Use updated map at each iteration of Monte Carlo Localisation class

Hi all, I am using Monte Carlo localisation from the robotics toolbox (robotics.MonteCarloLocalization). I am updating the map at each iteration and would like to be able to use the updated map to localise (effectively turning this into SLAM). I tried updating the map in the sensor model but that in itself doesn't seem to make a difference. Re-initialising amcl at each time step doesn't yield good results- and is not what I am after since particles are re-initialised at each time step. Any ideas?

Answer 1

[Ning Zhang]

Hello Christina,

It seems that you are trying to use the robotics.MonteCarloLocalization to localize the robot using map. The issue seems that the map is not updated. There might be a couple of possible places may cause the issue. R2016b was just launched officially. Have you tried this latest release and take a look at this shipped example? http://www.mathworks.com/help/robotics/examples/localize-turtlebot-using-monte-carlo-localization.html#responsive_offcanvas

You also mentioned that your approach is similar to a SLAM method, which actually will agglomerate its mapping along the localization. If you want to start with uniform distributed particles and remove unnecessary particles along the iteration. Check out this part of the shipped example and see if this fits what you requested: http://www.mathworks.com/help/robotics/examples/localize-turtlebot-using-monte-carlo-localization.html#zmw57dd0e7041

If you are building a SLAM, and a point cloud is what you want to use, check out this following link for a shipped model: http://www.mathworks.com/help/vision/examples/3-d-point-cloud-registration-and-stitching.html?searchHighlight=slam