ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question
0

Particle Filter without Laser Scans match

asked 2018-03-03 16:55:34 -0500

Andreluizfc gravatar image

updated 2018-03-06 13:34:58 -0500

Hi,

I'm trying to locate a robot inside a map. I have odometry information, IMU, magnetometer, pressure, a 2d map and the initial position inside the map. But I don't have the /scan topic to provide the correction step and use the AMCL package. Is there a way to do localization with those?

I was trying to use or implement a Particle Filter Approach, for the prediction steps I will use sensor fusion, and for the corrective step I wanted to use fiducial markers in the map.

I was looking at this tutorial for Particle Filter and this is something I'm looking want to have.

Thanks very much!

André Castro.

edit retag flag offensive close merge delete

Comments

2

As you've noticed, AMCL works by comparing laser scans to a map to figure out where your robot is. If you don't have laser scans, that doesn't work. The tutorial you linked to is about bfl, not AMCL, and looks like it might be more relevant.

ahendrix gravatar image ahendrix  ( 2018-03-03 19:20:37 -0500 )edit
3

I suggest you reword your question to remove the AMCL references and spend some time describing the sensors that you do have available for localization. It sounds like you might have some other sensors, but as you've described it now you only have an IMU, and you can't do localization with it alone.

ahendrix gravatar image ahendrix  ( 2018-03-03 19:22:49 -0500 )edit

Updated the question a bit more. Thanks!

Andreluizfc gravatar image Andreluizfc  ( 2018-03-05 18:14:31 -0500 )edit

1 Answer

Sort by » oldest newest most voted
1

answered 2018-03-06 13:18:19 -0500

Gayan Brahmanage gravatar image

In order to perform SLAM or Localization, you need both Proprioceptive sensors (Internal state like an odometer, IMU, wheel encoder) and Exteroceptive sensors (external state like laser scanners). You need to use internal measurements to perform prediction with a given motion model. For correction step, you need to use external measurements like laser scans.

In localization problem, particles are drawn according to internal measurements and motion model. Then, resampling is performed using external sensor measurements and observation model. If you are able to detect fiducial markers in the map, you can use a particle filter.

edit flag offensive delete link more

Comments

Thanks! For correction step I am trying the pose from fiducials. Actually there are no packages and I'm having a hard time trying to adapt this example with BFL.

Andreluizfc gravatar image Andreluizfc  ( 2018-03-06 13:33:16 -0500 )edit

I even don't know if it's possible... But thanks anyway! If there's a possibility to customize, I'm chasing that.

Andreluizfc gravatar image Andreluizfc  ( 2018-03-06 13:33:30 -0500 )edit

Question Tools

2 followers

Stats

Asked: 2018-03-03 16:55:34 -0500

Seen: 552 times

Last updated: Mar 06 '18