Glossary
This glossary defines key terms used throughout the Physical AI & Humanoid Robotics book. It will be maintained and extended as modules are written.
A
Action (ROS 2): A communication pattern in ROS 2 for long-running tasks with feedback. Actions combine the request-response pattern of services with the continuous data stream of topics.
ASR (Automatic Speech Recognition): Technology that converts spoken language into text. Examples include OpenAI Whisper and Google Speech-to-Text.
D
Digital Twin: A virtual representation of a physical robot or system that mirrors its behavior in simulation. Used for development, testing, and training before deployment.
G
Gazebo: An open-source physics simulation engine for robotics. Supports realistic physics, sensor simulation, and robot modeling via URDF/SDF.
H
Humanoid Robot: A robot designed to resemble and move like a human, typically with a torso, head, two arms, and two legs. Examples include Unitree G1, Boston Dynamics Atlas.
I
Isaac ROS: NVIDIA's ROS 2 packages for perception, navigation, and manipulation. Includes VSLAM, object detection, and other AI-powered pipelines.
Isaac Sim: NVIDIA's photorealistic robotics simulator built on Omniverse. Provides high-fidelity physics, graphics, and synthetic data generation.
L
LLM (Large Language Model): AI models trained on vast text corpora that can understand and generate natural language. Examples include GPT-4, Claude, LLaMA.
N
Nav2: The ROS 2 navigation stack for mobile robots. Provides path planning, obstacle avoidance, and localization capabilities.
Node (ROS 2): A process in ROS 2 that performs computation. Nodes communicate via topics, services, and actions.
P
Physical AI: The intersection of artificial intelligence and embodied systems (robots) that exist in and interact with the physical world. Contrasts with "pure digital" AI that operates only in computational spaces.
R
ROS 2 (Robot Operating System 2): The second generation of ROS, a middleware framework for building distributed robot systems. Provides communication, tools, and libraries for robotics development.
ros2_control: A ROS 2 framework for managing robot hardware interfaces, controllers, and real-time control loops.
S
SDF (Simulation Description Format): An XML format for describing robots, sensors, and environments in Gazebo. More feature-rich than URDF.
Service (ROS 2): A request-response communication pattern in ROS 2. A client sends a request and waits for a response from a server.
Sim-to-Real Transfer: The process of training AI models or developing behaviors in simulation and successfully deploying them to physical robots.
T
TF2 (Transform Library): A ROS 2 library for tracking coordinate frame relationships over time. Essential for robot kinematics and sensor fusion.
Topic (ROS 2): A named communication channel in ROS 2 where nodes publish and subscribe to messages. Uses a publish-subscribe pattern.
U
URDF (Unified Robot Description Format): An XML format for describing robot structure, joints, and links. Used by ROS 2 and Gazebo for robot modeling.
V
VLA (Vision-Language-Action): A robotics paradigm that combines computer vision, natural language understanding, and robot action execution. Enables robots to follow natural language commands.
VSLAM (Visual Simultaneous Localization and Mapping): A technique that uses camera data to simultaneously build a map of the environment and determine the robot's position within it.
This glossary is a living document and will be updated as new concepts are introduced throughout the book.