🧊 Robot URDF Model

1. Course Introduction

View the A1 Pro 3D URDF robot model in real-time in RViz2. The model pose is driven by IMU data in real-time — pick up the robot and flip or tilt it arbitrarily, and the 3D model in RViz2 will follow the motion in real-time, intuitively displaying the coordinate systems of various sensors, the wheeled chassis, and the joint structure of the 6-DOF robotic arm.

🎯 Course Objectives

1. Understand the basic concepts and structural components of URDF models

2. Master the visualization method of IMU-driven robot 3D pose in RViz2

3. Understand robotic arm joint state feedback (joint_state_publisher) and TF coordinate transform tree

4. Master the usage of visualization launch parameters (pose_source, gui, camera type)

2. Demonstration Examples

2.1 Launch Model Visualization

Run the following command in the terminal:

ros2 launch dofbot_pro_description display_A1pro.launch.py

2.2 Operation and Effect Description

After launching, RViz2 will display the complete 3D structure of A1 Pro, including the Ackerman chassis, 6-DOF robotic arm, and depth camera:

1. IMU 3D Pose Tracking:

   • In default mode (pose_source:=imu), pick up the robot and flip, pitch, or rotate it arbitrarily in the air, and the model in RViz2 will synchronously follow with a 1:1 tilt and rotation (supporting full 3D pose with Roll/Pitch/Yaw).

2. Robotic Arm Joint State Real-time Synchronization:

   • The low-level driver node (Ackman_driver_A1P) reads each servo angle in real-time and publishes to the /arm_status/arm_joint topic.

   • joint_state_publisher subscribes to this state and updates the TF coordinate transforms for each joint of the robotic arm, keeping the arm pose in the 3D model synchronized with the physical entity.

3. GUI Slider Simulation Control:

   • You can pass the gui:=true parameter to launch the graphical slider control panel. When not connected to a real robotic arm or when debugging joints, you can manually drag the sliders to observe each joint's movement:

   xxxxxxxxxx
   ros2 launch dofbot_pro_description display_A1pro.launch.py gui:=true
   

3. Launch Parameters Description

display_A1pro.launch.py supports rich parameter configuration to adapt to different debugging scenarios:

4. URDF Basics

4.1 What is URDF?

URDF (Unified Robot Description Format) is the standard file format in ROS for describing robot models in XML format. It defines the robot's geometric appearance (Mesh), joint connections (Joint), collision bodies (Collision), and physical inertial properties (Inertial).

4.2 Core Tags

4.3 Joint Types

4.4 A1 Pro Coordinate Transform Tree (TF Tree)

The core coordinate system hierarchy of the A1 Pro robot is as follows:

xxxxxxxxxx
odom (odometry reference frame)
 └── base_footprint (chassis ground projection reference)
      └── base_link (chassis center reference)
           ├── imu_link (IMU sensor coordinate system)
           ├── laser_frame (LiDAR coordinate system)
           ├── camera_link (depth camera coordinate system)
           └── arm1_Link (robotic arm base rotation)
                └── arm2_Link (upper arm pitch)
                     └── arm3_Link (forearm pitch)
                          └── arm4_Link (wrist pitch)
                               └── arm5_Link (wrist rotation)
                                    └── rlink1_Link (end effector gripper)

xxxxxxxxxx
ros2 run tf2_tools view_frames
evince frames.pdf

5. Source Code Implementation Principle Analysis

5.1 IMU Pose Computation and TF Publishing (imu_display_tf.py)

In pose_source:=imu mode, the imu_attitude_converter node is responsible for converting the IMU sensor's pose and publishing it as the odom → base_footprint coordinate transform:

​x
def imu_callback(self, msg: Imu):
    # Extract IMU raw quaternion and convert to Euler angles (Roll, Pitch, Yaw)
    roll, pitch, yaw = quaternion_to_euler(
        msg.orientation.x, msg.orientation.y,
        msg.orientation.z, msg.orientation.w
    )

    # Adapt to A1P chassis mounting orientation: swap Pitch and Roll, invert Yaw
    x, y, z, w = euler_to_quaternion(pitch, roll, -yaw)

    # Publish /imu/data for other algorithm nodes to use
    converted = Imu()
    ...
    self.imu_publisher.publish(converted)

    # When publish_tf is True, publish odom → base_footprint 3D pose transform
    if self.publish_tf:
        transform = TransformStamped()
        transform.header.frame_id = self.fixed_frame
        transform.child_frame_id = self.base_frame
        transform.transform.translation.x = self.position[0]
        transform.transform.translation.y = self.position[1]
        transform.transform.translation.z = self.position[2]
        transform.transform.rotation.x = x
        transform.transform.rotation.y = y
        transform.transform.rotation.z = z
        transform.transform.rotation.w = w
        self.tf_broadcaster.sendTransform(transform)

6. Source Code and Resource Path Reference