Elvira Part 1 - Intro/Chassis

Elvira Write-up Navigation
Next: Motor Driver

Introduction

When I was living in California before grad school I rented this big ‘ol house with lots of “character”. Among its many quirks, it had what we’ll refer to as a “mild” spider infestation. Like any average, totally normal 20-somethings, my roommates and I would collect black widows to keep as pets and display on our mantle in a pickle jar. They were generationally referred to as “Elvira I-IV”.

Unfortunately once I started grad school in Colorado my access to an arachnid wonderland in the basement was cut short. Feeling lost, confused, and malcontent, I decided to take matters into my own hands…

… and thus Elvira V was born.

Project Description/Goals

Elvira is based on the Darkpaw quadruped chassis from Adeept. Although originally designed to use a Raspberry Pi, I decided that wouldn’t be enough computing horsepower to handle the functionality I plan to implement. As such, I also got an Nvidia Jetson Nano for its brain. Although the Darkpaw kit has code developed already, I plan to write as many of the drivers/low level stuff myself as possible.

I plan to use this as a practical (and frickin’ sweet) platform to explore stuff like:

• Computer Vision
• Machine Learning
• Autonomous Navigation
• Embedded Systems
• State Estimation

Chassis Assembly/Retrofit

The actual assembly of the quadruped chassis was pretty simple and the instructions for the Darkpaw assembly from Adeept were clear enough.

After assembling the stock body, I took some aluminum angle/flat bar and brass standoffs to retrofit the chassis to hold the Jetson. The Jetson and its case/fan/WiFi+Bluetooth antennas are much larger than the Raspberry Pi so I ended up throwing everything on top of the stock Darkpaw chassis.