rodralez/NaveGo

Canada example for Visual navigation is not working

rodralez opened this issue · 6 comments

rodralez commented

The navego_example_canada_data.m has been tested and is not working.

If FusionCase = "inertial_gnss" option is selected, the system diverge.

If FusionCase = "inertial_visual_gnss", the following error arise:

Error using horzcat
Dimensions of arrays being concatenated are not consistent.

Error in ins_visual_gnss (line 499)
            kf.z = [zv' zp']';

Error in navego_example_canada_data (line 62)
        nav_e = ins_visual_gnss(imu_planetary,gnss_planetary,visual_planetary,'dcm');

Please, download the current branch visual-inertial-gnss and work using the files from these folders. Avoid using code located at other folders.

rodralez commented

Since .txt files were not provided but the .mat did, navego_example_canada_data.m script was modified to
run using the .mat files only. And these .mat data files were uploaded at folder NaveGo/examples/planetary-data/canada-planetary-data/data/ in order the Canada visual example keeps running.

johanndiep commented

Thanks Rodrigo for the input. I realized I forgot to add the .txt files to my local repository, and since I had .mat in .gitignore, the finished parsed .mat files were not pushed to origin visual-inertial-branch.

Your latest push to origin visual-inertial-branch comment out the need for the parsing process. I will add them back in since I find it important for future users how to translate their dataset into the required format needed for NaveGo.

rodralez commented

Johan, yes, please, remove all my changes to the original script if .txt are added.

On the other hand, please check if the error in ins_visual_gnss (line 499) keeps showing.

johanndiep commented

I assume the error should be gone once the required input files are provided.

Note: as of currently, I am not able to test it since I do not have access to Matlab anymore. However, with the provided .txt files, the results should be similar to the results displayed in the following video:
https://www.youtube.com/watch?v=sPnPMBs1gSU&t=135s

rodralez commented

Note: as of currently, I am not able to test it since I do not have access to Matlab anymore.

If you don't have access to MATLAB, this is a problem. Maybe you can try Octave instead (https://www.gnu.org/software/octave/download).

johanndiep commented

That should not be a problem, I was able to run the code couple months ago to produce the results for the conference, since then I have not changed anything. If the paths are correctly set, the same results should come up again.
In the worst case, we can do a screen sharing session and we can check possible bugs.

I updated the origin visual-inertial-branch with the new files and the changes.