Webots – Python - OpenCV
This project involves simulating two types of robots. The first robot controls and synchronizes two traffic lights using a Python controller. The second robot is a JetBot prototype used to represent two robots that can drive autonomously using the same Python controller, which includes the OpenCV and NumPy libraries. This robot uses a camera for navigation, processing each frame to determine the center of the road and thus follow its path. It can also detect red traffic lights and stop accordingly.
- Sequential control and synchronization system for two traffic lights using FSM
- Navigation system using artificial vision, including road following and red light recognition for traffic lights
This project uses a low-poly model created in Blender that represents a small figure-eight road, containing an intersection with two traffic lights. I've added some houses and trees to enhance the model (small town).
The simulation is based on a Webots world that features a small road within a small town.
At its center is an intersection controlled by a pair of traffic lights to prevent collisions between two JetBot robots.
-
WorldInfo Defines global simulation parameters such as time step, physics, and world behavior.
-
Viewpoint Controls the main 3D camera parameters used for visualization.
-
TexturedBackground Provides background textures and lighting conditions for the environment.
-
Small Town Contains the ground that supports the simulation (solid) and the 3D models that make up the small town, including the road, traffic lights, houses and trees.
World
├── Small Town (Solid)
├── Traffic Lights (Robot)
├── JetBot 1 (Green Robot)
└── JetBot 2 (Blue Robot)
-
Traffic Lights It is the robot that controls the operation of the traffic lights.
-
JetBot 1 & JetBot 2 Robots These are two small robots, models (PROTOs) available from Webots and based on the real open-source robot based on NVIDIA Jetson Nano. The controller for these robots allows them to drive autonomously using artificial vision (OpenCV), recognizing the center of the road, as well as the red light of the traffic lights.
This robot does not include distance sensors, nor a lidar sensor for navigation.
- If solid objects are added to the road that are indistinguishable from the road (color and/or shape), a collision is possible.
- Something similar happens if the objects are outside the areas of interest for machine vision processing; they are practically invisible and can cause collisions.
- If the robots are moving at a considerable speed, they may not detect red traffic lights (just like in real life).
- Speed limits
- Road free of objects or other robots during the journey, except for the intersection where the traffic lights are located.
The purpose of these limitations is to maintain functional behavior within the constraints of the robots..
├── Road length: 1.47 m
├── Radius of curves: 0.50 m
├── Average speed: 0.58 m/s
├── traffic_ctrl.py // Traffic light controller
├── Jetbot_ctrl.py // JetBot robot controller
The controllers are written in Python (fully commented), however the robot controller requires the installation of the OpenCV library as a prerequisite.
- Installing Python
- Open the terminal (Ctrl + Alt + T) and update the list of available packages:
sudo apt update
- Install Python 3 (the default version in most recent distributions):
sudo apt install python3
- Verify the installation by typing:
python3 --version
- Installing OpenCV
- Update the system:
sudo apt update
- To instalin OpenCV library you can use pip to get the most up-to-date version, or apt for a stable version that's well integrated with the system.
pip option:
python3 -m pip install opencv-python
apt option:
sudo apt install python3-opencv
- Verify the installation by typing:
python3 -c "import cv2; print(cv2.__version__)"
- Clone the repository
- Open the world in Webots
- Run the simulation
- Observe: synchronized operation of traffic lights → automatic driving of the two JetBot robots, as well as red light recognition to stop
git clone https://github.com/DrakerDG/Traffic-Lights.git
No manual tuning is required for basic operation.
- Host: Victus by HP Laptop 16-d0xxx
- CPU: 11th Gen Intel i7-11800H (16) @ 4.600GHz
- GPU: NVIDIA GeForce RTX 3060 Mobile / Max-Q
- Memory: 31727MiB
- OS: Ubuntu 24.04.4 LTS x86_64
- Simulator: Webots
- 3D design: Blender
Feedback is welcome in the following areas:
- Improvements to the control logic
- Enhancements and capabilities in visual AI processing
- Performance on different machines
- Improvements to the overall code structure
- Specific optimizations for Webots
Please refer to CONTRIBUTING.md before submitting changes.
The primary goals of this repository are:
- Explore control strategies in simulation
- Study basic driving using artificial vision in an environment that can be reproduced with real robots
- Provide a reference implementation for experimentation and learning
This project is conceived as a support and complement to the JetBot robotic ecosystem, offering:
- A simulation-based environment that can be replicated in real robots
- Controllers written in Python using OpenCV and NumPy
- Reproducible experiments without risk of hardware failure
Under no circumstances should the JetBot robot’s autonomous driving model be used in any real-world vehicle or system where its operation could pose a risk to human safety or well-being.
- All credit for the robot concept belongs to the open-source JetBot real robot project and its contributors.
- This repository aims to add value through simulation, not ownership
- Improvements developed here are intended to be shared openly
Contributors are encouraged to respect the spirit of open collaboration and attribution.
This repository prioritizes:
- Learning over optimization-at-all-costs
- Transparency over abstraction
- Stability and realism over raw speed
Any contribution should align with these principles.
“If this work contributes to improving the understanding or development of simulation in Webots or the real use of robots with the use of artificial vision, it will have fulfilled its purpose.”
-
Blender https://www.blender.org
-
Webots Simulator https://cyberbotics.com
-
JetBot PROTO https://cyberbotics.com/doc/guide/jetbot?version=R2021b#nvidia-jetbot
-
JetBot https://jetbot.org/master
-
HSV reference https://en.wikipedia.org/wiki/HSL_and_HSV