AutonomousDrivingCookbook

Why we included this project

Microsoft Garage's Project Road Runner put this cookbook together for people who want to learn autonomous driving techniques without owning a car or a test track. Everything runs on top of the AirSim simulator, so you can train a virtual vehicle using end-to-end deep learning or distributed reinforcement learning and watch the results play out in a simulated world. The Jupyter notebook format works well here: each scenario comes with its datasets, helper scripts, and binaries, so setup friction stays low and you can focus on the algorithms themselves. Beginners get a gentle path into behavioral cloning, while engineers with more experience can borrow the patterns to build their own simulation-to-real experiments. It is a learning resource, not production software, but as a hands-on map of how self-driving research actually works, it earns its place.