TrafficSenseAI - ROS2 Traffic Light Detection System

Prerequisites

• Ubuntu 22.04
• Internet connection
• Root/sudo access

Setup Instructions

1. Add User to Sudoers (if needed)

su -
usermod -aG sudo <your-username>
exit

Log out and log back in for changes to take effect.

2. Clean Existing Conflicts (if any)

sudo rm -f /etc/apt/sources.list.d/vscode.list
sudo rm -f /etc/apt/keyrings/packages.microsoft.gpg
sudo rm -f /usr/share/keyrings/microsoft.gpg
sudo apt update

3. Run Setup Script

Navigate to project directory and run:

cd ~/Desktop/TrafficSenseAI
sudo bash ./bashScriptCOnfig.sh

The script will:

• Install ROS2 Humble
• Install Gazebo simulator
• Install Python tools (numpy, scipy, matplotlib, pandas)
• Install robotics packages (URDF, XACRO, robot state publisher)
• Build the workspace
• Configure environment

4. Source Environment

source ~/.bashrc

Or manually:

source install/setup.bash

5. Build Workspace (if needed)

cd ~/Desktop/TrafficSenseAI
colcon build
source install/setup.bash

Running the Simulation

Launch Gazebo with Robot

cd ~/Desktop/TrafficSenseAI
source install/setup.bash
ros2 launch robot_description gazebo.launch.py

Run Traffic Detection (in new terminal)

cd ~/Desktop/TrafficSenseAI
source install/setup.bash
python3 src/traffic_light_robot/traffic_detector.py

Project Structure

TrafficSenseAI/
├── src/
│   ├── robot_description/      # Robot URDF, meshes, worlds
│   │   ├── launch/
│   │   ├── urdf/
│   │   ├── meshes/
│   │   └── worlds/
│   └── traffic_light_robot/    # Python detection scripts
├── build/                      # Build files
├── install/                    # Installation files
├── log/                        # Build logs
└── bashScriptCOnfig.sh        # Setup script

Troubleshooting

Permission Denied

Ensure user is in sudo group and logged out/in after adding.

Package Not Found

sudo apt update
source /opt/ros/humble/setup.bash

Gazebo Won't Launch

Check if workspace is sourced:

source install/setup.bash

Build Errors

Clean and rebuild:

rm -rf build install log
colcon build

Common Commands

Rebuild workspace:

colcon build

Clean build:

colcon build --cmake-clean-cache

Source workspace:

source install/setup.bash

List available launch files:

ros2 launch robot_description <TAB><TAB>

Check ROS topics:

ros2 topic list

Dependencies

• ROS2 Humble
• Gazebo
• Python 3.10+
• NumPy, SciPy, Matplotlib, Pandas
• robot_state_publisher, joint_state_publisher
• XACRO, URDF tools

Launching the simulation :

bashcd ~/Desktop/TrafficSenseAI source install/setup.bash ros2 launch robot_description gazebo.launch.py

being able to connect and drive the robot using the keybord.

sudo apt install ros-humble-teleop-twist-keyboard xterm -y

< commands to be able to run the different type of simulations that we can use to develop this project here >

Build the packages first:

bashcd ~/Desktop/TrafficSenseAI colcon build --symlink-install source install/setup.bash Autonomous mode: bashros2 launch robot_description autonomous.launch.py Teleop mode (AZERTY script): bash# Terminal 1 - Launch Gazebo: ros2 launch robot_description gazebo.launch.py

Terminal 2 - Run teleop:

source ~/Desktop/TrafficSenseAI/install/setup.bash ros2 run robot_description azerty_teleop.py If azerty_teleop not executable: bashchmod +x ~/Desktop/TrafficSenseAI/src/robot_description/scripts/azerty_teleop.py

the git issue , testing if it is solved

Rebuild:

bash cd ~/Desktop/TrafficSenseAI colcon build --packages-select traffic_light_robot source install/setup.bash Run: bash ros2 launch robot_description autonomous.launch.py

New terminal:

ros2 run traffic_light_robot visualizer_node

launching the controlla-Tunaa

cd ~/Desktop/TrafficSenseAI colcon build --packages-select traffic_light_robot ./auto_tune.sh

#implmenting a custom build gazebo data set collector, trainer , and implmentation

New work planification : 1/ implmentation of the Yolo small version of this system 2/ testing the yolo of this system 3/ implmenting the hybrid version YOLO +HSV 4/ adding those walking person for the world implemntation 5/ refactoring the world into a slighlty better version 6/ setting up the rapport of the system + documenting the different iterations of this sysetm