Glossary

This page contains definitions for key terms used throughout the book.

A-D

Action (ROS 2)

A communication pattern for long-running, asynchronous tasks that provides feedback and can be canceled. It consists of an Action Client and an Action Server.

Action Planning

In robotics, the process of generating a sequence of low-level robot actions (e.g., "move arm," "close gripper") to achieve a desired high-level goal, often informed by an LLM.

Digital Twin

A virtual, dynamic representation of a physical object or system. In robotics, it serves as a simulation sandbox for development, testing, and validation before deploying to hardware.

Dynamics

The study of motion in relation to the physical forces that cause it, such as inertia, gravity, and friction. It helps determine the torques and forces required for a robot to perform a movement.

E-K

Embodied Intelligence

A theory in AI and robotics stating that an agent's intelligence is fundamentally shaped by the interaction between its physical body (morphology, sensors, actuators) and its environment.

End-Effector

The device at the end of a robotic arm, such as a gripper, tool, or hand, which interacts with the environment.

Forward Kinematics (FK)

The process of calculating the position and orientation of a robot's end-effector based on the known angles of its joints.

Gazebo

A powerful, open-source 3D robotics simulator that provides a realistic physics engine and sensor models. It is deeply integrated with the ROS ecosystem.

Inverse Kinematics (IK)

The process of calculating the required joint angles for a robot's end-effector to reach a desired position and orientation in space. This is often more complex than Forward Kinematics.

Isaac ROS

A collection of hardware-accelerated ROS 2 packages developed by NVIDIA that leverage GPUs to deliver significant performance improvements for common robotics tasks, particularly in perception and AI.

Isaac Sim

A powerful, extensible robotics simulation application by NVIDIA, built on Omniverse, designed for developing, testing, and managing AI-based robots with photorealistic environments and high-fidelity physics.

Kinematics

The study of motion without considering the forces or masses involved. It deals with the geometric relationships of the robot's structure, such as position, velocity, and acceleration.

L-R

Large Language Model (LLM)

A type of artificial intelligence model trained on vast amounts of text data, capable of understanding, generating, and reasoning with human language. Used in robotics for high-level planning and understanding natural language commands.

Middleware

Software that provides services and communication for applications beyond what the operating system offers. ROS 2 is a form of middleware for robotics.

Navigation2 (Nav2)

The standard ROS 2 framework for mobile robot navigation, providing a modular stack of algorithms for global and local path planning, control, and recovery behaviors.

Node (ROS 2)

A single, independent, executable program within a ROS 2 system that performs a specific task (e.g., controlling a motor, processing camera data). Nodes communicate with each other over the ROS 2 graph.

Physics Engine

A software component in a simulator that computes the effects of physical laws (gravity, collision, friction) on objects in the virtual world.

Plugin (Gazebo)

A shared library that can be attached to a simulated world, model, or sensor to extend its functionality, such as providing a ROS 2 interface.

Publisher (ROS 2)

A node that sends (publishes) messages on a specific Topic.

R-Z

rclpy

The official Python client library for programming in ROS 2.

Robot Operating System (ROS)

A flexible framework and ecosystem of tools, libraries, and conventions for building and running robot software. It simplifies the task of creating complex and robust robot behavior across a wide variety of robotic platforms.

SDF (Simulation Description Format)

An XML-based format used by Gazebo and other simulators to describe every aspect of a simulation, including robots, environments, lighting, and physics. It is a superset of URDF.

Service (ROS 2)

A synchronous, request/response communication pattern between two nodes. A client sends a request and waits for a response from a server.

Speech-to-Text (STT)

Technology that converts spoken language into written text, a critical component of Voice-to-Action (VLA) systems.

Subscriber (ROS 2)

A node that receives (subscribes to) messages from a specific Topic.

Synthetic Data

Artificially generated data used to train and test machine learning models. Simulators like Unity and Isaac Sim are often used to generate large, labeled datasets for perception tasks.

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 an asynchronous, many-to-many communication pattern ideal for continuous data streams.

Unity

A professional, cross-platform game engine known for its high-fidelity graphics and intuitive editor. It is increasingly used in robotics for creating photorealistic digital twins and generating synthetic data.

Vision-Language-Action (VLA) Systems

Robotics systems that integrate visual perception, natural language understanding (often via LLMs), and robot action capabilities to allow users to interact with robots using high-level language commands.

Visual Grounding

The process of linking abstract linguistic concepts (e.g., "the red cup") to concrete entities or features in a robot's visual perception of the environment.

Visual SLAM (VSLAM)

A method of Simultaneous Localization and Mapping (SLAM) that uses visual sensor data (e.g., cameras) to concurrently build a map of an unknown environment and localize the robot within that map.