• Specialist Assembly Policies over Diverse Geometries [Paper] [Video]

  Bingjie Tang, Iretiayo Akinola, Jie Xu, Bowen Wen, Ankur Handa, Karl Van Wyk, Dieter Fox, Gaurav S. Sukhatme, Fabio Ramos, Yashraj Narang

• Affordance-Centric Policy Learning: Generalisable Robot Policy Learning using Affordance-Centric Task Frames [Paper]

  Krishan Rana, Jad Abou-Chakra, Sourav Garg, Robert Lee, Ian Reid, Niko Suenderhauf

• [Best Paper] FORGE: Force-Guided Exploration for Robust Contact-Rich Manipulation under Uncertainty [Paper] [Video]

  Michael Noseworthy, Bingjie Tang, Bowen Wen, Ankur Handa, Chad Kessens, Nicholas Roy, Dieter Fox, Fabio Ramos, Yashraj Narang, Iretiayo Akinola

• Image-Based Pose Estimation of Sub-Centimeter Industrial Parts for Automated Assembly [Paper]

  Hameed Shaheed Abdul-Rashid, Yangfei Dai, Holly Dinkel, Minh Quang Ta, Tan Chen, Junyi Geng, Tim Bretl

• Blox-Net: Generative Design-for-Robot-Assembly Using VLM Supervision, Physics Simulation, and a Robot with Reset [Paper] [Video]

  Andrew Goldberg, Kavish Kondap, Tianshuang Qiu, Zehan Ma, Letian Fu, Justin Kerr, Huang Huang, Kaiyuan Chen, Kuan Fang, Ken Goldberg