Skip to content

Particle Filter SLAM#67

Open
DoernerD wants to merge 98 commits intosmarc-project:noetic-develfrom
DoernerD:asko23
Open

Particle Filter SLAM#67
DoernerD wants to merge 98 commits intosmarc-project:noetic-develfrom
DoernerD:asko23

Conversation

@DoernerD
Copy link
Contributor

@DoernerD DoernerD commented Feb 7, 2024

Merging everything done for the visual particle filter SLAM

DoernerD and others added 29 commits October 11, 2023 15:45
The motion model assumed that SAM's on the surface or at least
horizontal. Pitching it messed with the motion model. Fully
integrating all angular velocities and transforming them into
the proper world frame solved the problem and we get good estimates
now.
- Fixed the motion model for the particles. Not sure why it was changed,
probably some of the Stonefish fixes.
- Added flags to the particle filter launch file depending on whether
we rerun simulation or experimental data. It's not nice but works
and we don't need to commment/uncomment lines in the launch file.
- Deleted some old code junk
- Renamed a few functions in the filter to make them clearer.
Updated the motion model in the particles. That fixed stuff.

Also dropped more pose estimation methods for computational purpose when running experimental data.
- Changed the motion model back to only look at the angles and only
integrate the yaw angle
- Changed the covariances to achieve the same performance as the DR has.
- Some refactoring of names in the PF
- Disabled respawning nodes in the DR launch file. Otherwise we cannot
use shell scripts to run the DR because every time you'd kill one of
these nodes, a new one would respawn.
- Fixed the pf motion model (again)
- Randomized initial conditions for sam and the docking station
@KKalem KKalem requested a review from ignaciotb February 7, 2024 21:12
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants