|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |
Launch files
Messages
Services
Plugins
Recent questions tagged mola_state_estimation_smoother at Robotics Stack Exchange
|
mola_state_estimation_smoother package from mola_state_estimation repomola_georeferencing mola_gtsam_factors mola_state_estimation mola_state_estimation_simple mola_state_estimation_smoother |
ROS Distro
|
Package Summary
| Version | 3.0.4 |
| License | GPLv3 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | MOLA implementations of the virtual state estimation API for robots / vehicles |
| Checkout URI | https://github.com/MOLAorg/mola_state_estimation.git |
| VCS Type | git |
| VCS Version | develop |
| Last Updated | 2026-10-13 |
| Dev Status | DEVELOPED |
| Released | RELEASED |
| Contributing |
Help Wanted (-)
Good First Issues (-) Pull Requests to Review (-) |
Package Description
Additional Links
Maintainers
- Jose-Luis Blanco-Claraco
Authors
mola_navstate_smoother
Sliding window Factor-graph data fusion for odometry, IMU, GNSS, and SE(3) pose/twist estimations.
This repository provides:
-
NavStateFuse: C++ class to integrate odometry, IMU, and pose/twist estimations.
See package documentation.
Build and install
Refer to the root MOLA repository.
License
This package is released under the GNU GPL v3 license. Other options available upon request.
Changelog for package mola_state_estimation_smoother
3.0.4 (2026-10-09)
- Merge pull request #83 from MOLAorg/fix/imu-attitude-and-georef-cov Smoother: allow ignoring IMU attitude, warn on placeholder orientation; fix zero-covariance geo-reference
- smoother: allow ignoring IMU attitude, warn if it contradicts
gravity
- imu_attitude_sigma_deg: 0 now disables IMU attitude factors, like
the other IMU sigmas already do.
- Warn when the attitude reported by an IMU disagrees with its own
accelerometer by more than 30 deg. Some drivers publish a constant
placeholder orientation without flagging it as unavailable, which
makes the attitude and gravity factors fight each other.
- New unit test, also covering a zero-covariance geo-reference.
- smoother: floor zero variances of a fixed geo-reference A geo-reference given with an all-zero covariance (the default for an "exactly known" T_enu_to_map, e.g. as loaded from a map file) produced a singular prior and GTSAM failures. Floor its variances to the same tiny value already used for YAML-defined geo-references.
- Contributors: Jose Luis Blanco-Claraco
3.0.3 (2026-10-06)
- Smoother: fuse every non-map odometry frame as relative increments (remove relative_factors_frame_ids_re; relative_pose_increment_sigma_lin/_ang now required)
- Add simulated-sensors + RViz demo launch and a 'try it yourself' docs section
- Fix launch files: env vars not reaching mola-cli, empty navstate args defer to YAML, consistent imu/gnss topic argument names
- Fix deprecated CPose3D::getYawPitchRoll() and unchecked nodiscard load results
- Contributors: Jose Luis Blanco-Claraco
3.0.2 (2026-10-01)
3.0.1 (2026-09-28)
- Smoother: work with GTSAM 4.3 (Kilted, Lyrical, Rolling)
- Anchor gauge freedoms ({map} origin and T_enu_to_map azimuth) instead of relying on weak priors
- test-gauge-anchors: require estimates in the late-map window; document anchor helpers
- test-two-odometries: fix covariance indices of the simulated sources
- Contributors: Jose Luis Blanco-Claraco
3.0.0 (2026-09-28)
- Average IMU readings when decimating instead of keeping one raw sample
- Random-walk increment sigma growth and a Huber kernel on pose factors; make a drifting pose source usable through fuse_pose()
- Always fuse wheel odometry as relative increments
- Implement estimated_trajectory(): read back the optimized keyframe poses
- Port to MRPT 3.x
- Never fuse a dataset's ground truth; add relative-pose factors to fuse_pose()
- Expose additional_isam2_update_steps as an environment knob
- Add a covariance floor on the pose reported to a front end
- Don't throw when writing back a keyframe the fixed-lag smoother marginalized
- Optional relative BetweenFactors for wheel odometry
- Wheel odometry's absolute covariance must accumulate
- MOLA_NAVSTATE_DUMP instrumentation and tunable integrator sigmas
- Never take a time difference through Clock::toDouble()
- Make wheel-odometry motion-model sigmas tunable
- Fix deprecation warning for fixed lag smoother, moved to stable
- Fuse raw IMU gyro readings as angular-velocity observations
- Covariance-aware short-term extrapolation
- Test coverage for the predict-twist low-pass filter
- Preserve predict-twist EMA history on same-stamp solves
- Smooth the predicted motion prior (predict-twist low-pass) and moderate default accel sigmas
- Default the shipped config to the real-time path
- Decouple the map->odom /tf child frame from the odometry source frame
- Publish map->odom from the estimator's own graph variable, plus optional base_link_fused
- Sub-ms estimated_navstate() via async backend and high-rate keyframe decimation
- Add nested profiler entries in the estimated_navstate path
- Don't let a late measurement drag the published stamp backwards
- Don't double-count the motion model when splicing a keyframe
- Actually rebuild the smoother on GTSAM failure
- Fix relocalize mode (GNSS+IMU init against a known map) crashing or never converging
- Fuse IMU absolute attitude into georeferencing and state estimation
- Default viz module to MolaVizImGui with a per-app imgui_app_name
- Serve the {odom_i} prediction frame-local, not via {map}
- Don't crash on estimated_navstate() for an unregistered odometry frame
- Contributors: Jose Luis Blanco-Claraco
2.4.2 (2026-06-04)
File truncated at 100 lines see the full file
Package Dependencies
System Dependencies
Dependant Packages
| Name | Deps |
|---|---|
| mola_lidar_odometry | |
| mola_state_estimation |