China Merchants Lab Unveils LiOS Infrastructure for Robotic Cloth Folding
量子位 · wechat · 2026-08-20
China Merchants Lion Mountain AI Lab showcased a flexible cloth folding solution powered by its self-developed LiOS end-cloud collaborative infrastructure at the 2026 World Robot Conference. The technology addresses the sim-to-real gap via a closed-loop data iteration system covering training, deployment, and Real2Sim teleoperation, boosting training throughput by 5x and evaluation efficiency by 4x.
The lab previously won the ICRA 2026 LeHome Challenge using its proprietary VLA model. The on-site demonstration featured three core capabilities: multi-platform folding, autonomous identification and folding of different garment types, and large-deformation handling. The LiOS architecture comprises a cloud side (supporting 10B+ parameter models), an edge side (real-time execution), and end-cloud collaboration (low-latency video transmission), effectively integrating cloud computing with physical robots.
More from Embodied
- Robotics in-context learning sparks debate over definition of gradient updates — NielsRogge · 2026-08-20
- Robotic 3D printing can now build complex multi-directional parts that are stronger and lighter — AryHHAry · 2026-08-20
- Trail Mate: An offline-first handheld GPS system over LoRa network — tom_doerr · 2026-08-20
- Maker builds a portable body for AI companion Mythos, takes it to hear wind and waves — repligate · 2026-08-20
- Matic Robot Adds Voice and Gesture Control: 'Clean This' Works in 75 Languages — lukas_m_ziegler · 2026-08-20
- SoftBank Leads $200M Round for Gravis to Automate Excavators — lukas_m_ziegler · 2026-08-20