ICRA 2026 in Vienna, Austria
Our team recently returned from ICRA 2026 in Vienna, Austria, where we had an incredible opportunity to engage with global robotics experts and share our work with the community.
Jinhee Yun presented Dr. Jeongmin Lee’s work on “Variations of Augmented Lagrangian for Robotic Multi-Contact Simulation,” which was also published at IEEE Transactions on Robotics, 2025 (Online, Youtube).
Doyoon Kong presented Dr. Minji Lee’s work on “Narrow Passage Path Planning via Homotopy-Preserving Collision Constraint Interpolation,” which was also published at IEEE Transactions on Robotics, 2025 (Online, Youtube).
Hyelim Choi presented her work on “Robust Hand Tracking from Visual-Inertial Fusion” (pdf).
Jinhee Yun presented her work on “Amortized NeuralSDF-Mesh Collision Detection for Robotic Contact Simulation” (pdf).
Yejun Song presented his work on “Differentiable Optimization-Based Modular Planning Framework for Pick-and-Place with Regrasp” (pdf).
mình có lần lướt đọc mấy trao đổi trên mạng شيخ روحاني thì thấy nhắc nên cũng tò mò mở ra xem thử cho biết. mình không tìm hiểu sâu chỉ xem qua trong thời gian ngắn để quan sát bố cục cách sắp xếp جلب الحبيب các mục và trình bày nội جلب الحبيب dung tổng thể. cảm giác là các phần được trình bày khá gọn, các hurentest berlin mục rõ ràng Huren Berlin nên đọc lướt cũng không bị rối với mình như vậy là đủ để nắm tin cơ bản rồi. جلب الحبيب
👍
Jinhee Yun's work on augmented Lagrangian for multi-contact simulation caught my eye since that's close to my own research. I also looked up the Deep Dish Dungeon Wiki & Database to learn more about the team behind these papers. Congrats to everyone who presented in Vienna. Found Deep Dish Dungeon Wiki & Database helpful on this too.
Researchers are moving past rigid, pre-programmed motion planning into generalizable neural policies that allow robotic arms and dual-arm systems to handle delicate, unfamiliar objects in real-world environments without failing. A robotics graduate student in an online automation circle over on MaanWin brought this exact advancement up yesterday, noting how live demonstrations of zero-shot robotic task planning in the exhibition halls proved that simulated reinforcement learning is translating into real-world hardware faster than ever.