Community¶
Here are some sites that are of interest:
- CoderDojo Twin Cities - where you can sign up to be a mentor or student
- Twin Cities AI Racing League Meetup Site - where we announce our public meetings
DonkeyCar Hardware¶
- DonkeyCar web site
- Donkey Car Nano Setup Page
- DonkeyCar Assembly Video - Chris Anderson's detailed assembly video from 2018.
Track Options¶
Hardware Options¶
Raspberry Pi 3, 4, the Nvidia Nano, the Nvdia DX2, and the Intel Mobius Neural Stick The base DonkeyCar today uses the Raspberry Pi 3+ which has a list price of $35. This hardware is just barly able to process images in real-time. Small changes in lighting will throw the car off the track. The new Raspberry Pi 4 with 4GB RAM is a new option.
The Nvidia Nano on the other hand has 128 CUDA core processors and has more than enough power to drive around a track in real time with varied lighting conditions. This is the hardware we have used for our first generation cars in the AI Racing League.
There are also college-level autonomous driving teams that use the more expensive Nvidia DX2 hardware.
Nvidia Nano¶
Jetson Nano References
- Joseph Bastulli PyTorch Nano
- Nvidia Jetson Developer Kit
- Nvidia Jetson Nano Kaya Video
- Adding a Joystick to your DonkeyCar - From Dan McCreary's Blog
Videos¶
- Video of Wide Track
- PID Theory and Steering - why using machine learning is easier than setting PID parameters. This is covered in control theory.
- Real time optimal control of an autonomous RC car with minimum-time maneuvers - nice video of optimization of driving algorithm using a "U" shaped track.
- Sparkfun Autonomous Vehicle Race from 2016
- Ed Murphy on Maker Faire