Are you interested in getting started with the NVIDIA Nano Jetbot? The sections below contain links and insights I have gained as I explored the Nano Development kit and Jetbots.
The NVIDIA Jetbot has an excellent set of documentation and tutorials built up over time. It includes video series for those who do not like to read the boring manuals and detailed step-by-step instructions with diagrams on how to build a Jetbot.
Documentation can be found here.
Great question - this is often the best way to get started quickly and 'hacking away' on some Machine Learning examples. Below are two highly-rated kits from Sparkfun for the Jetbot.
- Sparkfun Jetbot - complete robot with 2GB Nano
- Sparkfun Chassis - for those who just need a chassis with motors
For those interested in third party perspectives and other cool things to develop with an NVIDIA Nano
The Jetson Hacks website has an amazing collection of all things NVIDIA Nano-related and great how-to videos.
NVIDIA offers a Jetson Certification Course that validates your knowledge of the basics of the NVIDIA Jetson (the chipset that the NVIDIA Nano uses). Find it here.
- TensorRT - this optimization library from NVIDIA enables many image-based models to run significantly faster (3-4 times) and allows your Jetbot to process incoming camera images much quicker. The git repo for this project is here. Looking for a higher-level theory discussion on what TensorRT does - this is a great article.
- Training your robot in a simulated world - one method to gather training data for your Jetbot is to train it in a simulated world. This article from NVIDIA's developer blog details how to use NVIDIA's Isaac Sim software to gather training data for your robot in a simulated world. Additionally, this tutorial shows you how to get started.