Monte Carlo Localization with LEGO NXT & MATLAB

Loading...

Sign in or sign up now!
Alert icon
Upgrade to the latest Flash Player for improved playback performance. Upgrade now or more info.
2,673
Loading...
Alert icon
Sign in or sign up now!
Alert icon

Uploaded by on Jun 15, 2010

This is a Monte Carlo Localization demonstration using a LEGO Mindstorms NXT Robot. It uses an IR remote control to control the odometry and the sensors are an ultrasonic and a compass sensor from HiTechnic.

The robot runs an embedded NXC program to generate the odometry motion model, based on the servos tachometers, retrieving the information to a laptop through bluetooth.

The filtering step with the sensor model is processed in MATLAB and uses RWTH toolbox to get the odometry.

The map dimension is 122 x 160cm.
The algorithm estimates x, y and theta and reaches a fairly good stability with about 150 particles.

Link to this comment:

Share to:

Uploader Comments (ricardocrl)

  • What exactly is it supposed to do?

  • @HedgieRobot72

    The robot is trying to localize itself, based on both sensor measurements and control given from a remote control. It's always an estimate (better or worse), never the true position.

  • Call me stupid but what is the robot trying to achieve?

  • @boonce88

    why would you be stupid? for asking? :P

    It's trying to localize itself, based on both sensor measurements and control given from a remote control. It's always an estimate (better or worse), never the true position.

  • That's very cool!

    I'm just starting out a pathfinding project using A* pathfinding and a grid. I hope to be able to run it within the NXT, but I'm not sure I have enough memory for it. This is very cool though. Must have been a b**ch to tweak... :-)

  • @atledreier

    Actually it wasn't very hard to adjust to the best parameters, for instance, for the variances of the gaussians. It was harder to get the odometry model working fine. :-)

    Good luck for your project! I believe it's not easy to do it all in the NXT...

see all

All Comments (10)

Sign In or Sign Up now to post a comment!
  • nerds. 1000 times preferable business than low pay scientific people.

  • How did you decided the first position of the robot , from where the robot starts localization .. And what about filtering the scanned data. Did you use Kalman Filter or any other technique?

  • Okey Dokey, thanks for clearing that up!

Loading...

Alert icon
0 / 00Unsaved Playlist Return to active list
    1. Your queue is empty. Add videos to your queue using this button:
      or sign in to load a different list.
    Loading...Loading...Saving...
    • Clear all videos from this list
    • Learn more