Git Product home page Git Product logo

deepracer-pi's Introduction

DeepRacer for Raspberry Pi 4

This repository provides a port of the DeepRacer software stack to the Raspberry Pi 4. It runs on Ubuntu 22.04 with ROS2 Humble, but with most of original code. To make the image smaller the server image is recommended.

Features

Main features of the port:

  • Previously trained models will work
  • Uses Tensorflow Lite or OpenVINO for inference. Supports the Intel Neural Compute Stick 2 (NCS2/MYRIAD)
  • Can integrate with DREM
  • Single power source¹ - reduces weight and lowers center of gravity
  • Over-the-Air Software Updates

Parts

The following parts are needed:

  • Raspberry 4 or compatible board, recommended 4GB or 8GB RAM, but even 1GB will work as memory consumption is about 700MB.
  • Optional: Intel Neural Compute Stick (NCS) 2
  • Waveshare Servo Driver Hat or compatible PCA9865 servo boards.
  • USB camera with ~60 deg FOV, e.g. original DeepRacer camera or a Raspberry Pi camera.
  • 4 RGB light diodes
  • Misc cables

¹ The Waveshare hat comes with a step-down converter from 7.4V to 5V, and will power the Pi via the GPIO header. Other similar cards will require separate power.

Optional

  • Fan/heat-sink kit for Raspberry PI 4

Software Install

Installation of software is reasonably straight forward, as pre-built packages are provided:

  • Flash an SD card with Ubuntu 22.04 Server for ARM64 using the Raspberry Pi Imager.
  • Boot the SD card, and let it upgrade (this takes some time...)
  • Run git clone https://github.com/larsll/deepracer-pi
  • Run sudo ./install-prerequisites.sh
  • Reboot
  • Run sudo ./install-deepracer.sh

Once installed you can start the stack with sudo /opt/aws/deepracer/start_ros.sh. To ensure a smooth start a camera needs to be plugged in. The launch log will now display in the console.

To automatically start at boot do sudo systemctl enable deepracer-core and to start the service in the background sudo systemctl start deepracer-core.

Changes

Some changes have been made to the code to enable access to GPIO as sysfs layout is different on the Raspberry Pi than on the custom Intel board.

Improvements

Details

PWM Outputs

Channel Purpose Notes
0 Speed controller Remove red cable for stock DeepRacer ESC
1 Steering servo
2 RGB LED Tail light - Red
3 RGB LED Tail light - Green
4 RGB LED Tail light - Blue
5
6
7 RGB LED
8 RGB LED
9 RGB LED
10 RGB LED
11 RGB LED
12 RGB LED
13 RGB LED
14 RGB LED
15 RGB LED

NOTE: Remove the red cable on the stock DeepRacer from the speed controller into PWM channel 0, otherwise you are putting 6V into the servo hat.

LiPo can power both the board and car, 3 pin (balance lead) gets wired to VIN (black and red cables only) to power the board and RPi. The 2 pin power cable goes to the car as normal.

GPIO layout:

  • gpio1 - enables PWM (does not do anything, PWM is always on for the Waveshare board)
  • gpio495-gpio504 - maps to PWM7 to PWM15 on the Hat, to control three RGB leds (those originally on the side of the board)

USB host mode: To connect the USB-C port on the Raspberry Pi 4 to a USB on your computer, then you need to perform two changes:

  • Add dtoverlay=dwc2,dr_mode=host to your /boot/firmware/config.txt
  • Add modules-load=dwc2,g_ether to your /boot/firmware/cmdline.txt

With this you will see a new network connect appear in Windows (and Mac?). The Raspberry Pi will have IP 10.0.0.1; you can connect to the console via https://deepracer.aws. This will require your computer to be disconnected from all other networks (given a DNS priority issue). SSH directly to the IP will work.

What does not (yet) work

  • Battery gauge is not connected - red warning message persists
  • Device Info Node is looking in non-existent places - no real impact

deepracer-pi's People

Contributors

davidfsmith avatar larsll avatar

Stargazers

 avatar  avatar

Watchers

 avatar  avatar  avatar  avatar

deepracer-pi's Issues

Abount Hardware,such as camera

Greetings,

Your project has piqued my interest immensely, and I find the prospect of miniaturizing DeepRacer particularly intriguing. Could you please inform me if there are any specific requirements regarding the hardware composition and cable connections for this project?

For instance, is it imperative to use the official DeepLens camera?

The following image depicts a mini DeepRacer model vehicle I previously encountered, which base on Raspberry Pi.
image
image

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.