The Transform System (tf2)

• https://articulatedrobotics.xyz/tutorials/ready-for-ros/tf
  • Also see ROS Tutorials- tf2
  • tf2 package

Transforms in ROS

• 

• Coordinate transformations

  • First we need to assign coordinate systems, or frames, to the components of our system
  • Then we define transforms to tell us the translations and rotations required to go from one frame to another
  • Each frame is defined by only one transform from a parent frame, creating a tree structure

Broadcasting Single Frames

• Use a static_transform_publisher to broadcast a static transform from parent_frame to child_frame

ros2 run tf2_ros static_transform_publisher x y z yaw pitch roll parent_frame child_frame

• Transform from world frame to robot1 frame is 2m in x, 1m in y, 45 degrees yaw rotation

ros2 run tf2_ros static_transform_publisher 2 1 0 0.785 0 0 world robot_1

[INFO] [1736063840.417270047] [static_transform_publisher_zTFbM6lzvgTIEWJz]: Spinning until stopped - publishing transform
translation: ('2.000000', '1.000000', '0.000000')
rotation: ('0.000000', '0.000000', '0.382499', '0.923956')
from 'world' to 'robot_1'

• Transform from robot1 frame to robot2 frame is 1m in x direction

ros2 run tf2_ros static_transform_publisher 1 0 0 0 0 0 robot_1 robot_2

• View the frames and tranforms in RViz

rviz2

• Rerun the first command with larger rotation of 90 deg (1.57 rad)
  • robot_2 will move to since it’s a child of robot_1 (its frame is defined relative to robot_1)

ros2 run tf2_ros static_transform_publisher 2 1 0 1.57 0 0 world robot_1

Broadcasting a Moving Robot

• In a URDF file, each joint can be defined as
  • fixed: robot_state_publisher can just publish a static transform
  • movable: robot_state_publisher must publish a dynamic transform using external values (angles, linear distances, etc) in order to calculate what the transform needs to be at each point in time
    • It gets these values (position, velocity, or effort of a joint) from /joint_states

• JointState messages will normally come from (real or simulated) actuator feedback sensors on the robot (encoders, potentiometers, etc)
  • We can fake this with joint_state_publisher_gui

example_robot.urdf.xacro

example_robot.urdf.xacro

<?xml version="1.0"?>
<robot xmlns:xacro="http://www.ros.org/wiki/xacro"  name="robot">
 
    <material name="grey">
        <color rgba="0.2 0.2 0.2 1"/>
    </material>
 
    <material name="white">
        <color rgba="1 1 1 1"/>
    </material>
 
    <material name="orange">
        <color rgba="1 0.3 0.1 1"/>
    </material>
 
    <material name="blue">
        <color rgba="0 0 1 1"/>
    </material>
 
 
    <link name="world"></link>
 
    <joint name="world_to_base" type="fixed">
        <parent link="world" />
        <child link="base" />
        <origin  rpy="1.57 0 0" xyz="0 0 0" />
    </joint>
 
    <link name="base">
        <visual>
            <origin xyz="2 0 0" rpy="0 0 0"/>
            <geometry>
                <box size="5 0.1 3" />
            </geometry>
            <material name="grey" />
        </visual>
    </link>
 
 
    <joint name="base_to_l3" type="revolute">
        <parent link="base" />
        <child link="l3" />
        <origin  rpy="0 0 0" xyz="0 0.25 0" />
        <axis rpy="0 0 0" xyz="0 0 1" />
        <limit effort="0.1" lower="-3.1" upper="3.1" velocity="0.2"/>
    </joint>
 
 
    <link name="l3">
        <visual>
            <origin xyz="0.75 0 0" rpy="0 1.57 0"/>
            <geometry>
                <cylinder length="1.5" radius="0.04" />
            </geometry>
            <material name="orange" />
        </visual>
    </link>
 
    <joint name="l3_to_l2" type="revolute">
        <parent link="l3" />
        <child link="l2" />
        <!-- <origin xyz="1 0 0" /> -->
        <origin rpy="0 0 0" xyz="1.5 0 0" />
        <axis rpy="0 0 0" xyz="0 0 1" />
        <limit effort="0.1" lower="-3.1" upper="3.1" velocity="0.2"/>
    </joint>
 
 
    <link name="l2">
        <visual>
            <origin xyz="0.75 0 0" rpy="0 1.57 0"/>
            <geometry>
                
                <cylinder length="1.5" radius="0.04" />
            </geometry>
            <material name="orange" />
        </visual>
    </link>
 
    <joint name="l2_to_l1" type="revolute">
        <parent link="l2" />
        <child link="l1" />
        <origin rpy="0 0 0" xyz="1.5 0 0" />
        <axis rpy="0 0 0" xyz="0 0 1" />
        <limit effort="0.1" lower="-3.1" upper="3.1" velocity="0.2"/>
    </joint>
 
 
    <link name="l1">
        <visual>
            <origin xyz="0.5 0 0" rpy="0 1.57 0"/>
            <geometry>
                
                <cylinder length="1" radius="0.04" />
            </geometry>
            <material name="orange" />
        </visual>
    </link>
 
 
    <joint name="l1_to_grip" type="revolute">
        <parent link="l1" />
        <child link="grip" />
        <origin rpy="0 0 0" xyz="1 0 0" />
        <axis rpy="0 0 0" xyz="0 0 1" />
        <limit effort="0.1" lower="-3.1" upper="3.1" velocity="0.2"/>
    </joint>
 
 
    <link name="grip">
        <visual>
            <origin xyz="0.25 0 0"/>
            <geometry>
                <box size="0.5 0.5 0.05" />
            </geometry>
            <material name="white" />
        </visual>
    </link>
 
 
 
    <joint name="world_to_camera" type="fixed">
        <parent link="world" />
        <child link="camera" />
        <origin  rpy="-2 0 2.5" xyz="4 1 2" />
    </joint>
 
    <link name="camera">
        <visual>
            <origin xyz="0 0 -0.2"/>
            <geometry>
                <box size="0.2 0.2 0.4" />
            </geometry>
            <material name="blue" />
        </visual>
    </link>
 
 
</robot>

• Install packages if needed

sudo apt install ros-humble-xacro ros-humble-joint-state-publisher-gui

• Run the robot_state_publisher passing in the URDF file preprocessed by xacro to the robot_description parameter

ros2 run robot_state_publisher robot_state_publisher --ros-args -p robot_description:="$(xacro example_robot.urdf.xacro)"

• Publish the joint states

ros2 run joint_state_publisher_gui joint_state_publisher_gui

• Run RViz

rviz2 -d articubot-tf.rviz

• Add RobotModel and select /robot_description in RViz to view the 3D model of the robot

• Use view_frames from tf2_tools package to visualize the transform tree along with information about how it is being broadcast
  • This is easier than using ros2 topic echo to listen to /tf and /tf_static directly

ros2 run tf2_tools view_frames

[INFO] [1736068659.429880586] [view_frames]: Listening to tf data for 5.0 seconds...
[INFO] [1736068664.497444195] [view_frames]: Generating graph in frames.pdf file...
[INFO] [1736068664.502413522] [view_frames]: Result:tf2_msgs.srv.FrameGraph_Response(frame_yaml="base: \n  parent: 'world'\n  broadcaster: 'default_authority'\n  rate: 10000.000\n  most_recent_transform: 0.000000\n  oldest_transform: 0.000000\n  buffer_length: 0.000\ncamera: \n  parent: 'world'\n  broadcaster: 'default_authority'\n  rate: 10000.000\n  most_recent_transform: 0.000000\n  oldest_transform: 0.000000\n  buffer_length: 0.000\nl3: \n  parent: 'base'\n  broadcaster: 'default_authority'\n  rate: 10.198\n  most_recent_transform: 1736068664.496014\n  oldest_transform: 1736068659.494837\n  buffer_length: 5.001\ngrip: \n  parent: 'l1'\n  broadcaster: 'default_authority'\n  rate: 10.198\n  most_recent_transform: 1736068664.496014\n  oldest_transform: 1736068659.494837\n  buffer_length: 5.001\nl1: \n  parent: 'l2'\n  broadcaster: 'default_authority'\n  rate: 10.198\n  most_recent_transform: 1736068664.496014\n  oldest_transform: 1736068659.494837\n  buffer_length: 5.001\nl2: \n  parent: 'l3'\n  broadcaster: 'default_authority'\n  rate: 10.198\n  most_recent_transform: 1736068664.496014\n  oldest_transform: 1736068659.494837\n  buffer_length: 5.001\n")