The field of robot learning is experiencing remarkable advances driven by reinforcement learning, imitation learning, and the emergence of foundation models. Alongside the growing momentum of open-source software development in these areas, we are now witnessing an increasing number of open-source robotic hardware platforms being shared with the community—from quadrupeds to dual-arm manipulators and humanoids. This workshop aims to bring together researchers and developers who are actively involved in the design, development, and dissemination of open-source robotic hardware. The central goal is to foster an open, collaborative community where knowledge and experiences around hardware and its integration with robot learning can be shared across groups. We will structure the discussion around the following key questions and topics:

What kinds of actuators, sensors, and hardware designs are currently being used in open-source robot platforms?

What are the best practices for integrating open-source hardware with modern robot learning frameworks?

How should licensing and intellectual property considerations be handled in open-source hardware projects?

Which simulators and software environments are most compatible and useful for learning-oriented robotics research?

How can hardware developers communicate capabilities and limitations effectively to the learning community?

Are there sustainable models for monetization or community support in open-source hardware development?

This workshop will benefit the robot learning community by lowering the barrier to entry for experimenting with real-world robotics and by promoting the sharing of physical platforms in the same way open-source software has accelerated algorithmic innovation. The intended audience includes researchers and practitioners working on robot learning, robotics hardware development, open-source projects, and simulation environments. Presenters and panelists will be drawn from communities involved in open-source robot hardware development, reinforcement learning, imitation learning, foundation model-based control, and robotics systems integration. By building bridges between hardware developers and the robot learning community, this workshop seeks to catalyze new collaborations and accelerate innovation in embodied AI.

• Best Poster Award 1
  Very High Frequency Interpolation for Direct Torque Control
  Rafael Kourdis, Maciej Stępień, Jérôme Manhes, Nicolas Mansard, Steve Tonneau, Philippe Souères, Thomas Flayols
• Best Poster Award 2
  An Open-Source Platform for Dexterous Robotic Surgical Manipulation and Learning
  Jacinto Colan, Ana Davila, Yutaro Yamada, Yasuhisa Hasegawa
• Best Open-Source Hardware Award
  RL-based Giant Swing Motions on the Open-Source Robot MEVITA
  Ayumu Iwata, Kento Kawaharazuka, Keita Yoneda, Kei OKADA

• Switch4EAI: Leveraging Console Game Platform for Benchmarking Robotic Athletics

  Tianyu Li, Jeonghwan Kim, Wontaek Kim, DONGHOON BAEK, Seungeun Rho, Sehoon Ha

• RL-based Giant Swing Motions on the Open-Source Robot MEVITA

  Ayumu Iwata, Kento Kawaharazuka, Keita Yoneda, Kei OKADA

• Sanity: An Agile Brushless Quadrotor for Multi-Agent Experiments

  Heedo Woo, Kazi Ragib Ishraq Sanim, Keisuke Okumura, Guang Yang, Ajay Shankar, Amanda

• SwimGym: A Robotics Platform for Bio-Inspired Swimming Research

  Junzhe Hu, Jeong Hun Lee, Sofia Kwok, Guo Ning Sue, Carmel Majidi, Zachary Manchester

• Safe Robot Learning in Contact-Rich Manipulation via A Low-Cost Soft Gripper and Haptic Feedback Teleoperation

  Steven Oh, Tomoya Takahashi, Cristian Camilo Beltran-Hernandez, Yuki Kuroda, Masashi Hamaya

• KLEIYN: A Quadruped Robot Evolved from the Open-Source MEVIUS, Equipped with a Active Waist for Chimney Climbing

  Keita Yoneda, Kento Kawaharazuka, Temma Suzuki, Takahiro Hattori, Kei OKADA

• SOBIT LIGHT: An Accessible Open MoMa for data collection

  Keith Valentin Cardenas, Jiahao Sim, Yoshinobu Hagiwara, Yongwoon Choi

• Expert-Guided Imitation for Learning Humanoid Loco-Manipulation from Motion Capture

  Rohan Pratap Singh, Pierre-Alexandre Leziart, Masaki Murooka, Mitsuharu Morisawa, Eiichi Yoshida, Fumio Kanehiro

• RoboManipBaselines: A Unified Framework for Imitation Learning in Robotic Manipulation across Real and Simulated Environments

  Masaki Murooka, Tomohiro Motoda, Ryoichi Nakajo, Hanbit Oh, Koshi Makihara, Keisuke Shirai, Yukiyasu Domae

• An Open-Source Platform for Dexterous Robotic Surgical Manipulation and Learning

  Jacinto Colan, Ana Davila, Yutaro Yamada, Yasuhisa Hasegawa

• The GrandTour of Boxi

  Turcan Tuna, Jonas Frey, Lanke Frank Tarimo Fu, Maurice Fallon, Cesar Cadena, Marco Hutter

• CRS: An Open-Source, Low-Cost, and Modular Platform for Robot Learning Research

  Lukas Vogel, Marcus Aaltonen, Sabrina Bodmer, Rahel Rickenbach, Hao Ma, Michael Muehlebach, Melanie Zeilinger, Andrea Carron

• Deep Reinforcement Learning Controller for a Furuta Pendulum

  Maciej Stępień

• Very High Frequency Interpolation for Direct Torque Control

  Rafael Kourdis, Maciej Stępień, Jérôme Manhes, Nicolas Mansard, Steve Tonneau, Philippe Souères, Thomas Flayols

• Universal Low-Cost Force-Controlled Gripper for Learning Delicate Object Grasping

  Xuhui Kang, Tongxuan Tian, Sung-Wook Lee, Yen-Ling Kuo

• Mini TurtleBot: An Open-Source, Low-Cost, and Customizable Platform for Mobile Robotics Learning

  Hasan Shamim Shaon, Lucky Sah, Cecilia Tirkey, Bhanu Teja Giddaluru, Xiangxu Lin, Jong-Hoon Kim

• KAMI (Robot Head): An Open Source HRI Platform for Social Interaction and Embodied AI

  Irvin Steve Cardenas, John Edward Naulty, George Digkas, Kostas Kryptos Chalkias, Jong-Hoon Kim