AutonomousVehicleControlBeginnersGuide

AutonomousVehicleControlBeginnersGuide

Learn autonomous driving by building it. Python simulations and an open textbook of self-driving car algorithms (localization, mapping, path planning, path tracking, perception and control), and of the software architecture that connects them into one autonomous driving system. The system grows one stage at a time, and you can swap and compare the algorithms of each stage in the same system.

Read the book · See all simulations · Quick start

Table of Contents

• About this project
• Requirements
• Quick Start
• Examples of Simulation
• The book
• License
• Citing
• Use Case
• Contribution
• Author

About this project

Every algorithm is a small Python simulation with an animation, written to be easy to read rather than fast or complete. You can run each one with a single command, change its parameters and see what happens. Fork it and use it for studying, teaching or your own work.

Beyond single algorithms, the project shows how they work together: how the stages of an autonomous driving system pass data to each other, why the true state and the estimates are kept apart, and how small interfaces let you swap one algorithm for another. The architecture and roadmap explain the design, which the code is moving toward step by step (see the roadmap issue).

Requirements

Please satisfy with the following requirements on native or VM Linux in advance.
For running each sample codes:

• Python 3.13.x
• Matplotlib
• NumPy
• SciPy
• CasADi (for MPC path tracking)
• do-mpc (for MPC path tracking)

For development:

• pytest (for unit tests)

For setting up the environment with Docker (see Quick Start):

• Docker
• VS Code with the Dev Containers extension (only for option A)

Quick Start

1. Clone this repository

$ git clone https://github.com/ShisatoYano/AutonomousVehicleControlBeginnersGuide
$ cd AutonomousVehicleControlBeginnersGuide

2. Set up the environment

Choose one of the following options.

Note for options A and B: The container shows the animations through your host's X11 display.

• Linux: Allow local connections to your X server before starting the container. This setting is reset when you log out, so run it again after logging in (or run it automatically at login).

  $ xhost +local:
  

  This allows any local process on your machine to connect to your X server, so be careful on a shared machine.
• Windows / macOS: You need an X server that the container can connect to, such as WSLg on WSL2 or XQuartz on macOS.

A. VS Code + Docker (Dev Container)

1. Install Docker and VS Code with the Dev Containers extension
2. Open this repository's folder in VS Code
3. Run Dev Containers: Reopen in Container from the Command Palette
4. All required libraries are installed automatically. Run the commands in the following steps in VS Code's terminal

B. Docker only

1. Install Docker
2. Build and start the container, and then open a shell in it

   $ docker compose up -d --build
   $ docker compose exec simulation bash
   

3. Run the commands in the following steps in this shell. To stop the container, run docker compose down on your host

C. Local (without Docker)

1. Install Python 3.13.x
2. Create a virtual environment and install the required libraries
   • Linux / macOS:

     $ python3 -m venv .venv
     $ source .venv/bin/activate
     $ pip install -r requirements.txt
     

   • Windows (PowerShell):

     > python -m venv .venv
     > .venv\Scripts\Activate.ps1
     > pip install -r requirements.txt
     

3. Run a simulation

Each simulation runs with one command. For example, localization with an Extended Kalman Filter, where the vehicle estimates its position from sensor observations while following a course:

$ python src/simulations/localization/extended_kalman_filter_localization/extended_kalman_filter_localization.py

You can find all simulations in doc/EXAMPLES.md.

4. (Optional) Run unit tests

You can confirm that the environment is set up correctly by running the unit tests.

$ . run_test_suites.sh

On Windows without Docker, run run_test_suites.bat instead.

Add star to this repository if you like it!!

Examples of Simulation

Every algorithm has a runnable simulation with an animated demo, in these categories: localization, mapping, path planning, path tracking, perception and course generation. The full list with the demo GIFs is generated from src/simulations: see doc/EXAMPLES.md.

The book

The explanations are published as a free web book: https://shisatoyano.github.io/AutonomousVehicleControlBeginnersGuide/. Each chapter adds one stage to the system, and explains the theory with equations and figures, where each equation lives in the code, and experiments to try. It is being written in the open (see the documentation roadmap). Comments and ideas are welcome in the Book category of Discussions, and anyone can help write a chapter: see the writing guide.

License

MIT

Citing

If you use this project in your research or teaching, please cite it. Click "Cite this repository" on the right side of the repository page to get the citation in APA or BibTeX format, or see CITATION.cff.

Use Case

I started this project to study an algorithm and software development for Autonomous Vehicle system by myself. You can also use this repo for your own studying, education, researching and development.

If this project helps your task, please let me know in Show and tell.
Any paper, animation, video as your output, always welcome!! It will encourage me to continue this project.

Your comment and output is added to this list of user comments.

Contribution

Any contribution by creating an issue or sending a pull request is welcome!! Please check this document about how to contribute.
Questions about the algorithms or the code are welcome in Q&A, and feedback on the book in Book.
If you are looking for a place to start, see the good first issues and the roadmap issue.

Author

Shisato Yano