This video demonstrates the "cerebellum common platform" co-developed by ITRI and NexCOBOT.

Deployed with a VLA (Vision-Language-Action) model inference system, the platform integrates visual input from 4 cameras, joint motor feedback, and natural language instructions to execute high-precision dual-arm pick-and-place and cross-arm collaboration tasks.

[Key Features]
● Multimodal Perception Integration: Combines 4-camera vision, joint motor feedback, and natural language command inputs.
● Real-Time VLA Model Inference: Accurately interprets instructions to execute sorting tasks (e.g., placing yellow cups into the brown bin and green cups into the white bin).
● Dual-Arm Collaboration & Hand-to-Hand Transfer: Dynamically transfers objects from the left hand to the right hand when items are positioned on the left side.
● Industry Collaboration: Co-developed by ITRI's engineering team and NexCOBOT on the cerebellum common platform.

Official Website: https://www.nexcobot.com/