Fix perfect_odometry rosbag

[nahueespinosa]

Proposed changes

Fixes #232.

This patch fixes message timestamps and starting time in the rosbag to match the header stamps recorded with sim time.

• The contents of the messages were not altered.
• The storage backend was changed to mcap.
• The last messages where there was a collision against one of the walls were deleted.

Type of change

• 🐛 Bugfix (change which fixes an issue)
• 🚀 Feature (change which adds functionality)
• 📚 Documentation (change which fixes or extends documentation)

Checklist

• Lint and unit tests (if any) pass locally with my changes
• I have added tests that prove my fix is effective or that my feature works
• I have added necessary documentation (if appropriate)
• All commits have been signed for DCO

[glpuga]

The topics LGTM, but when I tried to run the perfect odometry example rosbag does not start because the MCAP plugin is not installed.

After installing manually ros-humble-rosbag2-storage-mcap it worked just fine. We are probably missing a dependency declaration.

I had started the docker with ./run --build .

[nahueespinosa]

@glpuga Good catch! I originally tested on Rolling.

I reverted to using sqlite3 instead of mcap so no need to add extra dependencies.

[glpuga]

I looked into the results of the beluga_benchmark that run on top of this. The data gathered looks good and can be plotted just fine for beluga_amcl, but nav2_amcl seems to be crashing right after starting.

This can be duplicated by running:

colcon build 
source install/setup.bash
cd src/beluga/beluga_benchmark/docs/reports/2023-06-03
mkdir test
cd test
ros2 run beluga_benchmark parameterized_run 250 300 400 500 750 1000 2000 5000 10000 20000 50000 100000 200000 --params-file ../ref_config/likelihood_params.yaml --package nav2_amcl --executable amcl
cd -

Nothing is left on the logs to indicate why nav2_amcl is giving up.

[nahueespinosa]

@glpuga I can reproduce the error in Humble, but everything works fine in Rolling. More investigation is needed.

[nahueespinosa]

@glpuga I think the failure is most likely related to what @ivanpauno found here: #117 (comment)

I just checked and we have adaptive recovery disabled in those benchmarks, but I don't know if it's worth digging too hard if we know that the Humble version may crash unexpectedly.

We can benchmark using Rolling where the nav2_amcl bug was fixed.
I don't think that the issue is related to the changes in this PR.

[glpuga]

@glpuga I think the failure is most likely related to what @ivanpauno found here: #117 (comment)

I just checked and we have adaptive recovery disabled in those benchmarks, but I don't know if it's worth digging too hard if we know that the Humble version may crash unexpectedly.

We can benchmark using Rolling where the nav2_amcl bug was fixed. I don't think that the issue is related to the changes in this PR.

I see. I think it's ok we benchmark in rolling for the time being. Let's merge this then.