L11: Coordinate Frames, TF2, and Mobile Robot Control

Overview

This lecture covers pose representation (position and orientation using Euler angles and quaternions), coordinate frames and the TF2 transform library, and mobile robot control. You will learn how orientations are encoded in ROS 2 using quaternions, why gimbal lock makes Euler angles unsuitable for continuous rotation tracking, and how TF2 manages the system-wide tree of coordinate frames that every robotic application depends on. On the control side, you will drive a differential-drive robot to a goal pose using a proportional controller on cmd_vel, subscribe to odometry, and visualize the robot’s state in RViz2. All hands-on examples use the robot_control_demo, frame_demo, and tf2_demo packages.

Learning Objectives

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

• Explain how position and orientation are represented in ROS 2 using geometry_msgs/msg/Pose, Euler angles, and quaternions.

• Convert between axis-angle and quaternion representations and understand the double-cover property.

• Describe the gimbal lock problem and explain why quaternions are preferred in robotics.

• Explain coordinate frames and the TF2 transform tree, including the standard ROS 2 frames (world, map, odom, base_link).

• Publish static and dynamic transforms, and look them up with Buffer and TransformListener.

• Inspect the transform tree using view_frames, tf2_echo, and rqt_tf_tree.

• Drive a differential-drive robot to a goal pose using a proportional controller on cmd_vel.

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 + odometry

  • Localization with AMCL against a saved map

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

  • Sending navigation goals via RViz2 and the NavigateToPose action

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

• Read Nav2 Getting Started and skim Nav2 Concepts.