naokiakai / mcl3d_ros Goto Github PK

Hello Naoki,
Your work is amazing and your results are great. I think your work can easily outperform other methods when no or noisy INS information is available :-)
While there exist other methods for localization tracking as well, there are not many options for solving the 'kidnapping problem' which is especially important in the robotics field: finding the initial position of the robot in the map (with just a 3D LiDAR).
Because you already use a global optimal solver (MCL) in your approach, do you think the code could be adjusted to solve the kidnapping problem? (finding the initial position in the map, like demonstrated here: https://www.youtube.com/watch?v=gTav7fsAyWc :-) And how could this work with your code.
Thanks in advance for some inspirations.
Regards, Alexander

Hi,

Can you give some more instruction on reproducing some localisation results? For example, which map / bag file we should use to reproduce some localization results, or share some simple demo.

Hello.
First of all thanks for your great research.

I wanted to do the localization in my own prepared environment, I could get the PCD file of the map and convert it to a distance map. However, I am not sure where to read the bag files taken by LiDAR.
Please let me know the path where the bag file is placed and where I should change it in the source code.

If possible, I would appreciate it if you could share the bag file and the map used in the demo shown in the gif image.

Thanks.