How can robots learn from humans?
Peer-Reviewed Publication
Updates every hour. Last Updated: 16-Jan-2026 21:11 ET (17-Jan-2026 02:11 GMT/UTC)
Can robots grasp diverse objects adaptively like humans? Published in National Science Review, researchers from Tsinghua University reports a human-taught sensory-control synergy approach that transfers human grasping experience to robots. Emulating brain’s neural cognition, multimodal tactile signals captured with tactile glove are encoded into semantic grasping states, then converted into grasping actions with experience-based fuzzy controller. The approach enables robots dexterously grasping like humans.
Researchers have reported a zero-dimensional metal halide that switches its fluorescence ON through two independent pathways: pressure induces a structural transition to bright orange emission, while DMF solvent exposure activates intense yellow luminescence with 97% quantum yield. The solvent-driven “ON” state is fully reversible by mild heating, offering excellent cycling stability. This dual-stimuli response enables high-sensitivity DMF detection, advanced optical encryption, and reconfigurable logic operations.