Install from Scratch on a Raspberry-Pi

Install from Scratch on a Raspberry-Pi

This guide describe how we generate our ISO. Its main use is for us to keep it up-to-date and always accessible for our team. You can freely use it and adapt it to your needs but we don't intend to explain every details here.

Prepare the Raspberry-Pi

  • Burn an SD-Card (16Go or more) using Raspbian Buster with desktop - February 2020.
  • Make sure to allow ssh (touch ssh on the boot partition) or directly work from the Pi board using a mouse, keyboard and screen.
  • Power your Raspberry-Pi and connect it to internet via Ethernet (WiFi is not configure at this point).
  • Once booted, connect to it using ssh and the pi login (ssh pi@raspberrypi.local password raspberry)

Raspi-config

Once booted and logged in the Raspberry-Pi, we need to run raspi-config to setup few things. So, on the Pi bash run sudo raspi-config (We will give the number/letter corresponding to the entry item for each command below)

  • Change user password to reachy: 1
  • Change hostname to reachy: 2-N1
  • Turn on WiFi (you can set the country and your own WiFi if you want but this is not necessary) 2-N2-Cancel-Cancel-
  • Finish and reboot

Install Reachy needed software

  • Reconnect ssh pi@reachy.local (password reachy)
  • Create our own work folder mkdir dev and cd dev
  • Clone Reachy software: git clone https://github.com/pollen-robotics/reachy
  • Checkout latest release cd reachyand git checkout tags/v1.3.0
  • Install atlas for scipy: sudo apt install -y libatlas-base-dev
  • Install deps for opencv: sudo apt install -y libhdf5-dev libhdf5-serial-dev libhdf5-103 libqtgui4 libqtwebkit4 libqt4-test python3-pyqt5 libatlas-base-dev libjasper-dev
  • Install Reachy dependencies using the system python3: pip3 install -r ~/dev/reachy/software/requirements.txt
  • Install Reachy software using the system python3: pip3 install -e ~/dev/reachy/software
  • Get the Access Point and dashboard libraries: cd ~/devand git clone https://github.com/pollen-robotics/RAP
  • Install them: cd ~/dev/RAP/ and sudo bash install.sh (Say Yes both time when asked during install)

Setup Coral toolkit

(see https://coral.ai/docs/accelerator/get-started/)

Install extra stuff

  • Install IPython, Jupyter, Matplotlib: pip3 install ipython jupyter matplotlib
  • Setup ipywidgets:
    • pip3 install ipywidgets
    • /home/pi/.local/bin/jupyter nbextension enable --py widgetsnbextension
  • Install dfu-util: sudo apt install dfu-util

Check if everything is fine

  • Reboot
  • Connect to reachy access point (SSID: Reachy-AP)
  • Connect to the dashboard: http://reachy.local

Generate ISO

  • You will need another Raspberry-Pi with lots of space (64Go) to do that
  • Connect the reachy SD-card to the other Raspberry PI (using an USB adaptor)
  • Create an img using dd (check the disk number using sudo fdisk -l): sudo dd bs=4M if=/dev/sda of=reachy-$(date +%F).img conv=fsync
  • download shrink: wget https://raw.githubusercontent.com/Drewsif/PiShrink/master/pishrink.sh
  • Run sudo bash pishrink.sh -pz reachy-$(date +%F).img
  • Share the create compressed img!