Human Following Robot using YOLOv5

This repository contains an implementation of a human following algorithm, allowing a robot to accurately track and follow a person. The algorithm leverages the power of the YOLOv5 object detection model.

Robot

The primary robot used in this implementation is Adeept AWR 4WD WiFi Smart Robot Car Kit for Raspberry Pi 4

Requirements

• Operating System: Raspbian or Ubuntu 20.04, tested on 32
• Middleware: ROS Noetic
• Python version: Python 3.9
• Torch and open CV: Compatible with ARM architecture

Features

• Real-time Human Detection: Using the YOLOv5 model to accurately detect humans in the robot's field of view.
• Smooth Tracking: The robot is capable of smoothly following a person without abrupt movements.

Setup & Installation

1. Setup ROS Noetic:

   • For Ubuntu 20.04: ROS Noetic Installation Guide
   • For Raspbian: [ROS Noetic Installation Guide for Raspbian](See below)

2. Clone the Repository:

   git clone https://github.com/khalidbourr/Human-Following
   cd Human-Following
   

3. Install OpenCV, Torch and Torchvision:

    cd Human-Following
    pip3.9 install opencv_python-4.5.1.48-cp39-cp39-linux_armv7l.whl
    pip3.9 install torch-1.8.1-cp39-cp39-linux_armv7l.whl
    pip3.9 install torchvision-0.9.1-cp39-cp39-linux_armv7l.whl
    - 

4. Install Yolo requirements:

   cd Human-Following/src/yolov5_ros/src/yolov5/
   pip3 install -r requirements.txt

5. Prepare ROS Workspace::

   cd Human-Following
   rosdep update
   rosdep install --from-paths src --ignore-src -r -y --rosdistro noetic

6. Build Workspace::

   catkin build
   
   

Distributed Architecture Overview

To optimize performance, especially when it comes to processing-intensive tasks like object detection using YOLOv5, this project utilizes a distributed architecture. This approach divides the computational responsibilities between the Raspberry Pi and an external laptop. Here's why and how it's set up:

Why Use a Distributed Architecture?

• Performance Constraints of Raspberry Pi: The Raspberry Pi, even in its latest models like the one we're using, is not equipped with a dedicated GPU. This limitation makes the object detection task using neural networks like YOLO quite slow and potentially impractical for real-time applications on the robot.

• Leveraging External Computing Power: By offloading the heavy computational task of object detection to a more powerful laptop equipped with a GPU, we achieve faster detection times. This way, the robot can react in real-time while following a human.

How It Works:

1. Prepare Three Terminals with Raspberry Pi Connection: In each terminal:

   ssh -X Pi@[IP_of_your_robot]
   export ROS_MASTER_URI=http://[robot_IP]:11311
   export ROS_IP=[robot_IP]
   

2. Prepare the Fourth Terminal (Without Raspberry Connection):

   export ROS_MASTER_URI=http://[robot_IP]:11311
   export ROS_IP=[laptop_IP]

Execution:

1. Start the ROS Master (First Terminal):

   roscore

2. Launch the Robot's Essential Components (Second Terminal):

   cd Human-Following/
   source devel/setup.bash
   roslaunch adeept_noetic system.launch
   

3. Run YOLOv5 on the External Laptop (Fourth Terminal):

   cd [path_to_yolo_installation]
   source devel/setup.bash
   roslaunch yolov5_ros yolov5.launch input_image_topic:=/img

4. Execute the Human Following Script (Third Terminal):

   cd Human-Following/
   source devel/setup.bash
   rosrun object_follower object_follower.py
   
   
   

Install ROS Noetic on Raspberry Pi (Raspbian 10 - Buster)

#Setup ROS Repository
sudo sh -c 'echo "deb http://packages.ros.org/ros/ubuntu buster main" > /etc/apt/sources.list.d/ros-noetic.list'
sudo apt-key adv --keyserver 'hkp://keyserver.ubuntu.com:80' --recv-key C1CF6E31E6BADE8868B172B4F42ED6FBAB17C654

#Install Dependencies
sudo apt update
sudo apt-get install -y python-rosdep python-rosinstall-generator python-wstool python-rosinstall build-essential cmake

# Download Dependencies
sudo rosdep init && rosdep update
mkdir ~/ros_catkin_ws && cd ~/ros_catkin_ws
rosinstall_generator ros_comm --rosdistro noetic --deps --wet-only --tar > noetic-ros_comm-wet.rosinstall
wstool init src noetic-ros_comm-wet.rosinstall
rosdep install -y --from-paths src --ignore-src --rosdistro noetic -r --os=debian:buster

# Compile ROS
sudo src/catkin/bin/catkin_make_isolated --install -DCMAKE_BUILD_TYPE=Release --install-space /opt/ros/noetic -j1 -DPYTHON_EXECUTABLE=/usr/bin/python3

# Add message packages
cd ~/ros_catkin_ws
rosinstall_generator navigation --rosdistro noetic --deps --wet-only --tar > noetic-navigation-wet.rosinstall
wstool merge noetic-navigation-wet.rosinstall -t src
wstool update -t src
rosdep install -y --from-paths src --ignore-src --rosdistro noetic -r --os=debian:buster
catkin_make

# Note: If errors arise, consider commenting out problematic lines, typically in CMake files of the Geometry package.

# Add ROS environment variables to your .bashrc
echo "source /opt/ros/noetic/setup.bash" >> ~/.bashrc
source ~/.bashrc