Skip to main content

Understanding URDF for Humanoid Robots

URDF (Unified Robot Description Format) is an XML format used in ROS 2 to describe all aspects of a robot. It allows you to model the robot's kinematics (links and joints), visual properties (geometry, color, texture), and collision properties. For humanoid robots, a well-structured URDF is crucial for simulation, motion planning, and visualization.

URDF Structure Basics​

A URDF file is essentially a tree of link and joint elements. A link represents a rigid body (e.g., a torso, an arm, a leg), and a joint connects two links, defining their relative motion.

The <robot> Tag​

Every URDF file starts with a <robot> tag, which gives the robot a name.

<robot name="my_humanoid">
  <!-- Robot description goes here -->
</robot>

A link element describes the physical and visual properties of a robot segment. It includes:

  • <visual>: Defines how the link appears (geometry, material).
  • <collision>: Defines the link's collision properties (geometry).
  • <inertial>: Defines the link's mass, center of mass, and inertia tensor, critical for physics simulation.

Example of a simple humanoid link (e.g., a torso):

<link name="torso">
  <visual>
    <geometry>
      <box size="0.2 0.4 0.6" />
    </geometry>
    <material name="blue">
      <color rgba="0 0 0.8 1" />
    </material>
  </visual>
  <collision>
    <geometry>
      <box size="0.2 0.4 0.6" />
    </geometry>
  </collision>
  <inertial>
    <mass value="10.0" />
    <origin xyz="0 0 0.3" rpy="0 0 0"/>
    <inertia ixx="1.0" ixy="0.0" ixz="0.0" iyy="1.0" iyz="0.0" izz="1.0" />
  </inertial>
</link>

<joint> Element​

A joint element connects a parent link to a child link and defines the type of motion allowed between them. Key attributes include:

  • name: Unique name for the joint.
  • type: Type of joint (e.g., revolute, continuous, prismatic, fixed). Humanoids typically use revolute for rotational joints (like shoulders, elbows) and fixed for rigid connections.
  • <parent> and <child>: Specify the connected links.
  • <origin>: Defines the joint's position and orientation relative to the parent link.
  • <axis>: For revolute/prismatic joints, defines the axis of motion.
  • <limit>: Defines the joint's range of motion and velocity/effort limits.

Example of a humanoid shoulder joint:

<joint name="shoulder_joint" type="revolute">
  <parent link="torso"/>
  <child link="upper_arm"/>
  <origin xyz="0.1 0.2 0.5" rpy="0 0 0"/>
  <axis xyz="1 0 0"/>
  <limit lower="-1.57" upper="1.57" effort="100" velocity="1.0"/>
</joint>

Building a URDF for a humanoid involves carefully defining all its segments and their interconnections.

Humanoid robots typically include a complex hierarchy of links:

  • Base/Torso: The central part of the robot.
  • Head: Connected to the torso via a neck joint.
  • Arms: Composed of upper arm, forearm, and hand links, connected by shoulder, elbow, and wrist joints.
  • Legs: Composed of thigh, shin, and foot links, connected by hip, knee, and ankle joints.
  • End Effectors: Grippers or specialized tools if applicable.

Common Humanoid Joints​

  • Revolute Joints: Most common for humanoids, allowing rotation around a single axis (e.g., elbow, knee, hip, shoulder).
  • Fixed Joints: Used for rigid connections where no relative motion is intended (e.g., connecting a camera to a head link).
  • Continuous Joints: Revolute joints without limits (e.g., a wheel that can spin infinitely, though less common for primary humanoid joints).

Integrating Sensors​

Sensors are critical for humanoids to perceive their environment and their own state. In URDF, sensors are typically represented as link elements (often small, fixed links) and then referenced in simulation environments or through ROS 2 components.

Common sensors for humanoids include:

  • Cameras: For vision and object detection. 
    <link name="camera_link">
      <visual>
        <geometry><box size="0.02 0.05 0.02" /></geometry>
      </visual>
    </link>
    <joint name="camera_joint" type="fixed">
      <parent link="head"/>
      <child link="camera_link"/>
      <origin xyz="0.05 0 0.05" rpy="0 1.5707 0"/>
    </joint>
  • IMUs (Inertial Measurement Units): For orientation, angular velocity, and acceleration.
  • Force/Torque Sensors: At joints or end effectors for interaction control.
  • Lidars/Depth Sensors: For environmental mapping and obstacle avoidance.

These sensors are often added as new links connected by fixed joints to a relevant parent link (e.g., a camera to the head, an IMU to the torso). Their actual data publication is handled by ROS 2 drivers and simulation plugins, not directly within the URDF.

Best Practices for Humanoid URDFs​

  • Modularity: Break down the robot into logical components (e.g., _macro.xacro files for arms, legs, etc.) using XACRO (XML Macros for URDF) for easier management and reuse.
  • Consistent Naming: Use clear and consistent naming conventions for links and joints.
  • Origin Placement: Carefully define <origin> tags to ensure correct assembly and joint axes.
  • Collision Geometry: Simplify collision geometries where possible to reduce simulation computational load, but ensure they accurately represent the robot's physical extent.
  • Visualization: Use accurate meshes (.dae, .stl) for visual representation rather than simple geometric primitives where detail is important.

Conclusion​

URDF is a powerful tool for describing the physical and kinematic properties of humanoid robots within the ROS 2 ecosystem. By understanding how to define links, joints, and integrate sensors, you can create accurate and functional robot models essential for simulation, control, and visualization. Mastering URDF is a foundational step in developing advanced humanoid robotics applications. The next steps would typically involve integrating this URDF with a simulation environment like Gazebo or a real robot.