Skip to content
Ogma Intelligent Systems Corp
Menu
  • Technology Description
  • OgmaNeo
  • Press
  • Contact
  • GitHub
  • YouTube
Menu

Self Driving Car Learns Online and On-board on Raspberry Pi 3

Posted on June 21, 2017October 18, 2024 by Eric Laukien

Hello!

We have been hard at work to create (to our knowledge) the world’s first fully online learning self-driving mini-car!

Using a stock RC car model, we equipped it with a Raspberry Pi 3 along with an Arduino to control the servos/speed controller. We used the latest iteration of our online learning software, EOgmaNeo, to provide the brains of the car.

All inference and learning is done right on the Raspberry Pi’s CPU, no pre-training, no connection to a more powerful machine. The car uses a PiCamera to provide visual inputs and a steam controller to provide steering targets when in training mode.

The model is able to run in real-time with ~10 million synapses at 60 frames per second on the Pi.

Stay tuned for more information and a source code release!

Here is a video of the car in action!

 

Share this:

  • Share on X (Opens in new window) X
  • Share on Facebook (Opens in new window) Facebook

Related

5 thoughts on “Self Driving Car Learns Online and On-board on Raspberry Pi 3”

  1. Kit says:
    March 9, 2018 at 1:53 pm

    Hi Eric,
    I am impressed of the self-driving toycar. I’m going to build a similar one with an old RC car (my son’s toy many years ago), but will use some pretraining data from DroNet (http://rpg.ifi.uzh.ch/dronet.html).

    Is your source code available from github ? I don’t see it, maybe I miss it. May i have a copy for my personal use? I will use RPi3 with a Pi camera. Seems RPi3 cpu is fast enough to do the learning & inference in real time.
    What ML engine do you use ?
    What RL algorithm do you use ?

    Any comments or suggestions for me ? I got this idea a few days ago & started purchasing parts. The Pi camera is still on shipping. 🙂

    Your help or suggestion would help a lot of my father & son’s project.

    Thanks -Kit

    Reply
    1. Eric Laukien says:
      March 9, 2018 at 2:11 pm

      Hello!

      The source code for the library itself is available here. It is free for non-commercial use.

      It shouldn’t be necessary to pre-train it – especially since the camera perspective is very different in that data from what a small car would actually see.

      We don’t use any – as I like to call it – “backpropagation libraries”, since our algorithm doesn’t use any backpropagation. Our algorithm is written from scratch in C++ with bindings to Python, Java, and C#.
      Self-driving is a supervised learning task, so we don’t use any RL.

      If you like, you can come to our Gitter chat room – we can answer any additional questions you have there!

      Reply
      1. kit says:
        March 9, 2018 at 11:57 pm

        Thanks for the quick & detail reply, and invite to the chat room. Would be an exciting place to learn more.
        Thanks 🙂

        Reply
  2. Jamal says:
    October 31, 2018 at 2:09 pm

    Hello sir, I am a student of computer science and I am in my final semester. I want to take your Self Driving car project as my Final Year Project. Can I use it for that purpose please tell me.

    Reply
    1. Eric Laukien says:
      October 31, 2018 at 2:36 pm

      Sure, the software is free for non-commercial use!

      Reply

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Recent Posts

  • AOgmaNeo Lorcan Mini robot demo + CLOgmaNeo
  • Unsupervised Behavioral Learning (UBL) revisited
  • Real2Sim with AOgmaNeo
  • Tutorial: RL with OgmaNeo2
  • Learning to walk (faster) with OgmaNeo2

Recent Comments

  • Eric Laukien on A Tutorial on OgmaNeo2
  • Eduardo on A Tutorial on OgmaNeo2
  • Eric Laukien on Beating Atari Pong on a Raspberry Pi without Backpropagation
  • Joseph Yacura on Beating Atari Pong on a Raspberry Pi without Backpropagation
  • Eric Laukien on Beating Atari Pong on a Raspberry Pi without Backpropagation

Archives

  • June 2022
  • July 2021
  • May 2021
  • June 2020
  • May 2020
  • March 2020
  • October 2019
  • August 2019
  • June 2019
  • February 2018
  • July 2017
  • June 2017
  • March 2017
  • December 2016
  • November 2016

Categories

  • OgmaNeo
  • Press

Meta

  • Log in
  • Entries feed
  • Comments feed
  • WordPress.org
Privacy & Cookies: This site uses cookies. By continuing to use this website, you agree to their use.

To find out more, including how to control cookies, see here: Cookie Policy
© 2026 Ogma Intelligent Systems Corp | Powered by Minimalist Blog WordPress Theme