Semubot ROS2 Integration

This project implements a modular ROS2-based framework for controlling Semubot, the first open-source humanoid robot developed in Estonia. The system is split between a laptop (for visualization and control) and a Raspberry Pi 4B (for hardware motor control).

Project Objectives

• Build a URDF model of Semubot using STL files.
• Visualize the robot in RViz2.
• Implement robot movement using ROS2 /cmd_vel topic.
• Deploy ROS2 nodes on Raspberry Pi to control real motors via GPIO.

ROS2 Workspaces

1. semubot_ros2_ws (Laptop)

Includes:

• RViz2 visualization
• URDF robot model
• /cmd_vel publishing
• Motion simulation logic

2. semubot_raspberry_ws (Raspberry Pi 4B)

Includes:

• Motor control node
• GPIO command conversion

Installation Guide

1. Install ROS2 Jazzy

Follow the official guide:
https://docs.ros.org/en/jazzy/Installation/Ubuntu-Install-Debs.html

Once installed, source ROS2:

source /opt/ros/jazzy/setup.bash

2. Clone Workspaces

On Laptop:

git clone https://github.com/SemuBot/Huseynli-thesis-2025-ROS-integration.git
cd semubot_ros2_ws
source /opt/ros/jazzy/setup.bash
colcon build --symlink-install
source install/setup.bash

On Raspberry Pi:

git clone https://github.com/SemuBot/Huseynli-thesis-2025-ROS-integration.git
cd semubot_raspberry_ws
source /opt/ros/jazzy/setup.bash
colcon build --symlink-install
source install/setup.bash

Running the System

On Laptop:

ros2 launch semubot display_robot.launch.py

On Raspberry Pi:

ros2 run semubot_motor cmd_vel_motor_driver.py

ROS2 Networking

Ensure both the laptop and Raspberry Pi are on the same subnet (LAN). No ROS master is required as ROS2 uses DDS discovery protocol.

To verify connectivity:

ros2 topic list

You should see /cmd_vel on both devices.

Topics Overview

Topic	Type	Description
/cmd_vel	geometry_msgs/Twist	Velocity commands
/joint_states	sensor_msgs/JointState	Simulated joint data
/tf	tf2_msgs/TFMessage	Transform frames
/odom	nav_msgs/Odometry	Odometry (simulated)

Author

Omar Huseynli
University of Tartu – Robotics & Bioengineering
Email: omar.2005.22.04@gmail.com