Embodied AI Glossary中文

Embodied AI Glossary

2926 terms14 categoriesFacts as of 2026-09-30

A glossary for people new to embodied AI. The 14 categories follow a learning path: first the robot body (form factors, parts, motion, control, perception), then tools and simulation, then data, training and models, and finally the companies and the industry's vocabulary. Each category is split into sections that go from basic to advanced. Only the 240 Essential entries are shown by default; search covers everything.

Loading entries…

Browse by category

  1. 1 Core Concepts & Tasks (150)
    Start with the map: what embodied AI does, the tasks robots perform, and what terms like ‘generalization’ and ‘cross-embodiment’ mean.
  2. 2 Robot Types & Products (179)
    Meet the robots: body types from robot arms to humanoids, and the notable models you’ll see on the market.
  3. 3 Hardware & Body Parts (195)
    Take the robot apart: what motors, gearboxes, lead screws, dexterous hands, and compute platforms each actually do.
  4. 4 Mechanics & Kinematics (168)
    The physics behind robot motion: representing pose, forward and inverse kinematics, Jacobians, dynamics, and balance.
  5. 5 Control & Planning (203)
    Making joints move the way you want: from PID to force control, MPC, motion planning, and whole-body control.
  6. 6 Perception & Sensors (307)
    How robots see and feel: cameras, depth, IMUs, force and touch, plus point clouds, calibration, and SLAM.
  7. 7 Software & Tooling (232)
    The software that ties body, control, and perception together: ROS 2, URDF, motion libraries, learning frameworks, and deployment tools.
  8. 8 Simulation & Evaluation (210)
    Practicing and testing on a computer: how simulators work, the sim-to-real gap, and the benchmarks used to evaluate.
  9. 9 Data & Collection (177)
    The raw material for learning: how demonstration data is collected, what datasets exist, and how data is processed and scaled.
  10. 10 Training & Learning Methods (203)
    How models are actually trained: imitation learning, reinforcement learning, pretraining and fine-tuning, plus tricks that make it more stable.
  11. 11 Models & Architectures (184)
    What the resulting models look like: Transformers, VLMs, how actions are generated, VLAs, and world models.
  12. 12 Landmark Models & Projects (332)
    Famous models in historical order: from SayCan and RT-2 to the π series, GR00T, and world models.
  13. 13 Companies & Institutions (237)
    Who’s actually working on embodied AI: tech giants, hardware and model companies worldwide, component makers, and research institutions.
  14. 14 Industry Jargon & Business (149)
    Making sense of press releases, pitch decks, and group-chat slang: embodiment, ‘big/small brain,’ data flywheels, mass production, and more.