This workspace is a curated set of concise notes and mini-tutorials covering the full workflow of building intelligent robots with ROS 2.
It follows a natural progression:
- Foundations of ROS 2: core ideas such as DDS, workspaces, packages, build tools (
colcon), client libraries (rclcpp,rclpy), node types, and QoS. - Communication patterns: topics, services, actions, parameters, launch files, and DDS configuration, so you can structure data flow and runtime behavior of your system.
- Robot description and control: modeling robots with URDF/Xacro, defining links and joints, and driving hardware or simulated joints using the
ros2_controlframework. - Simulation and navigation: using Gazebo Sim for rich physics-based simulation, and the Nav2 stack combined with Behavior Trees to achieve autonomous navigation in realistic environments.
- Perception and advanced intelligence: camera calibration, depth sensing, reinforcement learning with Isaac Sim / Isaac Lab, and LLM-based ROS 2 agents for natural-language-driven robot behaviors.
You can treat these notes as:
- A learning path: read from foundational ROS 2 concepts, through communication and control, up to navigation, perception, RL, and LLM agents.
- A reference: open the relevant markdown file when you implement a given subsystem (e.g., topics, Nav2,
ros2_control, Gazebo, RL, or LLM agents).
本工作区整理了一组围绕 ROS 2 机器人系统完整工作流 的精简笔记和小教程，方便系统化学习与快速查阅。
整体内容大致按以下逻辑展开：
- ROS 2 基础：包括 DDS 通信模型、工作空间与包结构、
colcon编译、客户端库（rclcpp/rclpy）、节点类型以及 QoS 等核心概念。 - 通信模式：围绕 Topic、Service、Action、参数、Launch 文件以及 DDS 配置，帮助你搭建节点间的数据流和运行时行为。
- 机器人建模与控制：使用 URDF/Xacro 描述机器人（Link 与 Joint），并通过
ros2_control框架对真实或仿真关节进行控制。 - 仿真与导航：利用 Gazebo Sim 进行物理仿真，结合 Nav2 导航栈与行为树，实现复杂环境中的自主导航和恢复行为。
- 感知与智能：涵盖相机标定、深度传感器、基于 Isaac Sim / Isaac Lab 的强化学习训练与部署，以及基于大模型的 ROS 2 智能体，实现自然语言交互和高层决策。
你可以这样使用这些笔记：
- 作为学习路线：从基础概念开始，依次深入通信、建模与控制、仿真与导航，再到感知、强化学习和 LLM 智能体。
- 作为速查手册：在开发具体子系统（如 Topic、Nav2、
ros2_control、Gazebo、强化学习或 LLM Agent）时，直接打开对应的 Markdown 文件查阅细节。