🤖 Person Following Robot

A ROS-based person detection and following system for TurtleBot3 robots using YOLO object detection and laser scanner-based obstacle avoidance.

📋 Overview

This project implements an autonomous person-following robot that can:

• Detect and track persons using YOLOv8 computer vision
• Follow detected persons while maintaining safe distance
• Avoid obstacles using laser scanner data
• Navigate around obstacles and return to following behavior
• Search for persons when none are detected

🔧 Hardware Requirements

• TurtleBot3 Waffle Pi robot
• Raspberry Pi camera module
• LiDAR sensor (for obstacle detection)
• Network connection between robot and control computer

💻 Software Dependencies

System Requirements

• Ubuntu 20.04 (recommended)
• ROS Noetic
• Python 3.8+

Required Packages

# ROS packages
sudo apt install ros-noetic-image-transport ros-noetic-cv-bridge ros-noetic-vision-opencv
sudo apt install python3-opencv libopencv-dev ros-noetic-image-proc

# Python packages
sudo apt-get install python-is-python3
sudo apt install python3-pip
pip install ultralytics

📥 Installation

1. Clone this repository to your catkin workspace:

cd ~/catkin_ws/src/
git clone https://github.com/C-H-E-N-Zhihao/Person_Follower_ROS
mv Person_Follower_ROS my_following_person_package
cd my_following_person_package/scripts
chmod +x *.py

2. Build the workspace:

cd ~/catkin_ws
catkin_make
source devel/setup.bash

3. Set up environment variables:

echo "export TURTLEBOT3_MODEL=waffle_pi" >> ~/.bashrc
source ~/.bashrc

🚀 Usage

The system requires multiple terminals to run. Follow these steps in order:

Terminal 1: Start ROS Core

roscore

Terminal 2: Robot Connection and Hardware Launch

# Connect to the robot
ssh <robot_ip>

# Launch robot base
roslaunch turtlebot3_bringup turtlebot3_robot.launch

# In a new terminal on the robot (or execute crtl+Z), launch camera
roslaunch turtlebot3_bringup turtlebot3_rpicamera.launch

Terminal 3: Remote Bringup

roslaunch turtlebot3_bringup turtlebot3_remote.launch

Terminal 4: Run Person Following System

rosrun my_following_person_package main_person_follower.py

🧩 System Architecture

Core Components

1. PersonDetector (person_detector.py)

   • Uses YOLOv8 for real-time person detection
   • Processes camera feed and provides person location data

2. ObstacleDetector (obstacle_detector.py)

   • Processes LiDAR data for obstacle detection
   • Monitors front sector (345°-15°) and left sector (75°-105°)
   • Configurable detection distances

3. RobotController (robot_controller.py)

   • State machine implementation with four states:
     • SEARCHING: Rotating to find a person
     • FOLLOWING: Moving toward detected person
     • AVOIDING: Executing obstacle avoidance maneuver
     • WAITING: Stationary when person is centered and obstacle ahead

4. OdomHelper (odom_helper.py)

   • Provides odometry-based rotation control
   • Enables precise angular movements for obstacle avoidance

5. MoveKobuki (move_robot.py)

   • Low-level movement interface
   • Publishes velocity commands to /cmd_vel topic

State Machine Behavior

• SEARCHING: Robot rotates in place looking for a person
• FOLLOWING: Robot moves forward while adjusting direction to center the person
• AVOIDING: Robot executes a 90° rotation, moves forward, then rotates back
• WAITING: Robot stops when person is centered but obstacle is detected ahead

📡 Topics

Subscribed Topics

• /camera/image (sensor_msgs/Image): Camera feed for person detection
• /scan (sensor_msgs/LaserScan): LiDAR data for obstacle detection
• /odom (nav_msgs/Odometry): Robot odometry for precise rotations

Published Topics

• /cmd_vel (geometry_msgs/Twist): Robot velocity commands

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

🔥 Demo

Demo_PersonFollower.mov

👥 Authors