Donkey Car is an open source DIY self-driving platform for small scale cars, originate from http://www.donkeycar.com. It uses a single board computer (Raspberry Pi) with a camera to control an RC car to drive thru tracks autonomously. The Robocar learns from your driving and it will drive autonomously after training.

How does it work?

1. Record training data: Drive your Donkey Car in the track for a few laps to record sufficient training data (camera images, throttle and steering data)
2. Train for a self-driving model: Import the training data into TensorFlow (a machine learning framework developed by google), it will build an autonomous driving model by extracting features from the images and drive according to the image features.
3. Apply the self-driving model: After training, the Robocar will know how to steer in response to new camera images.