This is the second post in a series of blog posts documenting the building of a Raspberry Pi Racing car. If you want to start at the beginning, check out this post here.
After a long summer pause and the beginning of cooling temperatures in the northern hemisphere mid-latitudes, J2G2 Version 2 is beginning to take shape. I am exploring the option of adding a third layer to the frame of the car. The third layer adds stability and additional area to house critical components like batteries and the Raspberry Pi Zero. It also give the car a higher vantage point for a future camera mount.
As it stands, Version 2 will not have artificial intelligence or machine learning; however, I anticipate developing the AI/ML capability in Version 3. We will explore what it takes to train an AI/ML model (possibly in Google Cloud) and deploy the AI/ML model to edge. In this case, edge computing means the J2G2 the racing car. In Version 2, I will mostly focus on remote control and chassis design to prepare for later versions of the car.
NVIDIA’s Jetson line of GPU developer kits seem very promising for developing a fully autonomous robot car. According to the NVIDIA website, the Nano is a “powerful computer that lets you run multiple neural networks in parallel for applications like image classification, object detection, segmentation, and speech processing.”
Version 2 is using the Raspberry Pi Zero. The Pi Zero’s technical specifications are shown in the left column below. Compare the Raspberry Pi’s specifications to the Jetson Nano and you will see substantially more computing power. NVIDIA also produces the TX and Xavier GPUs for edge computing, which are even more powerful.
Raspberry Pi Zero W
- 1GHz, single-core CPU
- 512MB RAM
- Mini HDMI and USB On-The-Go ports
- Micro USB power
- HAT-compatible 40-pin header
- Composite video and reset headers
- CSI camera connector
- 802.11 b/g/n wireless LAN
- Bluetooth 4.1
- Bluetooth Low Energy (BLE)