ROS 2 Human Follower

High-level control software for a mobile robot, leveraging the ros2_control framework to communicate with sensors and a low-level microcontroller, achieving real-time human tracking and following.

The algorithm was successfully implemented and tested on a custom-built mobile platform:

System Requirements

Hardware Specifications

• Main Processing Unit: Raspberry Pi 5 (8GB RAM recommended - handling ROS 2 nodes and computer vision tasks).
• Low-Level Controller: Any microcontroller capable of communicating with the RPi5 via SPI to control the motors (e.g., Raspberry Pi Pico 2, STM32).
• Vision Sensor: Raspberry Pi Camera Module HD v2 (connected via CSI port).
• Distance/Depth Sensor: VL53L8CX (64-zone Time-of-Flight sensor)

Software Specifications

• Operating System: Ubuntu Server 24.04 with Real-Time (RT) kernel enabled
• Middleware: ROS 2 Jazzy Jalisco.
• Control Framework: ros2_control

Installation & Build

1. Create a workspace and clone the repository:

mkdir -p ~/ros2_ws/src
cd ~/ros2_ws/src
git clone [https://github.com/Szymi039/ros2-human-follower.git](https://github.com/Szymi039/ros2-human-follower.git)

2. Install dependencies:

cd ~/ros2_ws
rosdep install --from-paths src -y --ignore-src

3. Build the packages:

colcon build --symlink-install

4. Source the workspace:

source install/setup.bash

Usage & Launch Modes

1. Standard Operation (Human Following)

ros2 launch my_robot_bringup my_robot.launch.py

2. Manual Teleoperation

ros2 launch my_robot_bringup my_robot.launch.py tracker:=false

On another terminal use:

ros2 run teleop_twist_keyboard teleop_twist_keyboard --ros-args -r /cmd_vel:=/diff_drive_controller/cmd_vel -p stamped:=true -p speed:=0.50 -p turn:=4.0

3. Sensor Data Visualization (configure RViz2 for PointCloud2 messages)

ros2 launch my_robot_bringup my_robot.launch.py broadcaster:=true

Example of the PointCloud data visualized in RViz: