[{"data":1,"prerenderedAt":110},["ShallowReactive",2],{"\u002Fen\u002Fglossary\u002Fembodied-ai":3},{"id":4,"title":5,"alternateName":6,"body":7,"description":100,"extension":101,"keywords":102,"meta":103,"navigation":104,"path":105,"seo":106,"stem":107,"updated":108,"__hash__":109},"glossary\u002Fglossary\u002Fen\u002Fembodied-ai.md","Embodied AI","具身智能 \u002F Embodied Intelligence",{"type":8,"value":9,"toc":94},"minimark",[10,15,22,27,40,44,86],[11,12,14],"h1",{"id":13},"what-is-embodied-ai","What Is Embodied AI?",[16,17,18,21],"p",{},[19,20,5],"strong",{}," (embodied intelligence) is the paradigm in which an intelligent agent has a physical body and accomplishes perception, decision-making, and action through interaction with the real environment. Unlike \"disembodied\" models that only process text or images, embodied AI must close the loop in the physical world: see → understand → act → observe the result → correct.",[23,24,26],"h2",{"id":25},"why-humanoids-are-the-primary-platform","Why Humanoids Are the Primary Platform",[16,28,29,30,33,34,39],{},"Human environments — stairs, door handles, tools, workstations — are designed for the human body. The humanoid form lets robots reuse this infrastructure ",[19,31,32],{},"without modifying the environment",", and lets vast human motion data (video, motion capture, ",[35,36,38],"a",{"href":37},"\u002Fen\u002Fglossary\u002Fteleoperation","teleoperation"," demonstrations) transfer directly into training data.",[23,41,43],{"id":42},"the-stack","The Stack",[45,46,47,64,70,80],"ul",{},[48,49,50,53,54,58,59,63],"li",{},[19,51,52],{},"Body",": high-dynamic hardware is the prerequisite — torque-dense ",[35,55,57],{"href":56},"\u002Fen\u002Fglossary\u002Fjoint-motor","joint motors",", high-frequency control buses, and whole-body ",[35,60,62],{"href":61},"\u002Fen\u002Fglossary\u002Fdegrees-of-freedom","degrees of freedom",";",[48,65,66,69],{},[19,67,68],{},"Perception",": multimodal sensing — vision, depth, IMU, touch;",[48,71,72,75,76,63],{},[19,73,74],{},"Decision",": foundation-model or reinforcement-learning policies, often transferred from simulation via ",[35,77,79],{"href":78},"\u002Fen\u002Fglossary\u002Fsim-to-real","Sim-to-Real",[48,81,82,85],{},[19,83,84],{},"Data",": teleoperated real-robot data collection for imitation learning.",[16,87,88,89,93],{},"The BXI ",[35,90,92],{"href":91},"\u002Fen\u002Frobots\u002Fhumanoid-robot","Elf 3 humanoid robot"," ships with a ROS2 SDK and MuJoCo simulation — an open platform built for embodied-AI research.",{"title":95,"searchDepth":96,"depth":96,"links":97},"",2,[98,99],{"id":25,"depth":96,"text":26},{"id":42,"depth":96,"text":43},"Embodied AI is the paradigm where an agent perceives, decides, and acts through a physical body interacting with the real world; humanoid robots are its primary platform, closing the perception–decision–action loop in reality.","md","embodied AI, embodied intelligence, humanoid robot, perception decision action, agents",{},true,"\u002Fglossary\u002Fen\u002Fembodied-ai",{"title":5,"description":100},"glossary\u002Fen\u002Fembodied-ai",null,"C6NRbR956x72lE_j5DM_QaS5e_lb5_lRYo0H3qL44lY",1783425867246]