The company focuses on the R&D of a new generation of tactile-enabled embodied large models, deeply integrating tactile information into its core technical roadmap.
Unlike conventional embodied models that rely on vision alone, we fuse vision, touch and force end to end, breaking the bottleneck of robots that can only see but cannot feel. This enables the leap from passively executing commands to actively understanding the world, giving robots autonomous cognition of the physical world, reasoning about object properties and interaction logic, and autonomous task planning and execution in complex dynamic scenes.
The model will provide new intelligent decision-making solutions for industrial precision assembly, compliant manipulation by service robots and the automation of scientific experiments.