L12: Namespaces, Remapping, and Behavior Trees

Overview

This lecture covers three topics that are essential for building production-quality ROS 2 systems. Namespaces isolate multiple instances of the same node by prefixing all their topics, services, and parameters. Remapping lets you redirect individual names (node, topic, service) at runtime without modifying source code. Behavior trees organize robot behaviors into a modular, hierarchical structure where leaf nodes (actions and conditions) do the work and composite nodes (Sequences and Selectors) make the decisions. All hands-on examples use the namespace_demo, remapping_demo, and bt_demo packages.

Learning Objectives

By the end of this lecture, you will be able to:

• Apply namespaces to isolate multiple instances of the same node.

• Remap node names, topic names, and parameters at runtime and from launch files.

• Explain the structure of a behavior tree: nodes, ticks, and return statuses.

• Implement a simple robot behavior tree in Python using py_trees and py_trees_ros.

• Drive a robot to a fixed goal in Gazebo using a BT with a condition and an action node.

Next Steps

• In the next lecture, we will cover Mapping and Navigation with Nav2:

  • Occupancy grid maps and the map frame

  • SLAM with slam_toolbox: building a map from LiDAR and odometry

  • Localization with AMCL against a saved map

  • The Nav2 stack: planner, controller, behavior tree, recovery

  • Sending navigation goals via the NavigateToPose action

• Complete the exercises from this lecture before the next class.

• Read Nav2 Getting Started and skim Nav2 Concepts.