Humanoid Robot in Gazebo

Introduction

Simulating humanoid robots in Gazebo allows developers to test bipedal locomotion, balance control, and manipulation algorithms in a safe, repeatable environment before deploying to physical hardware.

Learning Objectives:

• Load and spawn humanoid robot models in Gazebo
• Understand URDF structure for humanoid robots
• Control joints and monitor sensor data
• Test balance and walking controllers in simulation

Theory

Why Simulate Humanoid Robots?

Humanoid robots present unique simulation challenges:

Complexity:

• 20-50+ degrees of freedom (joints)
• Complex kinematic chains (legs, arms, torso, head)
• Underactuated systems (fewer actuators than DOF during flight phase)

Balance Requirements:

• Center of mass management
• Zero-moment point (ZMP) stability
• Dynamic gait generation

Contact Dynamics:

• Foot-ground contact modeling
• Friction and slip
• Impact forces during walking

Benefits of Gazebo for Humanoids:

• Test controllers without hardware damage
• Iterate gait parameters rapidly
• Simulate sensor feedback (IMU, joint encoders, force sensors)
• Debug balance failures safely

Humanoid Robot URDF Structure

A URDF (Unified Robot Description Format) defines the robot's:

<robot name="humanoid">
  <!-- Base Link (Torso) -->
  <link name="base_link">
    <inertial>...</inertial>
    <visual>...</visual>
    <collision>...</collision>
  </link>

  <!-- Joint connecting torso to left hip -->
  <joint name="left_hip_yaw" type="revolute">
    <parent link="base_link"/>
    <child link="left_hip"/>
    <axis xyz="0 0 1"/>
    <limit effort="100" lower="-0.5" upper="0.5" velocity="5.0"/>
  </joint>

  <!-- Continue for all joints: hips, knees, ankles, arms, neck -->
</robot>

Key Components:

• Links: Rigid bodies (torso, thighs, shins, feet, arms)
• Joints: Connections with motion constraints (revolute, prismatic)
• Inertia: Mass distribution for dynamics
• Collision: Simplified geometry for physics
• Visual: Detailed meshes for rendering

Gazebo Plugins for Humanoid Control

Gazebo uses plugins to interface with ROS 2:

1. Joint State Publisher

• Publishes joint positions, velocities, efforts
• Topic: /joint_states

2. Joint Trajectory Controller

• Accepts trajectory commands
• Controls multiple joints simultaneously
• Topic: /trajectory_controller/command

3. IMU Plugin

• Simulates inertial measurement unit
• Provides orientation, angular velocity, linear acceleration
• Topic: /imu/data

4. Contact Sensor Plugin

• Detects foot-ground contact
• Critical for gait state estimation
• Topic: /contact_sensor

Practical Example: Loading a Humanoid in Gazebo

Step 1: Install Dependencies

# Install Gazebo and ROS 2 packages
sudo apt install ros-humble-gazebo-ros-pkgs
sudo apt install ros-humble-joint-state-publisher
sudo apt install ros-humble-robot-state-publisher

Step 2: Create a Simple Humanoid URDF

Create simple_humanoid.urdf:

<?xml version="1.0"?>
<robot name="simple_humanoid">
  <!-- Base Link (Torso) -->
  <link name="torso">
    <inertial>
      <mass value="20.0"/>
      <inertia ixx="0.5" ixy="0" ixz="0" iyy="0.5" iyz="0" izz="0.5"/>
    </inertial>
    <visual>
      <geometry>
        <box size="0.3 0.2 0.5"/>
      </geometry>
      <material name="blue">
        <color rgba="0 0 1 1"/>
      </material>
    </visual>
    <collision>
      <geometry>
        <box size="0.3 0.2 0.5"/>
      </geometry>
    </collision>
  </link>

  <!-- Left Hip Joint -->
  <joint name="left_hip_pitch" type="revolute">
    <parent link="torso"/>
    <child link="left_thigh"/>
    <origin xyz="0 0.1 -0.25" rpy="0 0 0"/>
    <axis xyz="0 1 0"/>
    <limit effort="100" lower="-1.57" upper="1.57" velocity="10"/>
  </joint>

  <!-- Left Thigh -->
  <link name="left_thigh">
    <inertial>
      <mass value="5.0"/>
      <inertia ixx="0.1" ixy="0" ixz="0" iyy="0.1" iyz="0" izz="0.01"/>
    </inertial>
    <visual>
      <geometry>
        <cylinder radius="0.05" length="0.4"/>
      </geometry>
      <material name="gray">
        <color rgba="0.5 0.5 0.5 1"/>
      </material>
    </visual>
    <collision>
      <geometry>
        <cylinder radius="0.05" length="0.4"/>
      </geometry>
    </collision>
  </link>

  <!-- Add more joints and links for complete humanoid -->
  <!-- Right leg, arms, head, etc. -->
</robot>

Step 3: Create Launch File

Create humanoid_gazebo.launch.py:

from launch import LaunchDescription
from launch.actions import IncludeLaunchDescription
from launch.launch_description_sources import PythonLaunchDescriptionSource
from launch_ros.actions import Node
import os
from ament_index_python.packages import get_package_share_directory

def generate_launch_description():
    # Path to URDF file
    urdf_file = os.path.join(
        get_package_share_directory('my_humanoid_description'),
        'urdf',
        'simple_humanoid.urdf'
    )

    # Read URDF content
    with open(urdf_file, 'r') as file:
        robot_desc = file.read()

    # Spawn robot in Gazebo
    spawn_entity = Node(
        package='gazebo_ros',
        executable='spawn_entity.py',
        arguments=[
            '-entity', 'simple_humanoid',
            '-topic', '/robot_description',
            '-x', '0',
            '-y', '0',
            '-z', '1.0'  # Spawn 1m above ground
        ],
        output='screen'
    )

    # Robot state publisher
    robot_state_publisher = Node(
        package='robot_state_publisher',
        executable='robot_state_publisher',
        parameters=[{'robot_description': robot_desc}]
    )

    # Launch Gazebo
    gazebo = IncludeLaunchDescription(
        PythonLaunchDescriptionSource([
            os.path.join(get_package_share_directory('gazebo_ros'),
                         'launch', 'gazebo.launch.py')
        ])
    )

    return LaunchDescription([
        gazebo,
        robot_state_publisher,
        spawn_entity
    ])

Step 4: Launch Simulation

# Source ROS 2
source /opt/ros/humble/setup.bash

# Launch Gazebo with humanoid
ros2 launch my_humanoid_description humanoid_gazebo.launch.py

Controlling the Humanoid

Method 1: Joint Position Commands

# Publish joint commands
ros2 topic pub /joint_commands sensor_msgs/msg/JointState \
  "{name: ['left_hip_pitch', 'right_hip_pitch'], \
    position: [0.5, -0.5]}"

Method 2: ROS 2 Control

Use ros2_control for trajectory tracking:

# controller_config.yaml
controller_manager:
  ros__parameters:
    update_rate: 100  # Hz

    joint_trajectory_controller:
      type: joint_trajectory_controller/JointTrajectoryController

joint_trajectory_controller:
  ros__parameters:
    joints:
      - left_hip_pitch
      - left_knee
      - left_ankle
      - right_hip_pitch
      - right_knee
      - right_ankle

    command_interfaces:
      - position

    state_interfaces:
      - position
      - velocity

Monitoring Sensor Data

View Joint States

# Monitor joint positions and velocities
ros2 topic echo /joint_states

View IMU Data

# Monitor orientation and angular velocity
ros2 topic echo /imu/data

Visualize in RViz

# Launch RViz to visualize robot state
ros2 run rviz2 rviz2

In RViz:

1. Set Fixed Frame to torso
2. Add RobotModel display
3. Add TF display to see coordinate frames

Common Challenges

Issue 1: Robot Falls Immediately

Cause: Incorrect initial pose or unstable spawn position

Solution:

• Spawn with feet on ground (z = foot_height)
• Ensure center of mass is above support polygon
• Use initial joint positions for balanced pose

Issue 2: Jittery Motion

Cause: Physics timestep too large or PID gains not tuned

Solution:

<!-- In Gazebo world file -->
<physics type="ode">
  <max_step_size>0.001</max_step_size>  <!-- 1ms timestep -->
  <real_time_factor>1.0</real_time_factor>
</physics>

Issue 3: Unrealistic Contact

Cause: Default friction coefficients

Solution:

<!-- Add to foot link collision -->
<collision>
  <surface>
    <friction>
      <ode>
        <mu>1.0</mu>  <!-- Coefficient of friction -->
        <mu2>1.0</mu2>
      </ode>
    </friction>
  </surface>
</collision>

Exercises

1. Spawn a humanoid: Use an existing URDF (e.g., from robotics tutorials) and spawn it in Gazebo
2. Command joint positions: Move the robot's arms to specific angles using ros2 topic pub
3. Monitor IMU: Tilt the robot and observe IMU readings - do they match expected values?
4. Modify URDF: Change the mass of the torso and observe how it affects balance
5. Gait experiment: Research a simple walking controller and attempt to implement it

Summary

Simulating humanoid robots in Gazebo provides a safe testbed for developing locomotion, balance, and manipulation algorithms. By modeling robots in URDF, spawning them in Gazebo, and interfacing via ROS 2 topics and services, developers can iterate rapidly before deploying to physical hardware.

Further Reading

• Gazebo Tutorials - ROS 2 Integration
• URDF Tutorials
• Humanoid Robotics: A Reference - Kajita et al.
• Atlas Humanoid Simulation