Glossary
A list of key terms and acronyms used throughout this textbook.
A
Action (ROS 2) : A communication pattern for long-running, asynchronous tasks that provides feedback during execution and can be cancelled. Used for tasks like navigation or manipulation.
Artificial Intelligence (AI) : A broad field of computer science focused on creating machines that can perform tasks that typically require human intelligence, such as learning, problem-solving, and perception.
C
Colcon : The standard build tool for ROS 2, used to compile and install packages in a workspace.
D
Digital Twin : A virtual model of a physical object, system, or process. In robotics, simulators like Gazebo and Isaac Sim are used to create digital twins of robots and their environments.
Domain Randomization : A technique used in simulation to improve sim-to-real transfer. It involves randomly varying simulation parameters like lighting, textures, and object positions during the training of an AI model.
E
Embodied AI : A field of AI focused on creating intelligent agents that are situated in a physical or virtual body and can interact with their environment.
G
Gazebo : A powerful, open-source 3D robotics simulator that is widely used in the ROS community for testing and development.
H
Hardware-in-the-Loop (HIL) : A simulation technique where real hardware components (like motor controllers) are connected to a virtual simulation, allowing for testing of low-level hardware and firmware.
I
Isaac Gym : An NVIDIA toolkit for creating high-performance reinforcement learning environments. It leverages the GPU to run thousands of simulations in parallel.
Isaac ROS : A collection of hardware-accelerated ROS 2 packages from NVIDIA, optimized for NVIDIA GPUs and the Jetson platform.
Isaac Sim : A photorealistic, physically-accurate robotics simulator from NVIDIA, built on the Omniverse platform.
J
Joint : In robotics, a component that connects two links and allows for motion between them (e.g., revolute, prismatic).
L
Launch File (ROS 2) : An executable script (in Python or XML) used to start and configure a collection of ROS 2 nodes.
Link : A rigid body component of a robot that is connected to other links by joints.
LLM (Large Language Model) : A type of neural network with billions of parameters, trained on vast amounts of text data, capable of understanding and generating human-like language.
M
Machine Learning (ML) : A subfield of AI that focuses on building algorithms that can learn from data to make predictions or decisions.
Message (ROS 2) : A data structure used for communication in ROS 2. Topics, services, and actions all use messages to exchange information.
N
Node (ROS 2) : The fundamental processing unit in a ROS 2 system. A node is a program that performs a specific task.
NVIDIA Omniverse : A real-time 3D collaboration and simulation platform developed by NVIDIA, which serves as the foundation for Isaac Sim.
P
Package (ROS 2) : The primary unit for organizing software in ROS 2. A package may contain nodes, libraries, configuration files, and launch files.
Parameter (ROS 2) : A configurable setting for a ROS 2 node that can be changed at runtime.
Publisher (ROS 2) : A node that sends data by publishing messages to a topic.
R
Reinforcement Learning (RL) : A type of machine learning where an agent learns to make decisions by performing actions in an environment to maximize a cumulative reward.
ROS (Robot Operating System) : A flexible framework for writing robot software. It is a collection of tools, libraries, and conventions that aim to simplify the task of creating complex and robust robot behavior.
S
SDF (Simulation Description Format) : An XML-based format used by Gazebo and other simulators to describe the properties of robots, objects, and environments.
Service (ROS 2) : A communication pattern for synchronous, one-to-one request/reply interactions.
Sim-to-Real Transfer : The process of transferring knowledge or a policy learned in a simulation to a physical, real-world robot. Bridging the "reality gap" is a major challenge in this area.
Simulation : The use of a computer program to model the behavior of a physical system. In robotics, this is used for testing, development, and training.
Subscriber (ROS 2) : A node that receives data by subscribing to a topic.
T
Topic (ROS 2) : A named bus for messages. Nodes can publish messages to a topic or subscribe to a topic to receive messages. It is used for asynchronous, one-to-many communication.
U
URDF (Unified Robot Description Format) : An XML-based format used in ROS to describe the physical structure of a robot, including its links, joints, and sensors.
V
Vision Transformer (ViT) : A type of neural network architecture based on the transformer model, adapted for computer vision tasks. It treats an image as a sequence of patches, similar to how a language model treats a sentence as a sequence of words.
VLA (Vision-Language-Action) Model : A single, end-to-end neural network that takes visual and language inputs and outputs actions for a robot to execute.
W
Workspace (ROS 2) : A directory containing one or more ROS 2 packages, along with build and install artifacts.